Googler here. My opinions are my own; I have no non-public knowledge on this topic.
I do not understand the hyperbole around this report. Nowhere does it say that Google plans to shutter GCP if it doesn't reach #1 by 2023.
> said people with knowledge of the matter.
As in, people who may not have even been at said meeting which took place nearly two years ago, under a different VP and now a different CEO, and saw some meeting notes.
> The group’s leaders told staffers that if Google couldn’t reach a certain size with its computing and storage business—two of the most commonly used cloud services—the cloud unit might never become profitable, the person said. To reach such scale, they said, Google would need to be in a top two position in the market.
So it's not about being #1, it's about being profitable.
In other words, this was a conversation about literally every product at an executive level ever.
You set goals, you decide what happens when those goals are met. Furthermore, there is no concrete assertion here that what happens if this deadline is passed. "at risk of losing funding" can literally mean anything from changing FTE allocation to capital expenses to a million other things that get discussed at an executive level.
And finally:
> A Google spokesperson declined to comment prior to the publication of this story, but after it appeared released the following statement: "Reports of these conversations from 2018 are simply not accurate."
There may be some truth to these accounts--reality is often between the lines--but this is extremely soft on details and actual first person accounts.
> So it's not about being #1, it's about being profitable.
The problem Google has with the developer community is largely around the fact that Google has a long history of pulling the rug out from our feet. There have been 3 stories on HN in the past couple weeks alone about Google products which have either been dropped or changed significantly such that they are no longer viable for developers to use.
Google has lost a lot of trust in the community and if Google management wants to avoid this kind of reaction, they need to build a reputation for stability. Years of dropping products people rely has eroded much of the good will people used to have for Google.
Amazon doesn't drop products.
Microsoft has spent decades building a reputation for having a long term reliable product road-map.
Google is fickle and doesn't seem to care when they screw people over by changing terms underneath us. As the old saying goes—fool me once, shame on you; fool me twice, shame on me.
This is the entire problem. Swap “Google” for “AWS” in the article and you’d have legions of folks calling foul on the article. Instead, most folks are nodding along to the article going “yep, seems like something Google would do all right…”
Google’s problem remains its reputation for turning things off.
The issue is that folks here think GCP is the same thing as Google Reader. That's absurd.
Google is currently building a massive $500 million datacenter outside of Reno as we speak, and has 10+ billion invested in their datacenter cloud offering buildout this year alone.
GCP and the entire physical infrastructure isn't going to get "turned off" if it isn't #1 in the market in 3 years, that's just silly.
The issue for me isn't Reader itself but what it says about the disconnect between Google management and us, the users of Google products. That seems to have happened around the time Google went from _having_ a "user base" to being effectively ubiquitous. I can't speak for others, but Google of the 2000s "got" the hacker/nerd mindset in a way that was the closest thing to empathy I've ever felt from a corporation. That sounds naive to write down here in almost-2020, but that was how it felt at the time. When Reader died, so did that feeling, and now all of us are running around unsure if the Google products we personally care about are the ones Google management care about. Leaks like this one do not inspire confidence in that.
Free consumer products are not the same as paid enterprise services. The markets are entirely different, and ascribing the fate of Google Reader as a sign of anything on the B2B front doesn't make sense.
> Free consumer products are not the same as paid enterprise services.
They may have different revenue models, but there is little difference when it comes down to management killing services which aren't profitable.
If there is a difference, it is that Reader could be maintained at little cost, while GCE (when it gets the axe) could potentially be costing google hundreds of millions per year to keep running.
There have been plenty of other cases where google has burned bridges with the technical crowd which they should have embraced as champions for selling services. Killing off XMPP support was another prime one for me, and absolutely affected paid enterprise customers as well.
Reader could only be maintained at little cost if its dependencies were stable. But stable infra doesn't get you promoted, so every service needs to be staffed indefinitely (to keep it on the upgrade treadmill) or killed gracefully before prod explodes.
To be fair, the switch to mobile-first did more to kill XMPP than Google did. It's a lot better now a decade later with all the XEPs, but now there's nobody there to talk to so it's too late.
Just a cursory look at killedbygoogle.com will show you a list of products that fall in the category of not free and used by business (Hired by Google) and/or were products that developers integrated with their own software that was killed.
As a developer or a company, why should I trust Google as a platform?
It's not just Reader and all the other services. It's really crappy customer service too. Reddit's got a good thread going about someone that paid up on GSUITE that got their account locked. You have to hope a tweet goes viral to get help.
That's irrelevant. Its about trust, commitment and sticking it out. Google's penchant of throwing shit against a wall to see what sticks seems to prevent them from applying themselves to increase the stickiness of their shit.
You simply cannot trust Google to do right by me, the user. That's it.
See, the thing is though, that they kill stuff that DOES stick as well. There's plenty of stuff in the Google Graveyard that was still widely used, with large user base, and they kill it anyway. There's no consistent logic to their killing of products. Sometimes they simply don't want to update it, sometimes it's purely internal politics (G+), sometimes there's just no reason we can find.
What people believe about them matters, not whatever internal rationalizations may or may not have occurred.
Aside from their bad reputation for killing products, they also have a bad reputation for providing next to zero human support services for their products - even if you do pay for them.
AWS has launched a hundred small products which anyone can signup for on the spot, it’s not that different than consumer product releases. And it’s also not the opaque enterprise sales process of Oracle and IBM where things move at a way slower pace.
It’s somewhere in between and if Google started launching a bunch of new DB and other infrastructure products I’d be worried about adopting them individually - not worried that the entire data centres and business divisions getting shut down. Because Google will probably move on to another product without thought of the last one.
Agreed that Google isn’t going to announce one day that they’re going to demolish all the datacenters in 30 days, but that doesn’t mean that GCP has any guaranteed future either. If Google leadership decides to walk away from cloud hosting, it’ll come in the form of reallocated engineers, reduced headcount, smaller conferences, and a general transition to “maintenance mode”. (At which point it’s a corporate backwater and everyone who can transfer out will.)
But for companies picking a cloud platform tomorrow, that’s a fate as good as death! AWS and Azure won’t slow down development and new features, so a couple years after GCP stops being a priority, it’ll start also being a bad deal, and a couple years after that it’ll be a straight-up strategic liability. THAT is why anything other than full-throated support from Google leadership for indefinite GCP investment will lose them business.
And maybe they won’t kill it, but I ve had a number of times platform changes killed features I used. Thing is, with a company I pay for I can vote with my wallet. With Google it’s just sending a message and hoping it doesn’t end up in nullspace somewhere.
Google is just too big as a company, and if they replaced every human in it with drones no one would know the difference. There is no human face or point of contact or any accountability and it’s becoming a problem.
My girlfriend was able to get on the phone with a tech support person for her small business ads and I was shocked she was able to even talk to someone at Google over a phone without being a major company with a big monthly bill.
The woman was still snarky to her and unhelpful but still a real human was available.
A lot of this has to do with our own perceptions and trust. Purchasing decisions are never fully rational or information complete.
> that doesn’t mean that GCP has any guaranteed future either.
Nothing has a guaranteed future.
> AWS and Azure won’t slow down development
As long as it's paying off, they won't, and AWS maybe for longer even if it isn't. But Microsoft is no stranger to making investments in offerings and letting them stagnate or outright killing them if they don't pay off.
While Google Reader might be the most popular of the tools, that got offline'd, doing so is a common pattern with Google. According to gcemetery.co[1] Google has already killed 164 products, with each one having an average lifespan of approx. 4 years.
Google is always building new data centers to handle traffic growth for their own services. GCP infrastructure wouldn't be turned off, but re-purposed for internal growth.
On the other hand the, in comparison tiny, cost that Google Reader and other was in the budget, would have given Google so much good will. It was in other words cheap positive PR.
Killing Reader was massive strategic blunder, which in fairness to Google was probably not obvious to see at the time. But then to repeat that with Maps API,Google+, Sketchup and Wave (perhaps wave was first) just cemented Google's untrustworthy reputation.
Sketchup?? Huh. Glad I didn't get into it, then. This is the first time I've heard that, and I thought I was up to speed with the stuff Google dropped.
It doesn't matter, a datacenter can always be repurposed for their own internal product lines.
To be perfectly honest, i would rather go with a company that provides cloud as one of their main products rather than a company that has it as one of their many products.
Google piled massive resources behind Google+ and look where that is.
Personally, I don't think Google will yank the cord on this entirely, but they may just stop piling resources into it and/ or significantly increase prices to ratchet up profits. They don't have to drop the whole cloud to really screw up a developers life, just screwing with one API I rely on can kill a product and they've already done exactly that.
The problem is with Google you just don't know what's going to happen. Microsoft and Amazon are both known and neither plays games with their products.
Yeah, GCP is not the same as Google Reader, more like Google Maps. But that one recently had some changes that royally fucked some businesses over by changing pricing and quotas, so maybe not so absurd after all.
Exactly. If they just don't improve their pricing in line with AWS/Azure (which means massive capital investment, always subject to economy shifts in the cost of money) then people will be looking to move.
All the cloud offerings are massively more expensive than owning/leasing unless you have large dynamic scaling requirements. There will be middleware in the future to allow cloud agile / hybrid operation, so maybe cloud compute will become more commodity, as opposed to the current Rolls Royce pricing. At that point will Google be interested? Depends how much marginal cost is involved compared to running data centers just for themselves.
Investments in data centers and infrastructure can mean that
the Google Cloud may live on, but different offerings in the cloud might change with little notice.
Even if they don't just "turn off", why would I want to invest in a cloud platform that only staff enough to keep the lights on, while there are cloud platforms out there that continue to provide new features to make my development process less painful?
Those datacenters don't only house GCP you know. I'd be willing to bet GCP only makes up a comparatively small portion of their datacenter computing resources.
>Google’s problem remains its reputation for turning things off.
I'd put it this way: Google has a reputation for turning things off, but its _problem_ is an inability to stop doing that, to commit to its users/customers. In other contexts this could be written in terms of addiction.
If I were Sergey and Larry, this would be the thing I'd be retiring over, and I'm not sure Suchar can change it.
If Satya Nadella could get Microsoft to move beyond its Windows-all-or-nothing pathology, someone in charge at Google (including the two founders that vote-control the entire company directly - they can de facto hand Sundar the power to act as he needs to) can trivially put a stop to the relationship commitment issues at Google.
Its far from trivial to change an entire corporate culture on a dime.
From everything we have seen of Google's culture, working on new projects is exciting and is absolutely the way to get promoted.
Doing maintenance work, especially on products that aren't wildly successful gets you no attention, no staff, no funding, and probably more than a little tongue waging.
The CEO cant just change all that with directives, even if they do something extreme like a "No Depreciations" policy, products are liable to be depreciated in all but name (Joe the intern takes care of it every 5th Tuesday, technically not depreciated) by sheer corporate momentum alone, nobody wants to be stuck with a dead fish, even if its now illegal to openly comment on how much it stinks.
It would require change at every level of Google to place value into fixing/improving existing projects, rather than all the glory going to greenfield development.
Considering the kind of "kingmaker" culture Google's mentality breeds, ie what gets me ahead is gathering resources to focus on my project, changing things to a setup where contributing to other peoples silos is welcome and common is going to be an extremely difficult prospect because virtually all the people with power inside Google got that way by carefully tending their fiefdom, they aren't just going to throw that all a way, no matter what any executive says.
Its far from trivial to change an entire corporate culture on a dime.
You can say that, but when it's a response to the fact that Microsoft did that EXACT THING on the the scale in which they did it your statement doesn't hold any water. Go back in time and walk onto the Microsoft campus in Seattle and tell anyone that will listen in say 1995, 1999, 2001, 2007, or 2012 that Windows licensing is a dead end and see if you don't get lumped in with the loonies talking about the end of the world.
To say that Google can't do it is more an indictment of Google's leadership as opposed to the enormity of the task. To thinking a company with a 934 billion dollar market cap (Alphabet) can't figure out how to do enterprise software support if it wanted to, what a joke.
“All teams will henceforth expose their data and functionality through service interfaces.
Teams must communicate with each other through these interfaces.
There will be no other form of interprocess communication allowed: no direct linking, no direct reads of another team's data store, no shared-memory model, no back-doors whatsoever. The only communication allowed is via service interface calls over the network.
It doesn't matter what technology they use. HTTP, Corba, Pubsub, custom protocols -- doesn't matter. Bezos doesn't care.
All service interfaces, without exception, must be designed from the ground up to be externalizable. That is to say, the team must plan and design to be able to expose the interface to developers in the outside world. No exceptions.
I am not sure why you are saying Microsoft transformed, its still the same old one now trying to embrace Linux as it runs as the main workload on Azure instead of windows. This is just done for survival, otherwise Microsoft would have gone to irrelevance.
Windows and Office licenses are still the main driver for Microsoft, just with accounting tricks and office 360 made into cloud revenues. This is an investor story to jack up the stocks. They are no better than google when it comes to platform lock-in. On top of it having used Azure services, I can say without a doubt its inferior to Amazon and Google cloud experience.
Microsoft didn't transform it's just that enterprises who hold up upgrading Windows and Office did so after windows 7 debacle to Windows 10. This is just cyclical uptick, not really transformation, its just lucky Nadella. The only good thing Nadella did is, didn't take any disastrous decision to thwart the uptick by embracing Linux and trying open source, by understanding that if it's not done Microsoft will go the IBM path to irrelevance.
In the process Microsoft having an old relationship with enterprises helped them to move some workloads to Azure even though it is an inferior service, by supporting Linux and showing commitment to it for financial gain not out of goodwill towards open source.
Still today Microsoft relevance on mobile is due to its office products, otherwise it is as good as irrelevant on mobile application and development front. It's alive on mobile just due to proprietary platform lock-in to .NET. Now made it open source to try to increase the share as it's not as useful given the rise of nodejs, Python, rust and other platforms.
Windows has not seen a payed upgrade for 5 years - that is in itself a major change. We'll see if they continue on this road, but for now their announced plan is that there will not be a new paid update of Windows. This clearly cuts down on one of the revenue streams for Windows Licensing.
Microsoft is the largest single contributor to Linux development. They have a large hand in Kubernetes. .NET is open source and runs on Linux/Mac natively. They have released and are maintaining the most popular lite IDE/editor of the decade simultaneously for Linux/Mac/Windows, and it was open source from the start. MSVC has a Clang frontend.
That they are not relevant on mobile seems a non-sequitur. The discussion was on whether the Microsoft Windows-first culture has changed, and it clearly has. They have clearly embraced open-source as a good strategy for them, and are making massive investments in it.
I would also note that Azure, while having all sorts of unexpected limitations, is at least much more transparent about pricing than AWS. The Azure UI almost always includes pricing information on the resources you are creating, which AWS carefully segregates away.
This licensing does not apply to enterprise editions and volume licensing, I believe you are still looking at consumer market which Microsoft already lost to a Apple and lately google chromebooks.
The major revenue stream for Microsoft is still old long term licensing deal for office combined with windows and than related server technologies to support office applications.
The model changed from licensing+upgrade fee to new subscription fees. This can be office 360 subscriptions which get you office products and windows or volume license subscriptions. Depending on the type of subscriptions these are classified as cloud or windows and office revenue. Azure come into picture as those companies having volume licenses have applications in legacy .NET which they move to Azure.
Microsoft is supporting Linux today because they lost the battle of technical superiority and most of the azure workloads for supporting internal business apps run on Linux. They ported .NET to Linux just to have those old accounts do not move out of azure and Microsoft proprietary eco-system easily and move them to subscription with long term contracts.
Majority of the revenue effect you see is just because Microsoft abandoned product updates for old versions and forced customers to move to windows 10 and above by making old one economically very expensive. So Nadella was lucky because he got the benefit of the cyclical revenue increase due to enterprise upgrades which companies didn’t do for many years.
Obviously adoption of Linux as primary driver for Azure kept those old customers from moving out of Azure to Amazon or google, even though technically azure quality and services were inferior. A company will not move to amazon or google unless there is significant exponential gains. If the old vendor can still support and give something which is not bad but not the best, customer will still stay on and here Nadella was good not to bash Linux like Balmer.
Companies prefer to purchase services within an established contract than to go ahead and sign a new one with new provider and go through the bidding process.
Also .Net is synonymous with microsoft so there is no reason to trust AWS, it's not their priority, they are a Perl and Java shop in general.
AWS is always bragging that they have twice as many Windows VMs than Azure.
AWS a Perl shop?
All of AWS services treat C# .Net Core as a first class citizen.
- CodeBuild has prebuilt C# .Net Core images for building apps. For licensing reasons they don’t have prebuilt Windows/Framework containers but their instructions are so clear that I was able to build one before I knew anything about Docker
- Lambda has native support for .Net Core 2.x and the C# SDK for AWS has a template to build a custom runtime for C# 3.x
- Speaking of which, the AWS SDK and plug-in for C#/Visual Studio and the corresponding Nuget packages are excellent and fully featured.
- The CloudFormation Developer Kit supports C#
- There are CodeStar and ElasticBeanstalk getting started templates for .Net Framework and .Net Core
AWS was built to support Amazon website which is built in Perl and still internal AWS core code might contain perl. Probably over the years they might have moved slowly to Java. I don't work for AWS so can't say much.
Regarding support for .NET whatever AWS does microsoft will always be way ahead of them. Microsoft is not elastic.co that AWS leech on and put the original guys on the side and don't contribute back. AWS is really like a leech on open source now trying to leech on Kubernetes.
They can do this with small open source project, not with the projects supported by google or Microsoft. Like when it comes to Kubernetes AWS is playing catch with Google. So AWS Fargate and other services are still inferior to Google managed K8s. Indeed microsoft at present is ahead of AWS in managed kubernetes services. Its just the initial head-start and developers who are familiar with AWS proprietary API's find it hard to believe that in such a short time the skills they learned is no longer cutting edge and google or microsoft proprietary API's overtaken them in performance and ease of use.
AWS was never built to support amazons website, it was never the result of amazon.com's internal code-base of "hosting" getting made public, it was never amazon's excess peak server capacity, etc, it had virtually no relation to the hosting of amazon.com at all, it was an independent project that might have benefited from a lot of the general design principles of how amazon.com was built, but that is it.
All of these things are widely believed myths that have little to no basis in reality.
It certainly was not designed as some kind of Perl hosting environment.
It didn't even host Amazon.com at all until many years after the fact.
I did use AWS for good 5 years and every time it's been inferior, e.g. launching times of EC2 instances were double that of Google Cloud, better than azure though (might have changed with new instance type may be), N2 and N1 instances in gcp are usually faster in start, stop delete cycle than AWS as recently as last week. Networking stack between regional nodes is not as performant with higher latency than google cloud with its andromeda network stack, on top its more expensive than google and microsoft for bandwidth. Even though I put critical comments for GCP on this article, I feel AWS fare no better, indeed I feel its worse.
I have yet to see any good open source project by AWS team in the wild like Kubernetes, Linux Kernel Namespaces and Cgroups, Map Reduce from Google or .NET from Microsoft. So far what I have seen is that AWS just leech on other open source like PostgreSQL to build Aurora and RDS service without really contributing back to PostgreSQL, because postgreSQL comes with BSD license AWS can just leech on it. Same for elastic search and many other open source they leech on. I am still waiting if they are really just leechers or giving back to the open source community. Like Amazon Redshift is based on an older version of PostgreSQL 8.0.2, and Redshift has made changes to that version [1], but nothing is made open source because BSD license do not require it. This is just one big example but there are many. For open source community and contributions I feel Google and Microsoft are much better than Amazon and AWS by miles. It's just that due to early start AWS is reaping benefits by leeching on open source.
It seems like you were much more on the infrastructure side than the development side. The discussion was about .Net. What was your experience with the SDKs to judge where AWS was behind MS supporting it?
How does open source help an end user who is deciding which cloud provider to use?
AWS is infrastructure as a service with some priorietray API's to provide platform to build applications. These platforms are build from open source projects which does not have GPLv3 or AGPL license. So AWS is largely an IaaS service not really a platform, open source library or system development company like Google or Microsoft.
AWS does not develop .NET SDK, as I mentioned in my earlier post there is no contribution of AWS in development of .NET, so they just provide infrastructure support of what ever Microsoft and the open source community releases.
In general microsoft acts as a steward to develop .NET and related infrastructure on Azure, then AWS copy and launch it as .NET infrastructure as service. They play catch up with Microsoft in this area.
I will stand corrected if you can show me that AWS (i.e. Amazon) has active open source .NET framework core developers in their internal team developing and enhancing .NET SDK or CLR open source versions.
They do the same with AI and ML Services. It is built using TensorFlow, sci-kit learn, pytorch, CNTK and various python and C++ open source libraries. Again I have yet to see example of AWS (i.e. Amazon hiring core developers of these libraries and release open source versions). They don't develop these libraries just leech on it to create it's own API's. I have yet to see any framework from AWS like TensorFlow, Keras, Pytorch and CNTK.
Again, the discussion was about AWS being behind Azure with respect to support for .Net. When deciding which cloud provider one should go with, why should I care as an end user which company did more for open source?
There is nothing to copy as far as .Net. Microsoft releases a new version of .Net, all AWS has to do is make sure that its SDK’s support it. But there usually aren’t any breaking changes.
As far as the other services that I could see needing to support, the only ones that I can think of is CodeBuild and lambda. For those two, you can create custom Docker images and runtimes respectively.
> I am not sure why you are saying Microsoft transformed, its still the same old one now trying to embrace Linux
I'm one of those persons who upsold friends, families and at least one boss on to use Linux.
I'm not a business analyst but I can say that Microsoft has transformed. I still prefer Linux as does a number of fellow sysadmins and developers, but Microsoft has changed to the point where they are not a problem to us anymore. Azure portal works equally well from Firefox on Linux as from Internet Explorer or Edge on Windows.
> as it runs as the main workload on Azure instead of windows. This is just done for survival, otherwise Microsoft would have gone to irrelevance.
The best reason for transformation in fact: because the subject realizes it is in their own best interest to change and it improves their situation.
> Windows and Office licenses are still the main driver for Microsoft, just with accounting tricks and office 360 made into cloud revenues.
I don't know.
> This is an investor story to jack up the stocks. They are no better than google when it comes to platform lock-in. On top of it having used Azure services, I can say without a doubt its inferior to Amazon and Google cloud experience.
I've used Amazon (two years, narrow experience) and Google Cloud (two years, broader but possibly shallower experience, certified) and now Azure lately.
So far, inferior is not the word I would use. I'd possibly stick with different or something neutral.
One great book on how the culture of an entire giant company can change is "Who says elephants can't dance" by Lou Gerstner. He was the CEO of IBM who orchestrated IBM's transformation from a mainframe supplier to a service-oriented company. My take on the process is this - the change can be done and has been done when there is an existential crisis. The crisis may not be apparent from the outside, but company internals can notice a terminal decline of their core product/service and has an impetus to change the direction. Even then, it's not easy. I would argue that Microsoft had to make somewhat of an about-turn from Windows because first the Internet and second the Cloud was putting pressure on their core business threatening to make the OS a marginal player. I don't think Google is in that position, at least not yet. Google Ad business is not facing any terminal threat. Thereby it is much harder for Google to make a culture change because simply the pressure to make the change is not there at an organizational level.
[Edit: Prepositional corrections]
They're already there on my part. Between their firing employees who organize or complain that their corporate slogan apparently now is "Be Evil", their chaotic product decision making, and their toxic internal politics, I have absolutely no interest in working for Google. Furthermore, I've noticed that the Google workforce seems to be becoming less.... Googly.... recently (disclaimer: I work down the street from the Googleplex and see many Googlers as they do their daily migrations). As in, they're now hiring a broader selection of people who aren't just the top 5% in class. What that tells me is that they're having more difficulty on the recruiting front and having to cast their net wider.
When I first wrote that sentence, I said "until people", but it occurred to me that being shunned by a small group of engineers probably barely registers on any metric that a large company would pay attention to. It wouldn't be actionable until the well had been running dry for a while.
I mean, shit, I work at a place where many of my favorite people have moved on to other things, few have been replaced, and yet we're only just having sincere conversations about how to fix our culture problems. That had to wait for one of the ringleaders to quit.
So far as I can tell from my minuscule keyhole, Sergey and Larry are the ringleaders. They'll never quit, and it'll take hitting rock bottom for them to have a change of heart. Not unlike Microsoft, after their stock began to tank.
Every major tech company has changed its culture based on a changing landscape.
Apple - look at Apple’s revenue and profit mix over the past 10-15 years its apparent that they were able to change. Of course the change after Jobs came on is well known.
Microsoft - was “Windows first” during the Balmer era.
Facebook - was in danger of being made irrelevant after mobile. They were making a lot on games hosted on their platform.
Amazon: AWS and their move toward services even internally.
Netflix: went from shipping DVDs and had to become a tech company in the era of streaming.
Your comment really resonates with me. I worked at google for many years and I eventually got tired of exactly that. Inside the company we complained about all the problems they still have, canceling projects, no way to talk to a human not even paid support to talk about your whatever being turned off. It was really irritating.
I think Google needs to explicitly spin out a Google Product company that is customer focused and develops/maintains/supports platforms and applications. It would be distinct from Search and Ads and would instead sell them (and other customers) access to Google data centers and GCP.
I think Google needs to explicitly spin out a Google Product company that is customer focused and develops/maintains/supports platforms and applications.
Honestly, and I'm more of a Google side-eyer than a hater, but I'm not sure they're capable of this, it's just not in their DNA. I've been saying "Google is bad at product" for a long time, like a "high PR" long time, and this mindset has been greeted with continual killoffs and usability stumbles (current Android is kind of a mess UI-wise, e.g.).
I simply don't think they're capable of putting a division together that could be relied upon. Heck, Alphabet might simply kill that off after a few years!
Another perception is that of killing hierarchies of user accounts / pulling rugs on projects e.g. if links are established between adding someone who once broke the rules and whatever they might be working on now, with no way of reaching someone human for support...
... where one of the few ways of rectifying the situation is to get a top posting on HN in the hope that a Googler will see your plea for help.
I mean, that's probably mainly related to Android dev but I've seen plenty of stories to support that feeling:
My perception might be misguided / influenced by being in a perception bubble, however that is my reality and AWS or AZURE just seems a whole lot safer. So, why should I as a developer invest any time in GC? I did attend a Google dev day re machine learning++ in Melbourne a couple of years ago, and the products seem nice and all, but the competition isn't exactly lagging behind.
My recommendation would be to really address this issue of human support and trust.
Also AWS is just the best at marketing their services. They did what Google did with search. We always say we "Google" things.
Developers these days say "Chuck it in S3" or "Run it on EC2". I feel they have a strong advantage because they've kind of come to own the language of the cloud.
Developers these days say "Chuck it in S3" or "Run it on EC2". I feel they have a strong advantage because they've kind of come to own the language of the cloud.
After many years of working with AWS almost exclusively, I recently started a new $DAYJOB where our primary public cloud provider is Azure. A couple of days ago I was chatting with a co-worker about something, and he literally says "I think we should just chuck it in S3 and call it good" (paraphrased slightly, due to my imperfect memory).
It struck me as odd for a few seconds, and then I realized that exactly what you are talking about has taken place. S3 is so dominant that it's become the default label for object storage in the cloud.
Actually, most of the time we say "in the cloud" rather than naming a specific hosting provider. Definitely Amazon has the first mover advantage and accumulated reputation first, but it's hardly unassailable.
Still... I think the bigger problem isn't whether or not Google plans topull the plug or not, but that it has a reputation of pulling plugs to such an extent that people factor that into a decision of using their services.
When it's something as big a choice as a cloud provider that can cast a long shadow over their offerings.
In addition, I remain cloud agnostic as well. As a student, I stick to whichever platforms can give some free credits. At the moment I have some cloud infra on google and some on aws.
That’s a good thing not to tie in to proprietary api’s, indeed every cloud provider is doing it and after Google’s recent management shuffle to bring in management from Oracle making them alike.
Try using apache-libcloud to make your deployment and programs cloud agnostic. Also try using smaller alternatives like hetzner or scaleway.
Once you go down this cloud rabbit hole you will be hold hostage to whims of these providers, where they can shutdown any service at their will or ask to pay an arm and leg. Also try to have non USA cloud providers as another backup option in the long run.
Amazon and Microsoft are much better at support though. I doubt you’ll see a lot of Google cloud in the European public sector where I work for instance. Google simply doesn’t know how to work with us or give us the reassurances we need to trust them.
I’m by no means a AWA or Azure fanboy, I’m on the team that hopes we get a solid European alternative, but until we do both Amazon and Microsoft are leagues ahead of Google.
I remember vividly talking to a GCP saleswoman. I was asking her about technical things and process/pricing and she just kept saying things like "We're talking with $FIRSTNAME about that." And I was like, 'who is this $FIRSTNAME guy?' and then I realized it was the CTO. This guy is so far divorced from anything technical or policy it's absurd, he just produces content for his 1- and 2- up bosses to look good in press releases, and he has no input to the way anything is done. But GCP had identified him as the way into the organization, and was totally uninterested in engaging with anyone else other than giving away swag.
Contrast this with AWS. They have expert engineers, pre- and post- sales, helping you find good ways to solve your problem. I could ring a senior guy on his cell right now if needed. The only problem with AWS is cost and complexity, it's the fast path to developing a $20k/day cocaine habit. But the AWS dealer will definitely come to your house to deliver at 11pm.
(It is a bit of a change from 10 years ago though, when everything had to be paid by credit card. Gave the accountants a heart attack. Gave me an ulcer from thinking about everything halting because the fraud control might reject the bill without warning...)
Azure definitely isn't in that league, still want to sell to the whole enterprise, getting promo stuff requires them to sign off at five levels, lawyers are involved in everything. In short, their dev relations are shite.
To illustrate your point about AWS, at my own workplace there are a few AWS people who are actually in our office once a week, sitting nearby just so they can be easily available to answer questions and work with us to design our architecture or assist in troubleshooting issues. I honestly can't imagine Google providing that level of dedicated support.
Funny thing about plus, reader and wave: that was the kind of stuff that people who are now in positions to make infrastructure decisions got worked up about. It's the sort of image damage that may have seemed minor and acceptable at the time that can come back with a very large multiplier in the future, especially given the kind of prominence those products once had.
You're correct! I've never thought about that before but when they killed the RSS reader it really left a bad taste in my mouth. Now that I _do_ have the ability to make decisions at the company now, the bad taste did transfer over.
This was genuinely not a factor when we signed a 1E8 deal with GCP. What was a factor: (non-exhaustive)
* Access to their developers for core services
* Discounts available
* Quality of support
* Access Controls / Audit / Budgeting
* Dev Tools around Infra
* Future Roadmap
Not one of us ever would have chosen the shutdown of Plus/Reader/Wave as a factor. As a startup we hard pivoted in order to succeed. No surprise that members of the population of products have to die for the population itself to survive.
And the group making these decisions are a bunch of Unixheads who were some of the first members on The Old Reader. We made a few hundred million bet that GCP will be around and capable for the next five years (and probably for the five after). I am confident this will play out in our favour. I'd put money down except for the fact that I already did effectively by recommending the deal.
I agree that this seems to be a narrative that exists only inside the HN bubble. The fact that a multi million dollar business decision is being made not by running numbers and future roadmap but by one guy in the meeting saying "wait! they shuttered Reader, my favorite RSS reader. Oh no, these other guys killed Zune, my favorite music player. Wait a minute, don't tell me these are the same people who killed Firephone. Ok, we're going with Fujitsu cloud guys" and everyone nodding their head in agreement
True. But not everyone works at a multi million dollar business. Many start using cloud platforms in small companies and side/school projects and build up knowledge from there which eventually affects bigger companies: "Hey! Should we use the platform most of our people are already familiar with, or the one Bob used for his 500 line hackathon app last weekend?"
I'm not actually worried much about the multi million dollar companies using Google Products because then they start actually caring about you. When I worked at a place that made them significant ad money, I got splendid sub hour support. But Google is among the most disrespectful companies towards smaller actors while still trying to get as much money out of them as possible I have ever interacted with. E.g. Just recently it took us 3 weeks, 5 chat support sessions and 2 interventions from their supposed expert team to get an inappropriate "customer photo" shot 4'000 km away from maps/search removed. Or the whole Youtube ContentID disaster, or getting forced into G+ just for them to close it, etc etc.
Don't think anyone really cares about 1-2 products being closed that they were using, it's just a lot a things that add up and changed the perception of the company quite a bit over the years.
you're missing the point. Google has a bad habit of shelving services that people use. If it's never happened to anything you use, good for you, but if I was in a position to make a choice like that I'd be very hesitant to go with a company with such a history of shutting down services.
If people can't tell the difference between free consumer products and paid enterprise services (appengine has been running since 2008) then they have bigger problems when it comes to decision making.
Google Maps is a paid enterprise service, and while they didn't discontinue it, they did increase the price by 14x.
Of course, personally I chose AWS over Google Cloud well before that, because Google Cloud's authentication is so complicated and their examples so indecipherable.
Reputation does matter but depends on circumstances. We are discussing using a certain in-development chip from a giant company with a reputation of completely and utterly failed delivery of a chip in a related area (and lying to public). The first question I've asked - if they failed that thing just recently, how do we prepare for them possibly failing our new project? This is a serious problem to consider.
The thing is that a long term roadmap for a startup is 3 years. So you're going to look for things that are going to get you 3 years down the road cheaply, and then you still aren't that big, you have gigabytes of data, not hundreds of terabytes of data, so pivoting to another cloud provider isn't a big deal. Now, scale up to a company that has a monthly IT spend of 50 million dollars. You aren't moving that much infrastructure between cloud providers on a dime. Past a certain scale, there is a huge incentive to standardize on a vendor that has a 10 year time scale. We know AWS will be there in 10 years. Will GCP be there in 10 years?
I suspect the 50 mil / month guys will choose between AWS/Azure/GCP based on credits + discounts because all these clouds will be around 10 years from now. But it's true that I have no experience at that level (only from 0 to 9 mil / month) and the calculus could be completely different. I don't actually know if they could get greater commitments at that spend but it sounds likely. That's more than half a billion a year. Huge.
Perhaps someone who's made the decision on that can share.
Googles reputation was pretty much the decisive factor when a couple of years ago we decided to move from Heroku to AWS instead of GCP. Besides a prettier web interface GCP didn't really had anything going for them in comparison for our use cases, so why take the chance of having a rug pulled even if that risk might be quite small.
Absolutely. I like GCP more than AWS or Azure, but I don't use it because I don't trust that individual services will be available long term. Meanwhile, AWS sends me emails about minor service changes that may affect me well over a year in advance. (MS has a reputation for long term support, but the Azure UI was just so terrible. Maybe they've changed it by now?)
I used it about a year or two ago and remembered the UI as horrible. I then had to use it a few weeks ago and didn't have the same negative reaction I had the first time I used it.
Having said that, AWS also have a long track record of reversing charges where a dev fucked up and ran up a huge bill accidentally.
I am _extremely_ dubious you'd ever even get a person on the phone to talk to at Google if your CI/CD pipeline or platform automation had a bug that resulted in thousands of dollars of charges racking up over a weekend. I know people who've done that on AWS and had the entire bill waived, and then had AWS technical people call and help them debug and fix the broken (non Amazon) systems.
I'm not sure there's a company on the planet with a worse reputation amongst devs for "customer service" than Google. Pretty sure they're not going to be winning any developer mindshare on _that_.
If your 5-6 figure ad buying team get a say in which platform you deploy on, maybe _their_ experience with Google's legendary "customer support" is different. Buy the only known avenue for developers to get problems with Google accounts looked into by a human being is pretty much to make as much noise on Twitter/HackerNews as you can and hope some random Googler steps in to help...
Ah yes the "try getting someone from Google on the phone" trope.
I've used GCP at 2 different places. One with a 10k/month spend and another with closer to 200k/month and received the same level of service. In both instances I've had direct access to engineers to get shit fixed when needed, phone/email/gchat.
GCP has a free billing chat support open to every customer, big or small. It's open 24/7, they can reverse charges (although obviously you'll need a pretty solid justification for 4 figures), and they'll happily jump on a call if you ask them to.
I have no experience with Google support, let alone GCP support. But what I've heard both from inside and from customers, is that unless someone happen to want to help you with your specific problem, it can be a bit of hit or miss situation to get help. Maybe it's different with now GCP, but until I hear first hand experience telling me otherwise, I'm not even considering moving our cloud business to GCP.
We're running production systems making a significant amount of of our revenue, and when, not if, shit hits the fan we need someone with intimate knowledge of the internals to give us undivided attention until it's resolved.
Sometimes it's because we fucked up, sometimes it's the perfect storm of internals working unexpected - all of the time our systems are not working and our customers depend on them for their daily work.
Google letting the Google Maps API rot on the vine even as they raise their rates for developers and exactly zero support for effected users is a perfect example of how they are pulling the rug out.
Also, as has been mentioned, if your option is tolerating seriously crappy support or having a service pulled entirely, most people would take the crappy support.
It took us about half a year to get billing set up with them. Completely pointless shit show that basically every other company on the planet manages, with Google's own unique weirdness about talking with icky customers.
It was an unpleasant process, and then we dropped them anyway (but not because they suck at billing).
Poor customer service is not exclusive to Amazon, google directly shut the product down which cannot earn search scale revenue. If you see comments below being downvoted by google PR for examples.
I believe all these large cloud companies be it amazon, google or Microsoft deploy army of PR people and campaigns to downvote or try to remove comments which are critical. Especially I have seen this on HN, earlier when I commented on Amazon. Now can see here when it’s critical of google.
This is the same logic my extended family uses when they tell me that their phones eavesdrop on their conversations to target advertisements. Just because it's possible and fits your narrative doesn't make it true.
No brand or PR team that I've ever been near has the capacity to monitor and then manually downvote this kind of thing, even on major sites, let alone a niche news site like HN. Who has the budget for this?
> No brand or PR team that I've ever been near has the capacity to monitor and then manually downvote this kind of thing
Astroturfing and reputation management are a thing. Also, big companies have a large number of employees some of who hang out here and will simply take offense by association, which could be more than enough to explain the effect even if the company itself has nothing to do with it.
HN is no longer niche.
> Who has the budget for this?
Apple, Google, Microsoft, Amazon, Facebook, Twitter and whole pile more if they wanted to it would be a round off error on their budget but besides one of those I don't think there was ever hard evidence that this happens.
Reputation management is becoming more and more important to brands.
Pretty sure they wouldn't even notice the costs of downvoting every single negative comment about them on every single website they're already crawling and mining for keywords and sentiment...
It's possible, but the reputation management people also worry a fair amount about getting caught astroturfing or running bot nets. There are a lot less risky ways of accomplishing the same goal- pushing positive stories, honestly responding, using SEP techniques to get ahead of negative news in search results, etc.
What, you mean I could paid for reading and posting HN? Where do I sign up?
The claim of a covert army of astrotufers / reputation management firms was silly when it was made back in the day on Slashdot and it's still silly today. If nothing else, it would never pay well enough for people to keep such a juicy secret.
I keep seeing you blame any downvote you get on external factors. To me it looks like you're just trying to find excuses and cannot accept that people may be downvoting you for legitimate reasons.
Funny enough I just got an email from YoutubeTV last week basically saying the following:
Effective immediately and without warning, YoutubeTV will not have a dedicated channel for Seattle Sounders games (the whole reason I got YoutubeTV was because they were advertising they would do that) and all the previous Sounders games you’ve recorded are now gone effectively immediately (even though, once again, they sold me on their “infinite DVR capacity” that lasts for two years).
Google not just terrible here, it’s just void in terms of customer service. We had an outage due to GCP glitch , there were no way to report even. Recently I heard about bare metal offering , tried to talk most of sales i talked wasn’t aware of GKE term. I wasn’t able to get pricing eventually after 2-3 calls . It’s probably top secret
Eh, I have to point out that AWS has many services that receive such scant updates that they are effectively dropped products.
There are a lot of things that AWS produces which never receive adequate polish or timely updates. I’ve been in betas for more than a few new offerings and was stunned at how incomplete and buggy these offerings were. More than a year later, they remain in a half usable state.
Google dumps things it’s not committed to making great. AWS maintains numerous things which are half usable, poorly documented, and incomplete. It all depends on the context, but in some ways I prefer drops to half-assed products.
True. But if I build a tool that depends on a half-arsed service my tools keeps on tooling. Its about predictability. I'll soon find out that I can get a better service elsewhere, in the meantime my investment is not a complete write off.
> he problem Google has with the developer community is largely around the fact that Google has a long history of pulling the rug out from our feet.
This is merely a symptom of the real problem Google has: a contempt for humans that aren't Googlers. Everything about Google - as a company and their products - speaks of an institutional contempt for the normies outside their walls: everything from the interview process through the complete inability to respond to criticism with anything other than a "you're holding it wrong".
Google's high-handed behaviour around "oh your account got locked, fuck you", or "we shuttered this thing that people depended on, fuck you" is a product of that contempt. You're a Googler or you're shit.
It's absurd to compare loss-making consumer products with a business that's the only shot for Google diversifying beyond ads and that has companies paying hundreds of millions of dollars over decade+ contracts. Urs says enterprises don't care about this aspect of the Google brand. What they do care about is Google's reputation for technical excellence and security.
It's also interesting that Google is willing to suffer an unprofitable product for many many years (YouTube) without being fickle and cutting it, yet other products aren't even given a fair shot before getting the axe.
Other than profits, YouTube has many problems that no other product has - like dealing with bad content, detecting and enforcing copyright infringements, comment spam, storage costs. After going through all these headaches, the profits seem even smaller.
I agree your sentiments here. The goal for Alphabet is to make Cloud platform business profitable not shutting it down. From this article it looks like it hasn't been so since 2014 after GOOG gave it full attention. And to be profitable, Alphabet execs thought it has to be within top 2 to be profitable, I am unsure if that's true, given the cloud business is still a slow and painful transition process for Fortune 1000 businesses(some more advanced than the others).
Been using AWS for 8+ years since their first inception of S3, then 6 years of GCP. This year I had first hand experience with Azure. I just happen to think GCP has the most cutting edge cloud native technologies and services (GKE is much better than EKS or AKS, IMO; BigQuery is awesome). But AWS has a first mover's advantage and MSFT has the dominant enterprise customer base so pushing additional add-on sales is much easier. Those are GCP are up against.
Perhaps the market is big enough to accommodate top 3 players to be profitable. And GCP needs to do a better job in support their product & services not discontinuing them abruptly.
Google has cool things in GCP, but I struggle to imagine them being _popular_, cool, things.
If your product or project is big enough to require the scaling and capacity that GCP offers then you’re also probably big enough to be risk-averse about predicating your product’s viability on a platform you don’t control. The way I see AWS and Azure’s success: they replaced server colocation, dedicated server rentals and shared hosting for application hosting because they eliminated the costs and burden of systems administration and hardware maintenance. If you’re a company that did move to AWS/Azure - provided you stuck to their commodity offerings (App Service, EC2, RDS, Azure SQL, etc) you can move back to your own hardware just fine. For other things like S3/Azure Storage which are hard to move in-house, at least it’s easy to separate that from the rest of your application.
But what does GCP have that is cool - but without associated costs of what is essentially vendor lock-in? Without a solid contract/guarantee from Google that some binary I wrote that uses a GCP-exclusive service _will_ be supported for a minimum of 2-5 years it’s hard for me to justify risking it.
Google is not dumb and realizes this, which is why they're pushing so many tools that prevent vendor lock-in. K8s entire pitch is you can seamlessly go from a cloud offering to your locally managed K8s in a datacenter. Our (small) company has done it, it works very smoothly.
Their 'serverless' batch/stream processing works this way as well, with Apache Beam you can deploy effortlessly to GCP for large compute jobs (i've spun up hundreds of nodes in minutes), and also run it against on prem spark clusters with zero code changes.
So Google realizes being late to the game means they have to offer something more, and what they're offering is a lack of vendor lock in.
Yes, it could have been Google's competitor making up these stories so enterprise customers would stop and think before switching, it is also better for competitors's sales to pitch on the potential of a shuttered GCP, which you can tell full well if I was a sales I would be referencing this article.
And it wouldn't matter of there wasn't even an ounce of truth there, Google's reputation on future roadmap, communication, marketing and product dropping has always been on the negative.
Google needs to work hard to earn customer's trust. That is the basic fundamental of doing business. Microsoft is still working their asses off because of what they have done during IE era.
> Google needs to work hard to earn customer's trust. That is the basic fundamental of doing business. Microsoft is still working their asses off because of what they have done during IE era.
This. A thousand times this!
Once hero of the 2000's, Google is turning into the villain; meanwhile Microsoft has embarked on their redemption storyline.
Microsoft is a fantastic developer-centric company that has made their platform attractive to even Linux users. Google shuts down all the things you use, prices small shops out of freaking Google Maps, won't pick up the phone, ...
Google should hire Satya Nadella.
Doesn't Google understand that we should love them? That we should think they're a cool company? They've strayed so far from that. They might as well be Oracle.
Google cloud head Kurian [1] is a former Oracle executive, how can you expect him to work differently when he has spend decades in oracle.
I believe google cloud platform might be slowly turning to be similar to oracle, if things do not change to care more for users. As many people noticed google like oracle started changing their products in such a way as to have lock-in like oracle.
So it’s just a matter of time when they start behaving like the old Oracle to use the lock-in to ask for price hikes.
Additionally, he didn't have a particularly developer friendly reputation at Oracle -- and that's something they'll need within Google to attract the people who will continue to build the pipeline of features they need to attract customers.
Kinda make me consider the comparison with the changes to Map API pricing earlier this year. Exactly the sort of sales weasel and lawyer driven revenue booster you'd expect to have come directly out of Ellison's office...
You've totally drunk the Kool-Aid. Microsoft is on a redemption storyline? Windows 10 is a complete shitshow and this will likely be the last windows machine I ever personally own. Meanwhile LibreOffice suits all my needs fine where MS Office used to stand as they slowly tried to strip away the concept of software ownership and move to SAAS. Nobody uses edge anymore and Xbox is a joke. Maybe they are hustling on the B2B side, since they seem to have skyrocketing profits somehow, but I have almost no faith in them as a consumer at this point.
It’s not Google’s competitor making this story if you look at previous blog and critical comments from last few years you can see, google has internally been struggling on how to get more profit out of Google’s users. Also google PR is launching active campaign to muzzle voice by downvoting critical comments and blogs. Happened yesterday to my comments on HN and might happen to this comment as well.
Google used search for additional revenue by first make advertisement appear as legitimate search results by using a non prominent and small font text “Ad” in front of ads. But these are on top of search result. So asking users and company to pay to be on top of search results.
Launching new service and then offering free to attract developers and users and if it failed to generate revenue shut them down. [1]
As google is faltering on revenue in cloud, they started using security as pretext to scare the users of gsuite [2] and take away the functionality and force them to toe the google line and soon there will be change in charges and conditions.
Also shut down of gcp is a legitimate concern, google doesn’t care about users, they offer free to gain traction and if can not generate revenue shut down the service without regard for users.
I don't want to comment much on these threads because I am a googler and anything I said will be taken with pile of salt. But about your [2], from the thread it looks like Microsoft is doing the same thing -
> Microsoft’s Office365 will make a similar change in October 2020.
And I don't even want to comment on killedbygoogle.com. I don't see many Google Cloud or paid products there. And there was a similar Microsoft site posted here couple of weeks ago.
You can comment freely what you feel right, you work for google doesn’t take away logical arguments and that’s what the debate is about. I am not a Microsoft fanboy, they are no different and so is amazon when it comes to cloud lock-in.
I would have hoped in yester years that instead of Amazon api for cloud, companies would have adopted an open source standard, or just use API from openstack and extend them.
But that didn’t happen now cloud API are new windows kind of platform lock-in of old Microsoft. Personally I use apache-libcloud [1] to make my service as cloud agnostic as possible.
Also try to use a non USA company as backup mostly in Europe given strong privacy laws unlike US firms.
>Also google PR is launching active campaign to muzzle voice by downvoting critical comments and blogs. Happened yesterday to my comments on HN and might happen to this comment as well.
Source?
>As google is faltering on revenue in cloud
Source? They don't break this out in earnings so I assume you're just reading tea leaves here?
I don‘t believe they don‘t care about users per say. It‘s more the fact that they are blessed and cursed with the single most profitable business model the world has ever seen. Nothing is going to beat the profit margins of their ad business. It‘s very hard to then go to other models where the margins are much slimmer.
I still say cloud is their best bet besides ads. They absolutely do have to diversify. Once regulation hits it‘s gonna be tough. So cloud needs to be profitable when ads aren‘t growing 20% yoy anymore.
> Nowhere does it say that Google plans to shutter GCP if it doesn't reach #1 by 2023.
No, but it also doesn't say they're in for the long haul in case they don't end up on top and with Google's well documented and long history of killing products you really can't fault people for wondering if it is wise to build on top of GCP no matter what the advantages.
> Google's well documented and long history of killing products
How long will this tired meme continue to have legs for? Sheesh. Google is not going to shutter GCP right after spending billions of dollars on datacenters and fiber cables. The fact that HN is even entertaining the concept is embarrassing and speaks volumes about fall in the quality of discussion.
What you’re describing is actually the sunken cost fallacy. It’s entirely reasonable to assume Google execs treat the investments as sunken costs that cannot be recovered with GCP, because they rationally should.
Counterpoint, rational Google execs can't keep their jobs after spending 7 years sinking $Bs/year into unrecoverable investments in a product they eventually shutter. Google+ didn't require new physical infrastructure, and the other "Google kills xyz" stories are about small fry stuff.
Countercounterpoint, rational Google execs _definitely_ can't keep their jobs after spending 7 years sinking $Bs/year into unrecoverable investments in a product they eventually keep alive spending yet more $Bs/year.
Google spent quite a bit of money to run fibre in Louisville and then just abandoned the half buried cables. This kind of thing isn't unexpected for Google.
The problem is that Google has a credibility problem wrt supporting their products over long periods of time. Even if this particular report is wrong, it's still "true" in that sense.
For potential users, why should we take the risk when there are other competitors that don't have this problem?
If they'd even make some sub-brand explicitly for "shit we're serious about and promise we'll continue to invest in for a really long time"—and then actually stick by that—it'd go a long way. I think it's having no reliable indicator of their level of seriousness or commitment to a given product or service that's the main source of irritation for people. It's all just "Google" and who knows what the hell that means, this time.
Google could fix this credibility problem in a single step by simply saying "We won't materially change documented functionality of services A, B, or C in the next 5 years. Upon breach of this promise, we will pay any affected user $10M".
If they did that, I'd trust them. If they breached my trust, I'd get a cool $10M. They wouldn't shutdown the service because of the extreme financial disincentive to do so.
That opens them up to a ton of liability (what if something unforeseen happens), for very little gain. I can't imagine any company doing something like that.
The gain is hundreds of millions of dollars of new business that they are currently losing to competitors who don't have a reputation for dropping services.
I semi-seriously wonder if they just brought back Google Reader and seriously supported it how much goodwill they'd regain. Not that they haven't killed hundreds of other products, but that one just sticks out so much.
Given Google's track record of killing products people depend on, I can imagine the hesitation managers feel when they consider moving mission critical applications to Google.
Risking losing funding if they don't surpass AWS or Azure is a scary thought. It's a shame because GCP is very good from what I've read.
I will believe Google is serious about supporting GCP when they decide the job of interacting with their customers and partners is worthy of being labeled a 'Googler'.
Since it seems the majority of support and service at Google Cloud is made up of TVCs, with 2 year 'life spans', it seems Google is just hedging their bet, no need to bring people on permanently if you're going to downsize in 2-3 years. In fact, is it entirely coincidental that the goal is to be #1 in roughly the same timeframe as the TVCs are allowed to work? Seems like the same sort of lack of commitment to GCP that Google place in their workers.
Google has a graveyard full of products that have been shuttered. Imagine the risk in the decision - Do I bet my entire organization’s IT infrastructure on a product line that Google may just decide to disappear? It’s a trust issue. Trust is earned over time. I personally hope GCP stays and grows.
I think that what you’re seeing here on HN is a big problem that Google has — they’ve eroded the trust that they once had.
Perception matters. When you see the lack of trust Google has vs Amazon and Microsoft, you know you have a problem — both companies have track records of less than trustworthy actions.
Betting on Google’s offering is like betting your business on Bing. It might work but it’s an also-ran. And that’s before this leak. Time for leadership to clarify hard, ASAP.
Infrastructure should be built on rock, not sand. Google has sold too much sand.
Businesses need to focus on their next battle not finding a replacement API, web toolkit or cloud.
>So it's not about being #1, it's about being profitable.
Stop, do you mean that despite cutting the corners with technical support and having a weaker offering than the competition, GCP is not freakin' profitable?
Instead of spending time posting stuffs here, could your Googlers get back to your work to improve GCP's quality? Poor availability, of no help customer support, and slow fix of any problems is killing GCP.
> You set goals, you decide what happens when those goals are met. Furthermore, there is no concrete assertion here that what happens if this deadline is passed. "at risk of losing funding" can literally mean anything from changing FTE allocation to capital expenses to a million other things that get discussed at an executive level.
No customer is going to sink risk into a statement like that. Everyone wants to know that the investment increases and only increases.
In regards of profitable, I hope you also consider internal cloud computing usage as it were used by an external party.
E.g. if stuff like AlphaGo, Youtube, or the Search Engine used tons of cloud computing resources you should not count it as no profit made (actually loss because of power consumption, server housing, networking, ...), regardless if those cloud resources were anyways unused. Then it's up to your colleagues then to make the profit.
the problem is the timeframe. the decision is a sound one - at some point google has to cut its loses. but not in 3 years -- thats too early. if you want to be in the cloud business, you have to commit to trying to make a run for it for at least 5-7 years. it will take that long for the business to build
The article goes into more detail that it was a 5 year plan. Their source told them about meetings that took place in 2018 and they agreed on funding until 2023.
The article does say that there was a discussion about "exiting the business"; that's sufficiently terrible to be news, even without anyone explicitly talking about shutting it down.
I can't imagine a more damaging leak for Google Cloud. A company that's already notorious for abandoning projects now has a public date for when they'll abandon cloud. How could anyone in their right mind start building on GCP? They may as well shut it down today.
Maybe this helps GCP as the leak now makes this reputational for GOOG. They’ll have to keep it running to prevent people leaving because of the leak so they’ll do something drastic like pledge 10 years of support or something.
This is what has always terrified me about GCP. The business mode wasn’t there so I haven’t done anything serious since they killed/overhauled app engine years ago (aws seems to always make stuff better and I think Microsoft has bet the whole company on azure).
It always seems to me Google won the Search Engine battle, ads became their revenue source, and that is it. Android came out of necessity and stupidity of Microsoft at the time.
Comparatively speaking they are incompetent both Technically and Manageability, compared to Amazon and Azure, or to better phase it compared to Jeff Bezos and Satya Nadella.
It remains to be seen now Google CEO Sundar Pichai, now also Alphabet CEO will change. ( I could never understand how Google Cloud also had a CEO )
Google could very well be a one hit wonder. Coming from someone who works at a company with one successful product and many failed ones: Maybe they are too wealthy to want a second success. Life is cozy.
There's no evidence that YouTube is profitable[1]. In fact, there's no evidence any of this is profitable.
Android, Gmail, and Maps make money from ads, not from charging users directly. That makes them the same business as search.
Hardware businesses like Home and Chromecast usually aren't profitable, and Google (like Amazon) is likely using them to keep people in their ecosystem.
That leaves Drive, which might have some good margins, but they give it away for free to many users...
Every day that Youtube exists without competitors it further solidifies itself as the only video platform like itself that will exist.
Gmail makes money from business users and I would go as so far to say that if it doesn't make more money that Drive now it will make more money in 10 years from people who have lived their entire lives inside of a gmail account using it for business.
It said in the article G Suite might account for the bulk of Google Clouds revenue, since its combined in their financial reports with the infrastructure stuff and G Suite is extremely popular.
If the metric for success is revenue then, no, they are not by any sensible measure super successful. Their revenue combined it’s a small fraction of search/ads.
By that standard, practically all other companies are failures, since practically all other companies pale in comparison to Google search. So I would say your claim is absurd.
> Maybe this helps GCP as the leak now makes this reputational for GOOG.
Nope. Engineers know this product is dead. There's no way they'll risk their time and money on something Google doesn't believe in. We've been bitten time and time again, and adopting GCP is about the riskiest move you can make: hitching your entire project to a dying technology platform.
Since no one will be willing to make this gamble, GCP's growth rate will flatline. Exactly the condition Google said would cause them to shutter it. Self-fulfilling prophecy. Their platform just died today.
It's sad, because Spanner is actually really cool tech.
I bet Google engineers and PMs on GCP are reading this right now and already thinking about their future within the company.
In the future, we'll probably be able to look back at today as a major shift in strategy for Google.
I suspect this is down to how it's phrased by someone who may not be familiar with strategic planning.
It's common for businesses to work out several scenarios and estimate their outcomes. Then you look at the investment outcomes of those scenarios.
So quite plausibly there might have been these scenarios:
1. Invest More + High Growth
2. Invest More + Low Growth
3. Maintain Investment + High Growth
4. Maintain Investment + Low Growth
5. Divest
You pretty much always want to canvass the option to quit a business. You need it to keep your other estimates honest. Anything else is not only giving into sunk cost fallacy, it's flat out denying its existence.
So rejecting "Divest" doesn't mean it was ever a serious contender. It may be there mostly in the interests of completeness.
Dude, You're comparing apples and oranges. It makes no sense to compare GCP in 2017/2018 which is when the article refers to this conversation to AWS which predates it by 10 years. You don't think that AWS execs and Amazon execs had conversation about profitability in 2006?
Google has offered cloud computing with the release of app engine almost a decade ago. They kept it running this whole time and have only added to the platform.
The entire discussion in the memo seems to be more about how much they want to compete at the level of AWS rather than being in cloud computing at all.
Based on the developments and investment so far, it’s clear GCP wants to be a serious contender.
Again, having the option numbers crunched is not the same as discussing it or seriously considering it. "Quit" and "Do Nothing" should always be included in scenario planning. Especially the latter, because by ordinary momentum it's usually what happens.
I expect Amazon plans have included scenarios such as "quit machine learning" and "don't invest in on-premise units".
Why so much significance to such a minor line in a meeting years ago? What about the part where it was dismissed? Or does that part not matter?
I don't know how else to explain this. AWS and Azure are massive profit centers and GCP is slowly catching up. Cloud computing, machine learning, big data, IoT, and more have trillions in potential market cap.
It's completely nonsensical to consider that a company this big with this much investment already will exit that opportunity.
Interestingly we don’t even know if Azure is making a lot of money because Microsoft obscures it even harder than Google in their financial reports.
Microsoft has one hell of an enterprise sales and marketing machine though, probably the best in the world, so I’d assume they are doing quite well, even if they are mostly reinvesting it.
Nobody sent these signals. It was a minor discussion a long time ago. The billions in investment and growth of the platform are the exact opposite signal.
You seem to be reading this as if the CEO just announced that they're not sure about GCP anymore.
The article does not make it out to be a minor discussion and it was only 18 months ago, hardly 'a long time'. To me it is quite shocking that as short as 18 months ago Google would still consider pulling out of the market but then again, that is exactly the same play they did with other products so it probably should not surprise me.
> The billions in investment and growth of the platform are the exact opposite signal.
At Google's level I would not expect less than billions in investment, the growth of the platform is for now still substantially lagging behind the competition as far as I know, the bulk of the companies I look at are solidly in the AWS camp or in some cases running on Azure, Google is still the exception but getting more common over time. I look at about 50 companies / year so that's a relatively small sample but they are typically well funded start-ups or later stage established companies.
> You seem to be reading this as if the CEO just announced that they're not sure about GCP anymore.
The fact that it was mentioned in a meeting doesn't mean it was on the table. I'd more concerned with management if they weren't talking about exit points for a risky venture. A perfectly necessary point of consideration and I don't doubt that the same topic came up at Amazon and Microsoft.
For Google, if this is still a 'risky venture' then they are simply in the wrong business, they have the scale and should be able to offer their infrastructure at a cost that is quite profitable. There is plenty wrong with Amazon but they have never ever given off signals that they might withdraw their infrastructure offerings from the market. Microsoft also hasn't but then again, they do have a history of launching stuff and withdrawing it again (Xbox the notable exception, that took forever to take off and they stuck with it). Google has taken products off the market not because they weren't profitable but because they weren't profitable enough. That's the nightmare of early infrastructure adopters who will - by that time - be bound hand and feet.
If you're in competition with the likes of Amazon and Microsoft and you intend to go after the same customer profiles then you would do well to ensure that your messaging is what those customers expect and that every outward signal you give is that you're in it for the long haul, with a much longer horizon than the companies that are going to be your customers can oversee so that the trust becomes implicit.
That way if some internal document gets released it doesn't look as though you are still undecided.
Anyway, no need to take my word for this, let's just see how it plays out in the longer term.
Google obviously has a firm grip on what they’re willing to call unacceptable losses for an investment, but they need to realise that there is reputation damage every time they axe one of these projects that they don’t seem to be accounting for at all in their decision-making.
Something along the lines of "Google execs considered cutting GCP, later decided to set 2023 as a deadline for improved financial performance."
A few months ago, someone I know who is close to the issue mentioned something along these exact lines to me. I shortly moved all of my GCP projects to a self-hosted server. Glad I did it.
I'm curious what sorts of things you had running and how complex moving off was. I'd expect any cut to have some announcement period, which I'd expect to be longer than a few months, so you may not have needed to migrate so soon.
Agreed, this is a disaster. A lot of leaks simply don't matter all that much-- Apple might get annoyed at iPhone leaks, but they probably don't affect sales one bit-- but this is dead serious.
I hope Google's C-suite is treating this article as the PR equivalent of a Sev-1 servers-on-fire emergency, because it is. If they don't come out swinging today and announcing that they're all-in on GCP for the long haul, it's toast. It may be toast even if they do.
The Stadia pro tier fee is just $9.99/mo. The GPU and CPU you could consume playing a AAA game on that (yes, even on the purported "medium" settings) for just a few hours a day surely far exceeds what it could be sold for wholesale.
They don't have to match what it would be sold for wholesale though, just need to match the marginal cost. A graphics card that can run a AAA game is less than $500, so I don't think they are losing much on $9.99 a month. That's over $100 a year, and if each users plays 3 hours a day you can have 8 users using the same hardware at different times.
Now I don't think they are using consumer hardware, and there are other costs as well, but it seems realistic that they are making money off of it.
That's assuming people play games in well-defined shifts. I find it more likely to have 8 users play between 8pm and 11pm and zero users for the rest of the day.
Games are latency sensitive and thus region sensitive. Multiplexing opportunities here are poor. Stadia is a terrible idea business-wise as Google has defined it.
> (Incidentally, and you lose the game if you cancel your Stadia subscription.)
That is a factually incorrect statement. You lose access to any games you claimed for free as a benefit of your Stadia Pro membership if you cancel your membership; you regain access to those games if you renew, but you will not gain access to games that were offered for free during any interval where you were not a subscriber.
You retain access to any game you paid for, _even_ if you paid a discounted rate as a result of being a Stadia Pro member, even if your subscription is canceled.
Thanks for the info. I've been considering a new console purchase and had basically ruled out Stadia immediately due to the subscription requirement. This changes things a bit but I still don't like the fact that I can't play something without a connection to the big Google. I suppose this is just what it feels like to be part of a generation falling out of the marketing window.
> (Incidentally, and you lose the game if you cancel your Stadia subscription.)
What. The. Fuck.
First, they took our physical games and media away. Then they made us subscribe. Now they take away everything when we stop paying for our subscriptions? I can't even conceive of what the next step will be -- will they charge us for having memories of their IP?
We need to reduce copyright terms and kill this cancer of IP hoarding. The rent-seekers are taking ownership away.
Don't forget that they value new ways of spying on users at a non-zero dollar amount, and probably intended to slap ads all over the interface (maybe—maybe—not the games themselves) to make up the difference, or at least were holding that possibility in reserve in case it was needed or they just wanted to kick returns up a notch later (I'd be shocked if it didn't come up in any "decks" used to sell the project internally).
Fair, and that was certainly the YouTube model— years of free video hosting to saturate the market, now preroll and interstitials everywhere with "YouTube Premium" just a click away.
By this strategy, which disregard customer confidence completely, considering that Stadia isn't exactly well received, I seriously doubt I will put my money on it getting anywhere.
"That timeline was devised early last year, after an intense monthslong debate among senior leaders at Google and its parent company Alphabet over the future of the cloud business, a person with direct knowledge of the matter told The Information."
Can't imagine a leak like that happening. This leak could very well have been planted by MS or Oracle to help them sell their own cloud to clients.
For one thing, there is a lot of resentment about their lower quality bar. I am not a googler and never have been, but GCP has hired a lot of people I wouldn't hire even on a dare.
CTOs and CEOs who matter do not give shit about these sort of leaks. Google cloud is unlikely going anywhere because it is the same thing on which Google runs. Unless Google itself is shutting down the marginal cost of providing GCP to the rest of the world is less (still in billions though).
What would one prefer ? Google top brass not setting deadline to beat competition? Google not trying to be market leader ?
I want to put on end to the "you can't contact Google" comments that comes up every time Google Cloud makes the front page.
You can contact Google Cloud. They are very responsive. I'm currently at a small startup using GCP, and I've been at similarly sized startups with equivalent AWS spends. I have found it easier to have serious discussions with GCP, compared to AWS.
We have dedicated GCP contacts. We actually have a Slack channel we share with our dedicated GCP contacts, so we can easily ask questions. Two of my coworkers just got out of a meeting with Google PMs less than an hour ago. Another one will be meeting at a Google office tomorrow. I had a meeting with about a dozen Google engineers and PMs from their database team(s).
Yes, you have to pay for support. They have a couple different support options, depending on your needs. But you have to pay for support with AWS too.
The one bad thing about Google Cloud support is their Level 1 support. It is very easy to submit a support ticket, and they will respond quickly. But if you are highly technical, and know what you are doing, and can research yourself, Level 1 support it nearly useless. They do make it easy to to escalate, with a prominent Escalate button in the ticket. And you can always escalate through your direct contacts (Account Manager, Technical Account Manager, Customer Engineer). But it would be nice if Level 1 could be bypassed, or were more technical. I think they are trying to iterate on the process. They recently started offering to have a video call a lot of the time, which isn't my cup of tea, but I think it is a sign that they are trying to improve the Level 1 support.
If you are using Google Cloud as part of a business, you will be able to contact them. You will probably have an account manager that you can probably meet in person with. If you pay for support you will be able to submit support tickets, real humans will respond, and respond quickly.
I’ve had a similar experience where I could email PMs and we even had Google forward deployed engineers on-site.
The trouble is that their customer service is terrible.
We reported numerous issues and there was no tracking at all. For one UI bug in their webapp, a PM completely disregarded the reproduction instructions I gave him and made me have a half-hour kong Hangout with a remote engineer to prove the bug existed. The engineer even found more bugs while watching my screen.
The sales engineers (solution consultants?) are rather slimey, especially compared to the PMs who often admit when they don’t know something. This one eng kept pushing Dataflow even after we said no 100 times. It appears Google uses Dataflow / Beam internally now (instead of MapReduce) and so I guess at Google you’re dumb if you’re not doing the “correct” thing.
Also frustrating was how much money we wasted due to things like multi-gpu jobs taking 10-15 minutes to start and their GPU hypervisor pausing the whole system for 10-20 seconds incessantly.
If you use GCE, make sure Google is paying for the bloody edges. Get a ton of credits.
Cloud Dataflow is also vastly more expensive than using pyspark and Apache Beam is 10 times slower for most operations than Flink or Spark. I have to try the new flexible scheduling but DataFlow the last time I tried it 2 years ago was a massive rip off.
Also DataPrep was amazing until you run the DataFlow pipeline it produces and a < 1TB dataset cleaning costs $25...
Have had good experiences once the data is in bigquery and bq keeps costs low. But lately Ive been trying to figure out a semi-managed way of doing exactly-once stream processing for cheaper than dataflow. Possibly an architecture around cloud run might work.
$25 for cleaning a relatively small dataset does sound a bit high but not outrageous.. given the process abstracts out all the infra tweaking and setting up various systems. $25 an hour is probably a good estimate of how much a data warehouse like snowflake costs for a reasonable cluster as well, so unless you spin up your own spot ec2 instances and set up a spark cluster and perform the operations the cost savings are going to be marginal at best?
Do you need to run this pipeline more than once a day? Is it sufficiently important for your business case? I feel like 25 is a very small sum if any of them are remotely true.
On a related note, how practically reliable is the DataPrep->DataFlow workflow? Could a reasonably smart analyst with zero programming experience set them up and run them? Feel like that's who this workflow is built for, but I've not had the best of experiences with Trifacta demos in AWS with reliability.
In my experience $25 for 1TB of ETL is outrageously high. That’s 100 cores of GCP or AWS for 4-5hrs. I have some very very CPU intensive jobs that would take about that time for 300GB of data. But if it’s just parsing CSV / JSON and doing some joins then I’d think $5/TB would be more reasonable.
I agree a warehouse could get more expensive, but that’s where Athena or BigQuery is supposed to come in? BigQuery is rather expensive though. It can’t support reading Parquet natively (of course Google insists on their own format) so you have to pay their heavy tax versus just object storage and elastic compute.
Also the classic BigQuery UI is way less buggy than the new one, and the PMs won’t take bugs for the new UI.
“ Our benchmark results show that Apache Beam has a noticeable impact on the performance of DSPSs in almost all cases. Programs developed using Apache Beam suffered from a slowdown of up to a factor of 58 in the worst case. At the same time, there is one scenario where the query developed using Apache Beam is about as fast as its counterparts using the APIs of the corresponding DSPS. However, for most scenarios we observed a slowdown of at least a factor three.”
I had a 4 roundtrip email thread with their support over a speech API bug. Eventually sent them a narrated step-by-step screencast, and the guy is finally like "oh, it's not supposed to work that way." (Yes, that's what I said in e-mail 1!)
To be perfectly fair, I have had these types of experiences (plus more) with AWS as well. Their solution engineers are pretty useless, bugs in products never get solved and you get endless forwards and escalations on their issue tracker.
I think in general, scaling support for such complex a product that’s constantly changing is really difficult.
In one corner, I have the good experience of you, a semi-anonymous internet commentor.
In the other corner, I have the dozens and dozens of blog posts from HN front page from people who have gotten shafted by Google "support". And those are just the ones who had their ticket escalated, i.e. got lucky and made it to the front page of HN.
I'm not saying you haven't gotten great support. Hell, there are people out there who are happy with Comcast support. I'm just saying I've read too many horror stories to risk our business.
People, me included, just remember Google doesn't support their products. I won't be arsed to consult a list of Google products and properties to see if this particular one is going to get support or not.
If I buy something from Microsoft, I can contact Microsoft support. I don't have to roll the dice on it being actually supported by Microsoft, or only supported by other desperate users on an abandoned phpBB somewhere.
I don’t buy this claim that Microsoft has good support for all their products. If my Windows machine is running slow will Microsoft support help me? Why do my family members keep asking me for help with their computers if Microsoft support is so great?
To me, the idea of getting good support on any consumer software is laughable, and has no bearing on Google.
Because there aren’t a lot of stories of businesses struggling to contact Microsoft over a product they’ve paid for. Yet every other month there’s an article on the front page of HN of some paying business struggling to reach Google and the only way to do it is to go viral on social media.
The upgrade from Windows 7 to Windows 10 had similar consequences, but this could also be a reflection of the different types of audience the two companies serve.
For example, the person who wrote this book is probably never going to post a complaint on social media:
Figuring out whether a vendor will care about their customers in a given situation has a very simple, very universal "worst-case rule": if you're paying the vendor, and your "account" (in the financial sense) is typewise-indistinguishable from other accounts, of customers who pay the vendor ridiculously-huge amounts of money, then there will be a good support process for you.
Why? Because big-spender enterprise customers demand that vendors implement hand-holding red-carpet VIP support for them; and most of that work is, surprisingly enough, a capital cost, not an operational expense; and so you can free-ride on it as long as it's been set up.
On the other hand, if there is any way to typewise-differentiate you from the big-spender customers, then you'd better believe that they'll have support channels you don't have.
Well, maybe they should've called it "Alphabet Prism" or something and not "Google Cloud" so that it is set apart from their other product support in consumer's minds.
You're mostly likely getting lucky. We've gone through their highest tier startup program and have had so many account reps assigned that most aren't even with the company anymore. The ones that are don't reply to emails either.
GCP is the absolute worst at human support. Of course you get responses when you pay for support (since you're paying) but even then we've had several instances of refunded support costs because of a lack of reply over their SLA.
With Slack, anyone in the organization can file a ticket and get a human to respond in a few hours. If something isn’t great about Google Sheets (or of their cloud offerings), I have no idea how to submit and issue that will get attention of someone at google. They lose big time for this stance.
I don’t know if you still can but it used to be that with Microsoft premier support if you passed some specific exams you could bypass the first tier (who were excellent engineers) and go straight to escalation/ cpr engineers
If you create a culture of intense internal competition, where identities are tied to subunits (e.g. a product team) over the company, you incentivise leaks. Google's elitist, "product founder" driven culture, unmonitored by a detached leadership, almost guaranteed this sort of sniping and in-fighting.
This deadline should never have been set in the first place. They should have attached a massive bonus if they exceeded the target by that date instead. Of course, the CEO could have in their head the deadline, but that's be announced internally at earliest 6 months ahead of time.
anyone who will get money if Google Cloud doesn't, or anyone who doesn't like whoever is running Google Cloud or thinks that anyone associated with Google Cloud that pulls it off will be a competitor for the next raise in level etc. etc.
This is like corporate infighting for comedies set in the 50s really.
Good point. Google using GCP internally would be strong validation of the product and I think add more comfort for enterprising adopting it.
E.g. what coup would be for Google Search to run on GCP. Imagine the CIO conversation of, “well, if it is good enough for google search, it must be good for me”.
This article lists some of the AWS services that now power Amazon.com since they moved away from Oracle. So, while it may not have been true in the past, it does now appear that AWS powers a significant portion of Amazon.com.
Amazon using AWS today is entirely different from the origins of AWS or GCP or Azure being productizations of internal infrastructure.
AWS services have always been their own thing, they happen to be useful so today Amazon utilizes them just like anyone else. That doesn't mean the services started internally.
I think you're being a bit too literal here. Did AWS merely fork the infrastructure that Amazon.com runs on and start selling it, no. But are you telling me that it wasn't built on the lessons learned of amazon.com and that it doesn't share some aspects of how amazon.com worked a the time (with improvements to the bad parts)? I don't know the answer and maybe it was a clean-room implementation but that seems unlikely to me.
I don't doubt that Amazon.com runs on its own dedicated "installation" of AWS, but are you saying that it doesn't run on AWS at all? That doesn't jive with the infamous "API mandate" email that Bezos wrote that essentially kicked off AWS. Have I been misunderstanding that email all along?
If you think the API mandate necessitated or created AWS then yes, you misunderstand that memo. The API memo simply mandated a service-oriented architecture. That can run on any compute and indeed predates AWS by several years.
The origins of AWS are entirely unrelated, Amazon simply realized it was good at operating infrastructure. This led to the conception of offering infrastructure as a service. The initial AWS service offerings were built based on Amazon's experience and expertise but were not the literal internal services of Amazon.
Amazon (or Google, or Microsoft) is an anomaly, the things Amazon needs to operate Amazon.com are very very different from what most companies need. The internal services and architecture are purpose built although today they do heavily leverage AWS under the hood. Amazon.com is just a customer of AWS, there's no special installation.
Almost nobody needs Amazon.com's infrastructure (except Amazon) or could even afford to staff it. What people do need are generalized, managed services they can combine for their own use cases.
I'm confused because on the one hand you say that their needs are very different and then on the other hand you say that they are using AWS. Another commenter also pointed out that they are using AWS.
I think the only needs that are likely different are scale, so bigger instances, and the fact that they have a massive legacy infrastructure that can't be easily migrated off of. So that's pretty much what I would expect; Amazon.com does use AWS for new stuff but is still operating a lot of legacy as well.
Amazon uses AWS for some new stuff. New AWS services high up in the stack like AI services run mostly on AWS. Most infra-AWS services run on foundational AWS services like EC2 and S3. Foundational AWS services (EC2, S3, DynamoDB, etc.) don't run on AWS for obvious chicken-and-egg reasons. For all of the things that are done by latecomer AWS services you can count on there being an unrelated internal tool, since clearly Amazon was doing those things before they launched those services.
> Guthrie said Microsoft has long had a tradition to "eat our own dog food," which means it runs its own systems on top of the same software it offers to customers. Microsoft IT was an early customer for Office 365, Azure, and Dynamics 365, and has moved the vast majority of its internal servers to Azure, including its SAP implementation, payment systems, and test systems. This has helped Microsoft learn what it takes to run the technology on an enterprise scale.
That's normal progression. Hardly anyone at Microsoft used Azure in the beginning either, it takes a few years for teams to move infra and deprecate old stuff.
Same for Amazon, it took them very long also, including dedicated internal projects with focus on moving. But majority of Amazon and Microsoft workloads now run on their respective cloud offerings.
Google is probably still be at least 3-5 years before that would happen be aligned. GCP is quite immature and young, which this leak highlights also. I assume this would mean that also internal teams at Google won't be moving workloads to GCP.
In any enterprise competitor doesn't mean a similar product, it means any unit who can get headcount, budget etc. Google said in Q3 that they were spending a ton to grow cloud - which makes all the other kingdom building VPs concerned - they are the competition. Youtube, Gsuite, Ads, whatever else google hasn't abandoned that they charge for.
> The Google unit, which sells computing services to big companies, is under pressure from top management to pass Microsoft or Amazon—currently first and second, respectively, in cloud market share—or risk losing funding.
2) How Google treats all of its other "underperforming" products.
That only means that the team's head is on the chopping block. That's all.
> 2) How Google treats all of its other "underperforming" products.
This assertion is utterly absurd as you're describing the world's third largest cloud provider, and thus one of the dominant players, of being "underperforming" and justifying pulling the plug on a 8 billion-dolar a year business.
> Google is far from the world’s “third largest provider”
You're being a bit disingenuous. Your own source states quite clearly that Google is in the 3rd spot and arguably, if you cherry-pick results, you can place it in 4th.
By cherry-picked results I'm referring to posting 2018 results and ignoring how GCP's revenue doubled in the last 18 months, nearly matching Azure's revenue.
> So let’s look at the raw marketshare. It’s 10% of AWS.
Yes, AWS is 1st. What's your point? Azure is 2nd and GCP is currently 3rd and catching up with Azure. What does AWS has to do with whether GCP is 3rd or 4th?
I believe the 8 billion a year number includes its G-Suite products. My understanding is that GCP itself makes only around 1-2 billion a year (excluding G-Suite). That doesn't seem enough to cover the CapEx, OpEx and engineering costs they are sinking into keeping GCP running.
Only Amazon breaks out its AWS numbers completely today, but the other cloud providers bundle their cloud computing numbers together with other "cloud" numbers, so it makes it very confusing to keep track of what numbers people are talking about.
I'm fairly certain Google Cloud is not profitable. Even if it is profitable, Google will divest from business that don't hit a certain scale (billion $ businesses). Basically their goal is to never have employees staffed on projects which yield a low $ / employee ratio. They will invest for a while, but pull the plug after some arbitrary time if it doesn't hit the scale.
Exactly, which is the exact opposite of what AWS is doing. They have how many services that very few people use now? But at least they respect their customers and view the cost of running those services as goodwill for their customers.
They like to use small teams. I wouldn't be surprised if their least-used AWS product is still profitable (aside from brand-new stuff or obvious non-profits like Deep Racer). The idea is to get people into AWS and help keep them there. You don't do that by making life harder or unpredictable for your customers.
It's the opposite of what AWS is doing, but not what Amazon as a whole is doing. There was a recent article that said they're considering cost-cutting in the Alexa business unit [1].
And that's not a bad way to move at least in the short term. The problem is that longterm, people will avoid your products pre-emptively for fear of them being vaporware. Especially with something as serious as cloud implementation in an organization. I would hate to be the guys who now have to rearch all their applications to go from GCP to AWS.
That makes sense in the abstract but of the dozens, maybe hundreds, of projects Google is currently working on, how many outside of Ads make a sizable amount of profit?
A mildly successful cloud provider would probably make FAR more money long term than the median Google product. Pulling the plug on GCP to build a couple dozen projects that will mostly get killed themselves seems like a huge waste, even without considering all of the negative externalities of burning even more customers by cancelling a major service.
Breaking even doesn’t cut it since capital could be put toward something better. Unless it has some other strategic value (eg, bragging rights, growth potential, security, etc)
If you look at the way profits are distributed in a category, this goal makes perfect logical sense. What percentage of ad-tech profits do you think Google gets from being #1?
Unfortunately it's nearly impossible to be #1 in every business you touch.
"The Google unit, which sells computing services to big companies, is under pressure from top management to pass Microsoft or Amazon—currently first and second, respectively, in cloud market share—or risk losing funding. "
Since at least as of this writing, nobody else has mentioned this, and it's always someone's day [1] to hear something the first time, this is a variant of something that goes very far back in our industry called the Osbourne Effect: https://en.wikipedia.org/wiki/Osborne_effect
The traditional Osbourne Effect that earned the original namesake is to announce your v2 product a long time before it's available, thus trashing your v1 sales and eliminating the resources needed to make the v2 in the first place. But this is fairly similar; by announcing at what point Google is planning on throwing in the towel, everyone is going to think twice about going to GCP, because now everyone knows that everyone else is going to think twice about going to GCP. Consequently, the odds of this article actually creating the future in which Google pulls the plug on GCP are quite good.
Also, note how this is all about perception, and people's beliefs about what other people believe. Arguments like "But GCP is really good!" or "But GCP will surely continue to be very supported by Google right up until then!" would be missing the point. The point is that you need to change not only what everyone believes, but what everyone believes everyone else believes. Easier in some ways, harder in others, but none of those sorts of basic arguments will move the needle one bit. You need different arguments.
So, basically, in our industry, with this sort of folk wisdom widely known, Google damn near announced their plan to shut down GCP after 2023. They didn't. Quite. But damn near. Yes, it's a sort of second or third order conclusion from the literal text of the article, but it's a very easy one that lot of people can make, which ironically is the very thing that makes it true in the first place.
(Also, since PR is pervasively dishonest, putting out a PR release that says "No, srsly, we love GCP and would never dream of shutting it down" will not move the needle anywhere near as much as they'd like. Also in the "what people believe other people believe" department, Google has a reputation now of just shutting things down. GCP would be quite a bit bigger than anything else they've shut down, but, even so, they've shut down some pretty big "all hands on deck" initiatives before, like G+, so we're all going to believe that everyone else is going to find the idea that they might just pull the plug quite plausible. If they had a reputation for supporting things and gracefully shutting them down with care and concern, they'd be in significantly less trouble now.)
People with real knowledge to the matter knows that GCP is growing, and more companies are adopting GCP every month.
This is an attempt (which according to some of the comments here quite succeed) to create a story where there is no story.
Should Google Cloud aim to be better and bigger? Sure.
Should Google / Alphabet execs need to discuss challenges and strategy? Sure.
What and where is the story?
GCP produce new products and features every week [1]
At my company, we use all major 3 clouds.
The GCP support is super fast.
The GCP price is better than other two.
Google indeed has a learning curve as a company, and that is at the area of selling to the enterprise.
By its DNA, Google is an engineering company, and it takes years to understand how to "attack" the market, build teams of sales, marketing, and all the supporting infrastructure, and they do it.
Time will tell how successful they are, but I cannot see any option for them to shut down this great platform known as GCP. It simply make no sense.
There is a trend in the media in the recent months to rant about Google by all means and angles, and I tend to think this article is just one of those.
A rising tide lifts all boats. Are more companies adopting GCP because they are leaving Amazon/Azure? Or is it simply because the market is still immature and growing fast?
According to the title Google does. If the market matures and Google is not able to win customers from Amazon/Azure then they will not attain that #1 position.
Google did just that with G+, Hangouts, Checkout and a whole raft of others. Even if you don't become #1 you owe it to your customers to plan ahead and to ensure a business is 'unit cost profitable' at the scale at which you offer it so that you can sustain it.
Going for broke on projects and then canning them when they fail to attract the kind of success that a smaller company would be more than happy with is in the long run a losing bet.
> The group [which included Google CEO Sundar Pichai, Alphabet chief financial officer Ruth Porat and then-CEO of Alphabet Larry Page] even talked about—and eventually dismissed—the idea of leaving the market entirely
It's sounds like shutting down GCP at some point is definitely a possibility.
Why does “dismissed” mean so much but “eventually” mean nothing? That makes it sounds like it was a pretty serious conversation where most people up to this point may have been using GCP with the idea that shutting it down was never a real option for Google.
They're both meaningful. GCP still exists because they eventually dismissed the idea of shutting it down. But the fact that they discussed the idea means that shutting it down was a real possibility. I think it also means that if GCP's category position weakens out stays the same, they'll consider shutting it down again.
The Google unit, which sells computing services to big companies, is under pressure from top management to pass Microsoft or Amazon—currently first and second, respectively, in cloud market share—or risk losing funding.
That should give any Enterprise thinking about going all in on GCP a warm and fuzzy.
When I say how can you trust Google for your infrastructure seeing how they abandon projects, the usual response I get is that they would never do that with their “Enterprise” offerings.
On the other hand, while new accounts can’t access some AWS deprecated services, they still support features like EC2 classic years after they had a better offering.
Microsoft is also well known for supporting offerings for years.
>When I say how can you trust Google for your infrastructure seeing how they abandon projects, the usual response I get is that they would never do that with their “Enterprise” offerings.
And that's when I talk about Google Enterprise Search.
We're fully adopted and couldn't be bigger fans. After about 2014 I feel like AWS just got into a "how many services can we launch" game but most (if not all) of them have poor reliability and poor docs. Google has fewer options, but what they ship works as advertised.
I agree the cloud (pun intended) of doubt is unsettling when you're considering a switch, but I'd consider risking it anyway. They do have a tendency to kill projects, but killing a capital investment of this size is a bigger pill to swallow than not supporting xyz software anymore. So even if they do exit, I imagine it would wind down slowly.
Its kind of ridiculous that this is the moment in time they decided to pull the plug. It feels like they leaked this to ensure they could get out of cloud by 2024.
GCP cloud skills/certs have just started becoming marketable. Only recently has GCP become something you can seriously suggest to managers and point to BigQuery, Stack driver, firebase and cloud run as reasons to switch. Cloud Run, dialogflow, bigquery and kubernetes engine are literally at the bleeding edge of where the industry is going and Cloud Dataflow is seeing adoption among lots of companies tired of running their own clusters.
It really seems like they leaked this to ensure they couldnt compete with AWS and Azure so they can jump ship in 2023.
And you’re kind of demonstrating why enterprises don’t take GCP seriously.
How is their enterprise support? The decision makers would hear everything you say and you might as well be speaking Greek.
The one CYA question they are going to ask is where does GCP fall on the Magic Quadrant? If things go wrong with GCP they will have to explain why they chose it.
You wouldnt mention jaeger or istio either to decision makers.
You would say we can cut operational costs and modernize big data infrastructure while making our data warehouse more accessible(BigQuery). We can save money when APIs arent being used (Cloud Run). We can process real time data without increasing operational costs (Cloud Dataflow). We can debug systems failures faster (Stackdriver). We can launch products and iterate faster with a lightweight frontend team (Firebase). We can build conversational interfaces rapidly (Dialogflow).
You reword in terms of results rather than tools for decision makers. Of course this is all
moot because you have to convince them google cloud wont shut down by 2023.
Making operations google’s problem is precisely what drives BigQuery, FaaS and Dataflow adoption. Its operationally safer when you dont need a hadoop team to maintain your data infrastructure.
See Nytimes, Twitter and Spotify for examples.
Now if you’re talking about future-safety there was an argument...until a few hours ago.
Again, you’re talking about what’s “operationally safer” not what’s “reputationally safer”. If AWS goes down, no one is going to question your decision - and you’re in the same boat as a lot of other people. If GCP goes down and everyone else is up, people are going to ask a million question.
Every company you named isn’t the same thing as just arguing to the powers that be - AWS is what Netflix uses. AWS has many more “referencable clients” that matter.
They put the guy that ran ORACLE'S CLOUD STRATEGY in charge of GCP.
ORACLE'S. CLOUD. STRATEGY.
Let's hire Thomas Kurian, bang up job on Fusion and ... oh wait, what do you mean with 'total failure'?
Oracle bought Sun, it had the keys to be the third force against AWS and MSFT - and they totally blew it. Just an amazing fumble, worse than MSFT and smartphones.
Are we talking about the same Oracle Fusion that powers most of the world's corporation's accounting backends today (including Netsuite, which Oracle acquired several years ago)?
Because if so, he did a pretty good job of achieving his objective: establishing market dominance in the face of superior competing products.
Hiring Kurian is a message to investors that Google is serious[1] about trying to take over the cloud platform space, through any means necessary.
[1] At this moment, at least. Given Google's history, their actual commitment to their cloud platform remains to be seen.
I think the logic was a recognition of how badly Google sucks in enterprise. Therefore the goal was to hire someone who embodied "enterprise" albeit in all aspects.
It would be better to phase how badly Google sucks at doing business. From SME to Enterprise.
And that is not a surprise since Google's culture is to get rid of the human / customers interaction and wants to automate everything. But in real world all business are conducted by person.
To be fair, my understanding is that Oracle basically went in the opposite direction that Kurian wanted to go on cloud strategy. With that said, yeah, nobody's thinking outside the box when hiring an executive from Oracle.
Oracle Cloud in the last year has been completely changing and restructuring their Cloud and the actual organizational structure of their Oracle Cloud team to match AWS. They even started calling it Oracle Cloud 2.0
Didn't know this, pretty crazy hire... what's the reasoning behind that. Maybe they were desperate as the vmware founder wanted to retire and they needed to fill in the seat quickly?
Judging a decision based on its outcome is a very good way to get articles into the Harvard Business Review, and a very poor way to evaluate a decision.
AWS is unassailable as number one, but Azure is more fragile than they might appear in their #2 position. They use their Office online revenue to boost their numbers (EDIT: apparently they don't do that anymore), and their utilization is probably lower than the competition because they rely heavily on bundling new cloud services into legacy contracts at renewal to reach their targets. So lots of Office / Outlook / AD / Windows Server customers are getting hundreds of thousands of thousands of dollars worth of Azure that they’re not using, but count against their revenue number anyway.
On top of that, Azure’s technology stack is dreadful. Definitely the worst of the three.
What Microsoft does have is focus and dedication, they are clearly willing to spend billions of dollars in capex to strengthen their lead.
I think if Google were to get serious about communicating focus and dedication to their cloud business (the opposite of what they’re doing now), they could conceivably catch up to Microsoft.
> So lots of Office / Outlook / AD / Windows Server customers are getting hundreds of thousands of thousands of dollars worth of Azure that they’re not using
Sounds like an arbitrage opportunity. Have tons of unused Azure credits that were forced on you? We'll happily use them up and pay you back a fraction of that credit.
I'd disagree. Microsoft has an EA with every significant corporate entity, and basically taxes everyone. They have the ability to leverage those relationships unlike any other.
Leveraged relationships Oracle has, and look how that turned out for them. Universally hated have they became.
Microsoft after a decade of Ballmer is not far from where Oracle is now, despite some recent positive reputation gains. They are already squeezing those relationship by bundling services, and it will be too far too quickly.
While AWS has a strong #1 offering in low-hanging fruit products like vms, dbs. I find they have terrible MVP-like execution on niche offerings (and support sucks too) for things like Fargate, K8s, one-off job execution. Azure has a stronger, more thoughtful offering in niche solutions.
This is partially because the customers of AWS do not care too much about these solutions. I have seen other solutions used by large customers (still probably classified as a niche) getting improved over time significantly. Fargate and k8s offers very little over a well designed EC2 cluster. This might be another reason.
I think that’s right. ECS for example is mature and in my opinion still the most robust and scalable “container-as-a-service” out there. In spite of the kubernetes hype, I suspect that ECS is more successful than any hosted kubernetes hosted solution out there (in terms of aggregate workload volume and revenue).
EKS on the other hand still feels like a checkbox service.
I'd beg to differ.
Between Lambda, DynamoDB, Aurora Serverless, S3, SNS, SQS, API Gateway and now EventBridge, which cost basically nothing at MVP scale and all of which's usage is drastically simplified by tools like the Serverless framework or Amplify, AWS is my go-to for anything MVP
I tried Azure. I ran away. I liked GCP, but it seems I'll steer clear for the time being.
I think you're completely wrong. Azure is in the prime position and AWS is on fragile ground. Yes, they have a service for everything, but they've alienated their partners, they aren't viable in an entire industry (Retail), and a significant fraction of their IP is based around creating managed services out of OSS software.
My projection is that -- if everything continues along the same trend line we've seen through 2019 -- in five years MS will comfortably be on top and Google Cloud will have rough parity with AWS, but be significantly far behind MSFT.
Disclosure: I work for Google Cloud but this opinion is my own.
I recently did some consulting work with a trucking company who is modernizing their mainframe systems and moving to the cloud. They made the decision to cancel all their AWS migration plans and go 100% into Azure due to Amazon moving into logistics and basically becoming a competitor.
I don't think it's dreadful (but maybe no. 3) and I think Microsoft are the best out of the 3 at Enterprise sales. Google prob the worst of the 3 in that regard (and no. 1 with stack).
I've been quite pleased with the Azure stack myself. There is the issue of figuring out what the marketing team named the damned thing you are looking for, but I had the same problem with AWS.
I am focused primarily on .Net though, so your YMMV by a lot.
not exclusive to azure as aws lambda requires API gateway now, but if i wanted to create a cloud function, azure auto generates like 2-3 other separate requires services for storage, http endpoint, etc
You do not need an API gateway for an HTTP triggered Azure Function. You do need a storage account, because that's where it stores the files that back your Function.
I think the Azure stack is very solid - UI sucks maybe, but it's best not to use UI anyway.
Also, Microsoft naturally is the to-go place if you want a partner to trust as an enterprise. They have proven to support systems and software in the long run.
Also, Pentagon just picked Azure over AWS - so Azure has a good run and bright future I'd say.
The resource template format is inhuman gibberish designed for machines. It does idiotic things like manually hard-coding enums in the template to support drop-downs. This means templates go "stale" very quickly and are super fiddly to update to support things like new VM sizes.
You cannot export resource groups bigger than a certain size, so if some idiot in your org makes them too big, there's no backup even in this minimalistic sense.
There is no "backup" for anything but a handful of object types (Web Apps, VM disks, Secret Vaults, and that's about it). If you want to protect your data, you'd better have an on-premise copy and 100% automated deployment capability.
Speaking of automated deployment capability, 99.9% of the Azure docs mention manual deployment or procedural (PowerShell) deployment. Declarative is clearly an afterthought, and mentioned only in passing. Because of this, ~95% of Azure deployments I've seen in the wild were entirely hand built and very fragile. If some malicious admin ran Remove-AzResource on everything, those orgs may as well declare bankruptcy the next day.
Despite this fragility, Azure sales reps are pushing-pushing-pushing for their customers to go "all in" on cloud, be "cloud first", and to basically "lift and shift" their data centres. You gotta break a few eggs to meet your quarterly targets, am I right? Right? Guys?
Speaking of the PowerShell, it's 99% auto-generated from the JSON API binding definitions, and about 5% of it is documented even in the bare minimum sense.
IPv6 "support" is hilarious. They very generously provide a /124 static public subnet prefix. So many addresses! A whole 14 of them! Woo! It's the future now! No need for NAT! A routable address for every endpoint! Let me get right on that, soon as I figure out the fiddly scripting needed to allocate addresses from hundreds of tiny pools. Much fun.
If you delete a DNS Zone by accident, you can't properly recover it within 48 hours because they randomly pick one of 10 name server pools. Hence, your NS bindings at the registrar will point at the wrong servers and even if you update this, there's an inevitable propagation delay. I am aware of workarounds for this like Resource Group Pinning, but only because we jumped up and down and forced support to admit that it's a problem. This little "surprise" is still undocumented.
Speaking of DNS, until we forced Microsoft to fix it, the only way to back up Azure DNS records (az network dns zone export) would corrupt CNAMES and wouldn't round-trip.
Azure DNS uses an idiotic Zone->RecordSet->Record hierarchical structure, which makes small incremental changes hugely fiddly in scripting. You have to download the existing RecordSet, modify it, and then send it back with ETags intact. You can't treat each record as independent rows in a table, even though they effectively are.
The Azure DNS servers don't send "Additional" records (e.g.: the matching A records for the target of a CNAME record), which means that a) it's slower for clients, and b) they can charge you more. They have zero incentive to fix this, because it literally doubles their revenue for alias records of all types.
DNS Metrics are collected every 2 hours, but the graph displays only daily or hourly intervals, so you either get no detail at all, or a sawtooth graph that gives you no useful feedback at all at best, or is panic inducing at worst. Imagine making a small change, glancing at the graph, and seeing it hit zero. Then... staying at zero for an hour. A joyous time, for sure.
I could go on and on.
Azure has some neat stuff, but they're moving fast and breaking things, and half their products are basically MVP garbage authored by the lowest-bidding Indian outsourced teams.
To change the vnet of a VM, you still need to deallocate it and restore it into the new vnet.
Anything they don't offer directly, they offer a pile of canned VMs authored by Bitnami, who adds their own layer of management scripts to them, which (in our case) corrupted the filesystem on restarting from the great Shellshock cloud reboot.
Azure is, in part, an ad-hoc cloud offering made up of services offered by mutually incompatible business units. Adding analytics in the dashboard has a 50% chance of routing you off of Azure to a separate MS monitoring/analytics solution (OMS) requiring its own licence.
Their linux management agents have repeatedly locked up our VMs.
Because of the way Cosmo pricing works, we were billed ~$5,000 over 3 months for 50MBs of data that was basically unused at the time--because you pay for "reserved network units" on a per collection basis (as in a mongo collection), and the floor pricing drove it that high without any utilization.
As parent said, it's a bunch of MVP garbage, to which I'd add: to tick a bunch of checkboxes for Fortune 100s to sign 8-9 figure deals with promises of volume pricing. Operationally it's flea market.
Just to add to my previous post... this is stuff that has irritated me in just the last few minutes:
* When selecting a VM size in the new-VM wizard, it silently resets it to "DS1 v2" a bit later. You have to go back to the first tab and re-select the size you really wanted.
* This then promptly resets my "Already have a Windows Server license?" setting so then I have reapply that too.
* The default VM template suggested is Windows 2016, not Windows 2019.
* The selection box for the region keeps resetting to US East, even though that's the worst location for me. There is no default location setting in the portal that I can set for myself. This is an Enterpise-wide setting only.
I do appreciate the fact that the Azure team is here, noting feedback and routing it. I had a longstanding complaint about how Cosmos' mongo API was incomplete, and my complaints about it here were noticed and it did eventually get fixed. So, it may be slow, but the Azure team isn't unresponsive.
Thanks for providing us your specific feedback. I am part of the ARM Deployments Team. We are looking to support export for resource groups(RG) with more than 200 resources. One of the top customer requests with regards to export was the ability to multi-select individual resources in a RG to be exported which we added a few month ago.
We have renewed our focus on ARM Template deployments(our infra as code offering) and have a dedicated team working on addressing the pain points described in the thread as well as other deployment issues. We are not going to solve all the problems instantaneously but we are consciously working on it.
Would love to get your input on other ARM Deployment related feedback and areas of improvement. If you haven't had a chance you can check out some of the deployment improvements we announced recently in this video: https://www.youtube.com/watch?v=3D-JIKShrws
* Comments in JSON -- the standard explicitly bans this, making ARM templates no longer JSON, and hence not processable by standards-compliant tools. Including, hilariously, PowerShell. ConvertFrom-Json barfs, as it should. The correct approach is to have an explicitly non-JSON format like TOML or YAML that allows comments.
* "ARM has this whole time been secretly case insensitive" -- Someone clearly hasn't been notified about the harmful consequences of the Robustness Principle. [1]
* "People say that nobody authors ARM templates from scratch" -- Well, yes. This is because the template format is barely more than serialised RPC and wasn't developed like a proper programming language.
* The tab completion suggests strings that are too long for the dropdown UI and are cut off, making this feature unusable for a lot of ARM fields, such as the resource "type" values, which have long prefixes.
* "This shows as something different in the Portal" -- You have to memorise internal project names, aliases, or other difficult-to-discover names for things instead of simply using the display names. This is not user-friendly.
* A general issue I've seen with JSON-style configuration files is that things that should be side-by-side (e.g.: a bunch of related VM names) are scattered and separated by a lot of repeated "ceremony" text. Take a look at Altova XML Spy, which has a hierarchical grid view that moves such text fields into adjacent columns and rows, much like Excel does. This makes it far easier to spot mistakes or to make bulk changes.
* The "WhatIf" support is hilarious. I updated my Az module yesterday and I don't have that command. In the demo, it did something terrifying and the presenter had to close it. The output looks hideous, not something that looks useful to me.
This is what it boils down to: I don't want color-coded output that is almost-but-not-quite-JSON. I want tabular output. I want to see what happens if I pipe in 1K rows from a spreadsheet into a template. I want ConvertFrom-Json to work. I want to be able to validate things before deploying them. I want to see WhatIf support in the Azure Portal, with the final look. I want more than one tool to be able to process ARM templates. I want ARM templates to default to 2019, not 2015.
Tooling isn't going to fix this.
I want a language designed for humans.
This demo video to me felt like putting lipstick on a pig. Sure, it's better looking, but... still a pig.
Thanks again for the feedback and totally agree with you on some of the points you have made. We are exploring some options with regards to the language. Wondering if you would be willing to do a call with us and get your thoughts\feedback. My email is satyavel[at]microsoft[dot]com if you would be willing to chat.
Indeed, Azure to me seems to be focused on achieving success on absolutely anything else besides technical quality and competency. Their strategy seems to be acquiring every certification and qualification under the sun so that they can be the only legally viable option for many large customers.
Kind of sounds like the good old Microsoft that made Windows a stable in nearly every enterprise...
Thanks much for providing specific feedback on areas where Azure could do better. This is really helpful and I've followed up with owners for the components you mentioned. None of these points fall directly into areas that I work on, but I'll make sure teams are aware.
I did want to correct one misconception, which is that we do not outsource Azure product development (we do, from time to time, hire contractors to add force on projects, but these projects are led by FTEs who are involved on a daily basis). There are certainly areas where we strive to target lean MVPs so we can get new products into customers' hands faster and get feedback earlier. I'm sure there will be some sharp edges there and we greatly appreciate you giving these services a try.
My guess is that if Microsoft is able to hold onto the JEDI contract (after Amazon's lawsuit/appeal), Azure is going to be much more unassailable than you think. As it's a huge in to being the de facto US government cloud provider.
As a Google Cloud Platform customer, this is pretty concerning news and seems likely to drive me to start looking at alternatives and also likely to avoid placing any dependencies on Google-specific platform features.
So, if any Google Cloud employees are reading this, please report back that this just killed any interest I had in Anthos, Google Cloud Run, the managed Google "Kubernetes Apps", etc. Pretty much anything that isn't basic compute VMs, storage, or out of the box Kubernetes.
This is certainly not as true with Microsoft or AWS. You remember the old saying that “no one ever got fired for buying IBM”? If someone bet the farm on IBM in the 70s, they could still buy compatible systems.
Do you think you are at more risk betting your enterprise on Microsoft or Google?
> Do you think you are at more risk betting your enterprise on Microsoft or Google?
Microsoft? No.
Google? Absolutely.
I don't trust Google for anything that isn't related to selling ads. And I trust they will maintain their relationship for God-fucking-awful customer support. And I assume they will continue their tradition of sunsetting projects haphazardly.
After seeing how amazingly much better Facebook properties targets ads I won't even trust Google to do that.
Lately I've been using Instagram multiple times a month and the ads I get there are relvant, sometimes even useful (I once actually bought a product from one such ad that I actually like and wouldn't have been aware of without, and I am a typical HNer that have always said ads are wasted on me, because they have been for 10 years).
Contrast this to Google who has way more datapoints on me yet end up showing ads asking if I have old mens problems or suggest I join a dating site for elderly. And that is just the icing on the cake: they've been pushing dating sites ever since I got engaged over 5 years ago.
My account might be a fluke but I guess I am not allne.
I don't think you're a fluke at all. I've heard lots of "complaints" about how accurate Instagram ads are, because the products being advertised are so appealing. I've experienced it too, the ads I see on Instagram are almost all things I could see myself buying.
Even the newest mainframes are still compatible with code and binaries from the 1960s. You might pay out the ass for the privilege, but it's a legitimately impressive technical accomplishment.
> I think this is perfectly healthy and a stance you should take everywhere.
There is tons of value for non-VPS: serverless functions, messaging queues, object storage, managed databases, CDNs, monitoring infrastructure, load balancers, etc.
Unless of course they aren't going to disappear or gouge you in a couple years. But AWS is very unlikely to do that.
It makes sense to take advantage of platform features if you expect stability and for those features to ease your development process in some way that justifies the cost of vendor dependence.
There are certainly features of Google Cloud that meet the latter criteria, but stability? This leaked report suggests we shouldn't expect that past 2023.
Cloud Run is compatible with the Knative open source project, so you should be able to take your apps to a different cloud like you would with Kubernetes.
Ive never been interested in Azure until now and Ive heard CosmosDB, their FaaS product, and their client libraries are all pretty nice. If their docs, client libs and cli tools are all as nice as GCP I will have to switch. I dunno if they have a BigQuery alternative though.
Actually I started using Office365 instead of Google Docs recently and have been quite happy with the results. Especially OneNote’s live sync.
I'm astounded that they would seriously consider closing GCP after committing so much time, money, and effort into building it.
Every time there's a GCP story on HN commenters say "I won't use it because Google can just pull the plug." I thought this opinion was silly as GCP is the #3 cloud provider in the world, not a free service like Reader.
Guess I was wrong, nothing is safe. Google wants #1 or nothing at all.
In my opinion, it's just a consequence of the culture of stack ranking. The only thing that promotes you is if you deliver something. Not supporting something well. Not having an established, loyal customer base. Or, even if you provide good support/bug fixing. It's make something go public. Doesn't matter in what state. It's why I wasn't in the least shocked about Stadia's problems.
Amazon has stack ranking and hasn't fallen into this particular trap. It all depends on what you prioritize and what you reward. At Amazon, I got the sense that it was actually harder to get a promo if you were on a secret new product team. I'd heard younger engineers talk about doing stints in less fashionable departments as the ladders seemed easier to climb.
Google just seems to care about the launch. Not how happy customers are. Not how much money is pulled in. Not how well it fits in with the overall strategic vision of the company. Just launch points.
I never thought they would shut down Google+ either. It may have been a “failure” in the sense of not beating Facebook, but it had enough traction among hacker circles especially that I assumed it was here to stay just as the cross-Google identity system if not a standalone Facebook-alike.
However tempting (and it is very tempting), how much they’ve already spent should never figure in this kind of decision (aka the sunk cost fallacy). What matters is the return they will get from future commitments.
More precisely, always have a backup plan. Nothing wrong with running on GCP if you're using it in a platform-abstract way (either by using it as a simple hosting provider, or by going through a layer like the "serverless" framework).
GCP can seriously be #1 because their kubernetes platform and cloud platform is much better than other cloud providers I have used. I started off with their Google Cloud Engine (?) and Big Table they provided and even though it was not really pure Java, I loved it.
And then they pulled the plug on my accounts because some of my android apps had "likeness or similarity to other apps" which I have absolutely no idea what it meant. Then they pulled the plugs on my colleague's account because of equally vague reason.
I'm surprised that the overall sentiment of the comments indicates that gcloud's offerings are inferior to aws / azure. I've worked with all of the clouds and have found Google to have the best product. Their GKE with built in Istio support is a killer feature. Spanner is best in class. Competitors do not provide a useable alternative to Firestore realtime database. Pizza Hut is the most popular pizza in the USA, doesn't mean it's the best.
Endorsed. If GCP goes away, I'll be sad to see it, because the tech itself is great. Google don't necessarily have to be the ones running it, maybe. Perhaps another bigcorp could acquire GCP from Google?
I couldn't help but laugh out loud when I saw this headline on Twitter. It's just kisses fingers.
So now that everyone knows they are waffling on whether to stay in the business (and have a deadline), Microsoft and Amazon just need to increase the competitive intensity for the next few years to drive Google out and instead of cloud computing being an oligopoly it'll be a duopoly.
If you're a startup you should seriously consider colocation for as many of your workloads as possible, because the long-term future of this market is AWS and Azure being an extremely high margin duopoly with massive barriers to entry. You might see aggressive competition before 2023, but 10 years out it's not going to be that.
The NYT wrote a piece a few days ago about certain startups considering antitrust complaints against AWS, and more of them should consider that [1]. The FTC is investigating AWS [2].
Microsoft and Amazon don't even need to do anything. Knowing that there might only be 3 years of service remaining, I definitely won't use Google Cloud and I'll tell everyone I know not to use it too.
Microsoft Azure: The cloud service that wont abruptly close out of nowhere with no transition plan leaving you stranded, and a service that actually comes with that thing you call when you need help and support.
Mostly all you need to know about Microsoft's support lifecycle is that my Windows Mobile phone just got a security patch on Patch Tuesday, and apparently Microsoft also extended its support an extra month into January just because.
Microsoft platforms most people think are long dead and buried still get maintenance updates years later.
Like the Microsoft Azure "German Cloud", assured to be GDPR compliant. We spent quite some time and money to migrate from AWS to that solution in 2017.
Guess what? It has been shut down by the end of 2019.
We're back on AWS now. We were (I guess) not an insignificant customer, paying around 8k €/month, with the prospect of growing significantly (which has happened in the mean time, but on AWS).
If you still don't believe me, please provide your public key or email address so I can publish a response.
Maybe it doesn't stand out: Microsoft Germany delivered the promise that US authorities are technically unable to access stored data. The "new" data centers were without that promise, making the USP obsolete.
> Existing customers can continue to use the current cloud services available today.
> Existing customers are not affected by this and can continue to use the Germany cloud without restriction.
You don't see any Windows Phones being sold, but they still years later being supported with security updates.
No one can possibly fault a business for stopping continuing development of an unpopular offering. It's when it becomes unsupported and leaves people stranded that the the real problems happen.
Microsoft have plenty of platforms and products that were abandoned overnight. It’s hard for me to even think of Windows Phone, for example, without thinking of what happened to WP7 buyers when they abruptly pivoted their mobile OS/runtime strategy with the incompatible WP8. The most they got was a WP7.8 update that skinned the immediately-dead platform to look like the new one.
It’s a matter of visibility and what kind of customers own the products being shut down. From a PR perspective there’s a big difference between killing an enterprise product with few large customers, and killing a consumer product with many small ones. Even if the few complaints are louder, the many complaints around the latter would still totally drown it out. I’m sure they have analytics to know how many people still use Windows Phones and weighing the possibility of an outcry e.g. if all their unsupported phones got 0wned and they collectively took to Twitter about it.
Windows Phone was like the opposite of Microsoft's strategy over the last several decades. They decided to move fast and break things like we're all told to do and they were roundly punished in the marketplace for it.
They just lost my business. I was planning to use them over AWS for a new project. Now that I know they might cancel it like the other Google experiments, I'm not going to risk that investment. I'll stick with AWS.
They have a great platform, but I think they've just lost the war by the top brass waffling on this. That's a shame. There could have been a lot of money in third place, and as long as you're in the game there is a non zero probability of moving up the ladder.
If you're a startup you should seriously consider colocation for as many of your workloads as possible
The vast majority of startups do not succeed or fail based upon their ability to control their datacenter spend. They succeed or fail based upon their ability to attract customers and grow revenue. Colocating and managing your own servers instead of taking advantage of the ease of using the cloud will only be a distraction for almost all early stage companies.
Hi, startup here, with a colo and an AWS bill. The amount of time spent per a month on tasks that "AWS Handles" is almost 0 month to month. Maybe 4 trips a year into the DC. Calculating our AWS bill for fully replacing the colo is well over 4x using small 4-8gb instances. Our virtualization hosts(plural) have 64gb ram+ and 24 cores. That allows me not to focus on any scale issue since I can just shove everything into ram for no extra cost. I can buy more DDR3 ram for the servers for less than my hourly bill rate at this point.
I don't know what your company is but knowing it's sensibly renting rack space rather than jumping on the cloud bandwagon because it's the path of least resistance would give me confidence if I were a potential customer.
Perhaps that's true for a startup still trying to prove their concept, but it might not for one that has achieved product/market fit. "Hybrid cloud" is starting to get traction for large enterprises adopting the cloud [1]. They're going from on-premise in their own data centers to colocating at internet exchange data centers, which act as on-ramps to the cloud oligopoly.
There are options other than "put your entire infrastructure on Amazon" and "hire someone fulltime to manage your physical infrastructure".
Unless you have some truly massive amount of hardware (in which case it's _definitely_ cheaper to hire someone than pay Amazon), you can generally just set it up and forget about it save for periodic maintenance and emergencies.
So what you really need to pay for is the time of someone who can do these things for you and funnily enough, such a service exists. I believe it's common in the not-startup world: a managed services provider (MSP).
And in the end this really depends on what the business is actually doing. AWS is _really_ price inefficient but in particular for bandwidth. If you're doing anything bandwidth-heavy like video, images or backup, you want to keep the hell away from any kind of cloud.
If Jeff Bezos sent an email to the company saying "We want retail revenue to increase by X% by 2023" would you assume that they'd shutter the whole thing if they didn't meet that goal? Of course not.
Why is a goal and strategy for Google treated differently?
How is it ridiculous? The article said GCP would receive reduced invesment if that target isn't met. It's a bit hard to grow without investing, so it smells to me like a precursor to sustained engineering.
Google migrated from a weird, Apple like "come buy stuff from us, but first solve this riddle" sales process with reps with huge territories to a more traditional enterprise model. Based on Linkedin updates, the Oracle class of 2015 has basically moved in on the sales side. Those guys usually make premium bank, so they need to bang out deals or GTFO.
> The article said GCP would receive reduced invesment if that target isn't met.
And the comments upthread said that GCP would be terminated instantly if that target isn't met. You genuinely don't see the disconnect in that interpretation?
Google has built a strong reputation as a company that shuts down any product or service that isn't #1 in the target market. it is not unreasonable to believe that might extend to their third- or fourth-place GCP product.
So I just counted my RSS feeds via Live Bookmarks (dang you Firefox for killing them shakes cane) and I have 60+(! I expected like 30 tops) RSS feeds that I check regularly.
Most sites have RSS feeds, even the heavy hitters like YouTube.
That's a statement that cries out for analysis, although I'm not doubting you could be right still - it's not the kind of thing one just states without backing it up.
Reader was written using a bunch of proprietary Google tools, deployed on proprietary Google infrastructure and used a proprietary Google system for authentication.
It's hard to even imagine how it could have been split apart from the mothership at all.
If he sent an email to the company saying "We want retail revenue to increase by X% by 2023 or we will decrease our investments in the retail parts of the company" I might be worried that the retail parts of the company were going to enter into a death spiral given Amazon's problems with quality control already.
If I also thought it was unlikely that they would hit their target I would be really worried (assuming of course I had any reason to actually worry about how Amazon's business was doing)
If we were talking about a company that had a history of shutting down successful divisions because they were not successful enough I would think "hey 2023 is when the shutting down process is likely to start"
Of all the comments about damage in this thread I think this is the most damaging. In other words if I were AWS and Microsoft sales I would be calling all the big clients on Google Cloud and scaring them about their future prospects and wooing them with some sweet deal.
Even if Google does an about face in tomorrows news cycle the time for MS and AWS to strike is now.
>Locking yourself into a colo architecture is just as risky as locking yourself into AWS
That's the game theoretical predicament: Colo's can take advantage of all this and increase their prices as well.
We just don't know how this whole thing will shake out. Given that uncertainty, I think you're proposing the only reasonable way forward for a little guy.
Be agnostic. Meliorate your market position by differentiating in some key area of your business where you can actually secure a competitive advantage.
"The Cloud" is just managed hosting and SaaS/PaaS. There are literally thousands of such providers in the world, and until everything about providing managed hosting and SaaS/PaaS becomes impossible to compete on, there always will be.
The only "duopoly" will be in terms of the hyper-scale size of such providers, but very, very, very few customers actually need such hyper-scale. Not to mention, you literally get to pick and choose how much of their tech you use, limiting any potential risk of lock-in. Not to mention, the hyper-scale providers are probably the most expensive ones!
I think this is all very over-blown. AWS's services are fantastic for scaling up services quickly, and their customer focus is great, but if they were all gone tomorrow we'd still be doing the same things as before, just slower and with less reliable infrastructure.
There are literally hundreds of services offered by AWS and Azure outside of what you are thinking of "rented servers". These things cannot be replaced by "managed hosting". Not that they would be slower and less reliable, these services don't even exist outside the cloud providers.
My job is managing infrastructure, services and products for large corporations in the cloud. It is managed hosting, and SaaS/PaaS, and yes, there are plenty of providers who have alternatives that we regularly use, in addition to being able to build our own with mostly open source tools. It's part of my job to find alternatives so we don't get into a jam when the one solution we use becomes unusable.
The most "cutting edge" things I remember announced at re:Invent 2019 were machine learning and AI stuff, but it was all stuff integrating with existing 3rd-party services and products. They know you want to be able to work with somebody else's tool, and they want to make it easier for you to spend money using that tool on their systems. The bulk of AWS's income is still EC2 (and data transfer); the rest is window dressing to keep you from looking elsewhere.
I worked with GCP on a small (<6 weeks) project last year and the experience left me with a bad taste. Their reps were fairly condescending, questioned my requests regularly, over-engineered what was a fairly straight-forward ask (twice in a week updates, for a 6 week project) where the specs were clearly laid out at the beginning. I then realized that the people who noted the specs and the people who actually had to work on it had not spoken to each other at all so I had to explain the entire ask all over again.
Somewhere down the line, they came back to me to let me know that I had not given them clear instructions and what I had asked was impossible. In the meanwhile, an intern in my team had got the entire solution up-and-running on a small, self-hosted containerized service. I showed them and they mumble about something to be fixed and how they'll get back to me. All in all, a complete waste of time, resources and bandwidth. I am glad it was for just six weeks and in hindsight, I'm happy I didn't get the rest of my team involved in it.
We have worked with Azure since then and I am super impressed with their professionalism. There are hiccups but the understanding I now hold is this - Google (probably still) builds great consumer products but have zero clue on how to work with enterprises while it appears to be the opposite with Microsoft. Their razor-sharp and sole focus on enterprise appears to help them excel instead of having a diffused, unclear strategy that is Google's currently.
I have done a few GCP projects and this is the typical OP of the google team. I think their intentions are good, trying to make sure you get everything perfect, but customers do not really care to spend all the corresponding time updating the google team twice a week for something that would be better served with as-needed support.
And there can be a flavor of condesending attitude when they assume the customer doesn't know anything.
Well this is one way to kill your cloud business. Even a hint that execs were considering or debating this move is enough to kill any serious company's interest in GCP.
Why would you ever consider GCP over AWS/Azure knowing that the Google scythe could kill it at any time, with your suspicions confirmed by a leak this? Bonkers.
It's so stupid that they've got this reputation now. A lot of the things they've killed could've instead been sold/spun off instead. Shutting down successful services that are not successful enough by Google's standards has given them a bad rep at the same time as destroying value.
Reader could've been a successful company on its own for sure, for example. I wasn't a user of it, but there was enough people who loved it enough.
I think google builds stuff in a way where it depends on the google stack. That makes it harder to spin off or sell, as you'd basically have to re-write it.
The crazy part about this is that in the meeting of setting targets one option that came up was folding the cloud unit entirely. And this was already after billions invested.
I don't really believe that they will leave the cloud wars, but still goes to show you the scale and profitability of their ad business and how central it is that even after billions of invested they can still decide to shutter something.
Doesn't exactly instill confidence in your customers to trust you for the service you provide.
All cloud computing put together is less than 10% of total enterprise computings.
Again are they profitable? If the article is to be believed, there own management doesn’t think so and as I posted below, they are distant third at less than 5%.
Reminds me of Ray Noorda[1]. Nothing good ever happens when companies focus entirely on their competitors. If Google really wants to beat Amazon & Microsoft in "Cloud", they should focus entirely on the customer and the value they can provide that customer.
From the Everything Store,
“We are genuinely customer-centric, we are genuinely long-term oriented and we genuinely like to invent. Most companies are not those things. They are focused on the competitor, rather than the customer. They want to work on things that will pay dividends in two or three years, and if they don’t work in two or three years they will move on to something else.”
I interviewed him a few times and I don't think that was strictly true. Perhaps near the end where Windows NT and OS/2 were taking chunks out of Netware
If they pour all of their efforts into customer experience, maybe they can stay in the game. I personally am terrified of using Google Cloud. I'm afraid something will go wrong and I won't be able to talk to a human. I've heard so many stories about the "stone wall" that is Google customer support. Also, I'm afraid the rug will get pulled out from under me if I decide to go all in given their history of axing projects that they don't want to support anymore.
Google sure is nowhere near as capable or even willing as Microsoft or Amazon; heck even IBM through Red Hat could prove a much better steward for its enterprise customers.
But I think the morality is to never put all your eggs in the same basket, because any of them can fail for external reasons — e.g. horror stories about people locked out of any service because payments failed due to some stupid action by their bank; months to fix issues... don't be that guy/business with a single point of failure.
So, contingency plans, and a healthy distribution of your flows.
Have several banks/cards, several clouds, front load balance with a cross-platform HA solution, then for everything behind peg providers against each other; optimize for cost or convenience or whatever.
This is the resilient way; and possibly even antifragile as you develop new synergies in this meta-space. It also sends a message that interop is value to you, to the market.
When you operate at that level, there may be room for GC in your system — typically where you can afford it to fail / disappear but it's better/cheaper as long as it lives. You're already ready for the day it dies, because you make that risk assessment a parameter in your arch.
The thing that scares me the most, is some automated system deciding that I did something wrong without telling me and banning not only my Google Cloud account but also all of my personal google accounts.
Or, alternately, some automated system linking personal and work identities, a misunderstanding occurring on a personal side project of any of our developers (say, a misconfigured email send that's flagged as spammy, or someone's personal blog that's hacked and has malware uploaded)... and our entire startup going offline as a result.
It's bad enough that we use G Suite and that's vulnerable to this. But a cloud lockout, with tooling built for a specific cloud, database restores required, etc. would be a whole different level of devastating. I almost have an obligation to shareholders to steer clear of them just because of that.
GCP has the same support model as AWS. You don't get to enter cases with either free plan, your support is posting on forums and hoping for answers. If you want to pay then you have different levels of access to varying response times.
The "Google has no support folks which can be reached" for GCP isn't applicable here.
We're on Google because we're not betting the farm on them or anyone else. Google has the best Kubernetes hosting, but using Kubernetes allows us to move if anything drastic changes.
This whole idea of basing your entire company’s infrastructure on a non responsive provider is so foreign to me. We have the regular old Business Support Plan with AWS and we can always reach a live person by opening a ticket and starting a live chat. It’s only $100/month minimum or 10% of your monthly spend with discounts. We go through an APN so I doubt we are paying that much even though we are a small company.
It’s interesting seeing the rapid turn-around of opinions on HN. Before everyone said GCP offered the best experience and disregarded Azure. Now people bag on GCP and praise AWS and Azure for their customer/service support.
Having gone through a 6 week long process just to get a working replacement for my Pixel 3, I vowed not to use Google Cloud for fear of a similar opaque customer support experience. I can't fathom the amount of organizational change that would be required to make Google a customer focused organization nor how they could possibly communicate that change to all the people they've burned.
You're confusing Google customer support for their free consumer products (nonexistent) with support for paid services.
Cloud has extensive support, on par with AWS and Azure.
Google isn't a monolithic entity. It doesn't have a single "culture" of being anti-support. Rather, support policies are tied to individual product areas.
I agree, support may be lacking for the Pixel, certainly nothing anywhere near Apple's excellent support. But again, zero to do with Cloud.
It's literally as silly as saying you won't use AWS because Amazon won't give you a refund on a shirt you ordered.
Oh yeah? I remember a story about GCP shutting down an entire large enterprise customers GCP account and turning off all their services over some minor billing thing.
No notice. No warning. No email or call. Just bam - your infrastructure is offline.
Now maybe they've gotten better. But this was only a year ago. You have to at least acknowledge that they are quite frequently not at all on par with organizations known for customer service in this time such as Microsoft.
Part of Google's problem is that Amazon will give you a refund on a shirt you ordered and then have customer support walk you through setting up AWS services to run your website.
Amazon is obsessively focused on keeping customers happy and will bend over backward to do so. Google does not have a similar reputation or history.
Sorry, I left out a key part of my argument. Google scales by automating everything they can, keeping people out of the way. Making customers happy isn't something you can automate, it requires the human touch.
If your company has a widespread perception to have some of the worst support in the industry you have a problem, even if that perception isn't entirely accurate in all cases.
Any inaccurate perceptions users have are Google's fault anyway. They name every product "Google {{service_name}}" and do nothing to differentiate which services have real support and which ones only let you yell into the void.
The whole point of branding is that people expect similar levels of quality for products under the same brand. Google wants the positive effects of using it's name on everything, but leaves it totally up to the end user to figure out which services will deign to let you speak to a real live human when you run into trouble.
We get good support. But recently I think I have heard about big/huge potential customers of us tell us that they are scared of moving because Google suddenly starts ghosting them.
Their support is garbage. Their only answer to getting better support is to get a TAM. Which is completely ridiculous. Plus their SLA isn't worth the paper it is written on. They deny any SLA compensation unless your application is FT over 3 or more AZs.
They deny any SLA compensation unless your application is FT over 3 or more AZs.
For what it's worth, the 3 AZs requirement is their internal policy as well, and they intentionally run exercises where they randomly shut down important looking data centers to verify that failover is working correctly.
If you haven't set up your application to continue working if the plug is pulled on the data center, there is a non-zero chance that it will be shut down due to intentional action.
To be fair, there is also a non-zero chance that it will be shut down due to other people's actions. My two favorite examples are, "Gypsies stole the cables to the data center so that they could sell the metal in it" and "Hunters used the shiny supports for the cables for target practice." These happened in Eastern Europe and Oregon respectively.
On the flip side, their tools make it really easy to have your software run extremely robustly. Even if Gypsies or hunters attempt to ruin your day.
I'm not an Apple dev, but you just reminded me of Apple's excellent support. I got a problem on my development account and immediately received a phone call from a human helping me solving the issue.
As an Xoogler and someone who used GCP professionally I had numerous issues with their support, especially when helping customers launch brand new projects.
Their payed support is terrible and getting worse. We've had no end of issues with being told incorrect advice, being advised on how to administer our systems (when its very clear the problem is with the infrastructure). We've had disks corrupted, which is theory impossible. The list goes on and on an on.
Im also wondering what “best” or “winning” means in this context. Theres no way they’ll surpass AWS in revenue in 3 years so whats the dimension of quality here?
Every company should question the businesses they're in periodically. But having that discussion leak to the public is disastrous. Trust in executive decisions at Google must be extremely low if all their internal business keeps becoming public knowledge.
Ha! A year ago I steered away from GCP for a long-term personal project because I had a small but persistent worry that they'd shut it down, and switching clouds has a cost. I considered myself near-crazy for making the decision on that basis -- they wouldn't do that, would they?! that's silly! that's ridiculous! -- and was a little embarrassed to reveal my rationale to others. Saying it out loud sounded so alarmist, so far-fetched.
And yet here we are. Glad I kept clutching those pearls.
#2 by 2023? Good luck. I wouldn't bet a dollar on them knocking Amazon or Microsoft off their perches by then, and I won't spend another dollar on GCP.
I no longer use any of Google's services and neither does my employer for this reason.
We did have one application running on their cloud, but decided to move it to aws because we just can't trust them to keep anything around. Every year we had to rebuild some part of our app, because they replaced or completely removed some feature we were using. It got old and we were done with it...
This a self fulfilling prophecy. They set a deadline to be a top player, but now that the strategy has leaked, lots of customers will abandon or not take up the service since they who wants to invest in a cloud service that could be abandoned in a few years? The writing is on the wall now.
Wow, this is a real indictment of goog’s plan to take the loot from ads and spend to diversify. They basically invented the cloud model, and if they don’t have the culture and the ability to expand to this most naturally of expansion areas.. they are doomed.
There is something seriously broken at google. This puts Larry quitting in more context. He’s lost confidence in google doing this, then there is nothing left for the org to do. Over—compensated, over-entitled, over-talented engineers that, when put together can’t actually accomplish anything.
Maybe gutting the product org all those years ago was a mistake.
I agree with the general sentiment. It is staggering that they have spent at least 10s of billions if not > 100b trying to diversify away from ads and still do not have meaningful alternative revenue stream. It is one of the biggest mysteries to me. How is it possible to throw so much money and so many smart people at so many different problems and have no results? Microsoft and Amazon don't seem to have this problem but Facebook seems to suffer as well.
That being said, their advertising business is insanely successful. I suspect they could at least double their stock price (and thus the wealth of their investors) if they dumped their speculative bets and focus on ads. It is extremely profitable and revenue keeps growing > 20% from a large base.
Also, I feel like they are in an almost impossible competitive position. Their main competitive tool against aws is further differentiation by ‘moving up the stack” with more and more “do the tough stuff for you” platform services that fewer and fewer people use. It’s getting so baroque with no developer “pull”, people aren’t begging for these services to solve a real problem, more like goog saying “here’s another piece that could be helpful”
Meanwhile, aws’s long history of walking people up the ramp of IAAS makes it more of a fit when they do start adding PAAS service and they fill a real need and get great adoption(for some/a lot of them)
You have to wonder whether "must be number 1 or 2" came out of some pop-management text.
What exactly is wrong with being on the list of potential cloud providers for almost every project? Have you ever been on a non-MS project where someone didn't at least mention GCP instead of AWS?
Google being as profitable as it is can afford options. Why not sit around in number 3 spot?
The article claims, if I understood it correctly, that GCP will not be profitable unless it reaches the scale of being at least #2. The implication is that economies of scale are necessary for profitability.
For example, in 2014, James Hamilton spoke about how every day that year, AWS was adding enough new server capacity to support all of Amazon's global infrastructure when it was a $7B annual revenue enterprise (2004). That scale of build-out, server purchasing, etc. presumably justifies and enables significant cost reductions, such as buying equipment from vendors in bulk with steep discounts, or buying or building large facilities at once, etc. AWS is custom designing many parts of its stack including networking chips and software, ARM-based CPUs, and more; that kind of investment is only worthwhile at a certain large scale. This scale presumably motivated AWS's purchase of chipmaker Annapurna Labs in 2015. Some operator of a few colocation facilities, by comparison, could not justify designing custom silicon since their scale would not allow them to recoup the R&D costs. Before AWS reached whatever scale made these kinds of investments worthwhile, they were purchasing commodity hardware.
To pick another example, a company like SpaceX can only be profitable if it launches a certain minimum number of rockets per year. Without enough rocket launches, and demand from customers, you can't recoup the R&D budget needed to design them.
Google has a head start in cloud because their company's existing facilities are already huge. However, public infrastructure has different demands than internal infrastructure, and it sounds like the necessary investments in public have been costly.
But there's a subtle difference between being big enough to make profits and ranking. If you need X amount of business to make profits, why does it matter whether one or two guys make X+ ?
If you take your SpaceX example, they need to launch some number of rockets. So what if someone else is launching more?
Why isn't the minimum expressed as a % of the market for example?
No CTO who valued their career would go with GCP. “No one ever got fired for choosing AWS/Microsoft”. Even if Azure isn’t held in as high esteem as AWS, businesses trust Microsoft.
We also have Jack Welch to thank for stack ranking (ranking the bottom X% of employees in any given team, no matter how elite they are, as "underperforming").
The self-inflicted damage management fads (and Welch's Ilk in particular) have caused to our economy is incalculable.
These executives have no one to blame for erosion of trust in Google but themselves. Here is a company that appears to be driven by transient market share in their decisions while appointing itself the caretaker of the open internet and the dictator of open standards.
Nothing makes me want to subscribe to a company's offerings more than timelines, or threats it might be eliminated. That is one of the many reasons why GCP has not taken off; outside of Amazon and Microsoft just having much better salesmen/women.
The Google unit, which sells computing services to big companies, is under pressure from top management to pass Microsoft or Amazon—currently first and second, respectively, in cloud market share—or risk losing funding.
This is insane, GCP has been a great experience for me and I have even been working on getting their certs now because I like working with it that much despite them currently being worthless.
Its crazy that they would even consider closing google cloud after gaining so much traction. Especially with BigQuery, Dataflow, Firebase, Kubernetes engine and Cloud Run being at the bleeding edge of where the industry is going. Dialogflow is also hitting at the right time where smart home products and home automation are becoming mainstream.
How do they expect to do in 7 years (launch of compute engine) what Amazon did in 13? Its crazy that they would be so unreasonable.
The analysis from the Gartner Magic Quadrant circa July 2019, states some of it's concerns/cautions with regards to cloud based offerings from Alibaba v Google, which not only highlights how far ahead the top two contenders really are in the race, but also set to continue their meteoric rise.
"Alibaba Cloud earns 90% of its revenue in China and has not appreciably grown its enterprise customer base outside of China. Alibaba has limited capabilities in terms of an MSP ecosystem, third-party enterprise software integration and operational tools; and small global field teams limit adoption outside of China."
"Google demonstrates an immaturity of process and procedures when dealing with enterprise accounts, which can make the company difficult to transact with at times. This can be attributed to its nascent focus on the enterprise market. The immaturity of process is most pronounced in areas such as contract negotiation, discounting, independent software vendor (ISV) licensing, integration with enterprise systems and support."
Google Cloud could have easily dominated the market if they were willing to share their internal technology instead of knock-offs they offered. Now it's too late, competition improved their mediocre offerings and outside TPUs there is no advantage anymore.
From what I have heard, a ton of google internal infrastructure is actually aging and not as nimble by todays standards. It was innovative in 2010, but has been largely commoditized.
My point was that had they initially released cloud based on what they had at the time, AWS would have had a very hard time competing. These days as you mentioned their competitors moved on and even K8S is surpassing Borg in some ways.
It was my impression that the publicly-released versions are better than the originals. It's not that k8s is a knock-off of borg; k8s is a sort of borg 2.0, etc. Am I mistaken?
Yeah, why not play the long game by focusing on providing a great experience for developers? Enable today's hobbyists to start the companies of tomorrow on your platform, and let go of chasing the #1 spot with enterprise customers of today. Cloud computing as a market isn't going anywhere, and demand is only going to grow for the foreseeable future.
Google already has a leg up on this, their cloud offerings and APIs make a lot more sense than AWS.
I think the main reason they lag behind is because the service is perceived as worse. Amazon and Microsoft have well-established customer-development practices and they learn and resolve customers' issues. Having a good accessible support function is a great start to start listening to your clients but somehow in google culture they only want to look at aggregate statistics of a large userbase to learn something that customers are already trying to tell them through non-existing support channels.
For example, I've heard engineers complain about bad instances and intermittent networking issues. Concerns that I have never heard about in AWS except in the very early days of it.
I haven't looked at GCP in a long time, but I remember everything being very python-centric and very lightly documented.
Winning in technology often comes down to having a serious customer-centric product management function.
This whole story seems weird and inconsistent, but particularly: “The pressure to show more progress very soon in the business extends to Google Cloud product teams, which the company is now asking to develop detailed multiyear strategic plans, a former Google Cloud employee said.”
Detailed multi-year strategic planning as a change from 18-month plans seems more consistent with a recognition that significant progress on the obvious indicators is a long slog and that positioning for it is going to take tracking less direct indicators over a longer period than it is with increased pressure for short-term results. Increased pressure for immediate results tends to be accompanied by reduced long-term planning.
You're cattle to Google. Their lack of customer support in literally every one of their offerings shows their approach to making money. They will shut you off in a heartbeat and you have no recourse except to email a bot. Ultimately your puny little shop doesn't matter to their bottom line.
I'm confused; is it realistic for any company to deprecate a product with $8B annual revenue? Compare to AWS, it is small but still significant even for Google. Compare this to YT, which annual revenue estimation is around $15B ~ $25B and CapEx/OpEx is supposed to be very large.
I assume the thinking is that if the ROI is not high enough, then you might as well just invest that money into the more profitable portions of your business.
The problem is that there's no real way to measure and appropriately attribute expense and profit to cloud; most infrastructures are shared across Google products and cloud. Simply put, none of ads, search or YT cannot exist without the cloud infrastructure.
> If the company fails to achieve this goal, some staffers reportedly believe that Alphabet could withdraw from the market completely.
If I was a paranoid person, I might think this article is a sinister plan from Amazon or Microsoft. A quote like this can easily become a self-fulfilling prophecy: managers are already wary about GCP because of Google's flaky reputation, hearing a quote like this could easily make it worse, and the lack of new business makes Google all the more likely to shut it down.
Google Cloud was always a "become number one or die" thing. Same with Google Plus.
Google has assigned so many engineers to building and running the platform. That means the platform has a high ongoing cost to provide. Unless the platform grows rapidly and grabs a significant chunk of the market, it will never be profitable at Googles current level of investment.
You can't scale a project down from 5000 engineers to just 50 and expect it to carry on running as before.
Bringing in an Oracle exec to run it seems pretty ill-advised. Oracle is a services company of the past - GCP needs to be a services company of the future as cliche as that is.
The tech is there - but alignment isnt. I dont know if Google is capable of anything outside of search.
I would be soooo pissed if this happened. So much for marrying GCP. AWS and Azure are so much worse though. The price of good software is clearly very high.
Time and money aren't Google's scarce resources. Engineers are. Every Google project would love nothing more than to steal other Google project's staffing for themselves.
I guess google values intelligent people, and since intelligence involves
- thinking outside the box
- considering consequences
- estimating impacts of 2nd and 3rd order
this should lead to the realization this may not be a good idea: why didn't the managers (which quite probably have been engineers before becoming managers) consider that?
If there was a betting marketplace for this, I would bet all my money against Google. Three years is nowhere near enough time for Google to catch up when the current situation is that they're falling more behind every day.
They also have no way to fix the customer service issues plaguing every product of theirs, including their cloud stuff. So, so many stories about how bad they are.
This feels like a threat. Are they going to graveyard Google cloud? Probably not. But they've discontinued so many random, good services that I'm cautious to 'learn' a new Google tool. I wouldn't want to invest in a product only to be forced to transition to AWS in 3 years. Might as well just skip to AWS.
Even prior to this supposed leak, it would be really difficult as a developer to build anything reliant on Google's technology stack. There is a huge number of killed projects; abandoned platforms; and projects like maps where low/ no fees have become prohibitively expensive.
Building atop Google is betting the farm on a fickle master.
Well Amazon isn't exactly the king of great UI either and I'm not in love with Azure's cloud services platform management tools. All of these platforms have an extremely dense set of options which tend to change fast which leads to a very tough UI challenge and rapid UI rot.
These statements just made me remember a quote from Will ferrels movie.
Ricky Bobby : Wait, Dad. Don't you remember the time you told me "If you ain't first, you're last"?
Reese Bobby : Huh? What are you talking about, Son?
Ricky Bobby : That day at school.
Reese Bobby : Oh hell, Son, I was high that day. That doesn't make any sense at all, you can be second, third, fourth... hell you can even be fifth.
Ricky Bobby : What? I've lived my whole life by that!
GCP is easy to use when compared to AWS according to me. I have used both the service providers. Starting with Firebase, Firebase hosting, App Engine, Compute Engine, GKE, BigQuery and monitoring via stackdriver. Everything just fits very well.
Number of people visiting their yearly Sessions are increasing and their training platforms(Qwiklabs & YouTube videos) are very much precise. Most of the issues are solved in Stackoverflow itself and you have Google groups for each products.
I'm just flabbergasted this would even leak out. It will be a self fulfilling proficiency as prospective clients will shy away from Google. They must be doing damage control right now with all their latest enterprise clients who switched over. May as well start the clock on the Stadia death march. I give it 13 months.
The purpose of Fiber was to push Comcast, ATT, and other incumbents to start installing fiber internet, which they have.
Same as how Fi is focused on pushing cell phone providers to lower data costs and provide pay as you go, which they have.
Cloud is selling extra server space in the same data centers that support Youtube.
Google Fiber’s attempt to roll out its gigabit internet across the city of Louisville, Kentucky has apparently failed so spectacularly that the company has decided to completely shut down
The Information is a fantastic source, and I subscribed to support the excellent original reporting that comes out of it. Most of the news will get disseminated by other blogs and news journals in the next day or so, but The Information is pretty good if you want tech news first.
A big nugget from behind the paywall: "This person said the group’s leaders didn’t explicitly state what would happen to the cloud division if it didn’t reach a top two position by 2023. A commonly held view inside the group was that Google wouldn’t continue investing money if it failed to it meet its goal, the person said."
The article also claims Google Cloud isn't profitable under their current business plan: "The group’s leaders told staffers that if Google couldn’t reach a certain size with its computing and storage business—two of the most commonly used cloud services—the cloud unit might never become profitable, the person said. To reach such scale, they said, Google would need to be in a top two position in the market."
Curious about what others think of the price. $399/year is pretty steep. Compared to, say Ben Thompson who I do subscribe to for $120/year. For younger (<=30) folks, there's a $199/year discounted price.
It's super steep. I think that their target market is businesses and people who can afford it. But I jumped on it when they offered the young professionals plan. Really just to support them over the exclusive access side of things.
I signed up to the information for the same reason. Plus they actually have their paywall well designed. I get most of my news from them via email and I can read it right in my phone's gmail, versus everything else that has me sign up somewhere and then re-login every time. I wouldn't mind paying if the experience was at least seamless post payment.
Another way of looking at this. If you are shopping for a cloud platform Google is a really hard sell. On one hand you have Amazon who really paved the way here and literally built their company atop their platform and currently powers a huge piece of the web. On the other hand you have Microsoft who has decades of experience figuring out how to appeal to CEOs and enterprise customers and can dangle the prospects of integration with existing corporate infrastructure.
Google offers the Google name and a reputation for ditching services when the wind shifts. They really have little or no competitive advantage here. I'd personally rather build atop Linode or Digital Ocean than Google or at most, stick to the middle of the road, most generic parts of Google's platform.
Did you miss that the complete headline is actually 'Google execs reportedly debated getting out of cloud computing, but instead set a goal of being a top-two player by 2023'?
The thread is mostly concentrating on the negatives here. I'm hoping that GCP will offer some incentives for the next 3 years that I could take advantage of to help my side gig grow on the cheap! If they die in 3 years, well I don't rally care as I have no clue if I'll have a business in the next 12 month.
Of course, this article does say "The Google unit, which sells computing services to big companies" with emphasis on "big", but maybe there will be some trickle down for the small fry amongst it all.
There's no way for Google Cloud to achieve this. Even if Google does everything perfectly, both AWS and Azure are just too entrenched to be dethroned in such a short period of time.
Therefore, this story is almost equivalent to saying 'Google will ignore cloud from 2023 onward'. Were I ever to consider GCP, this alone would cause me to run far far away.
If Google is serious at all about cloud, it would issue a quick denial. If it does not, well, it would be evidence in favour of the story's credibility.
"Last quarter, nearly all of the company’s $40.5 billion in revenue came from advertising. Meanwhile, a collection of newer businesses that Alphabet lumps together as “other bets” in its financial statements, including the self-driving car unit Waymo, generated a mere $155 million in revenue while incurring operating losses of nearly $1 billion."
"Google other revenues consist primarily of revenues from: Apps, in-app purchases, and digital content in the Google Play store; Google Cloud offerings; Hardware; and YouTube subscriptions."
"Other revenues" is up nearly 40% q/q compared to Q3 2018:
"Our Google other revenues increased $1,788 million and $4,639 million from the three and nine months ended September 30, 2018 to the three and nine months ended September 30, 2019, respectively. The growth was primarily driven by revenues from Google Cloud offerings as well as revenues from Google Play, largely relating to in-app purchases (revenues which we recognize net of payout to developers)."
If a company like Westfield decides to get out of the leasing business, it makes sure its anchor tenants know there is some continuity of lease for their locations, because the value inherent in the property is the lease stability of the anchor tenant.
I begin to wonder if cloud has a parallel here: Does Google actually need, for selfish reasons, to inform significant cloud tenants in GCP that continuity of their business is core, and will be factored into any business development, or retiral, or sale options?
Elastic Search comes to mind: If you build to a dependency on Elastic, in GCP, its comforting to know it works in AWS but they are not an entirely identical product. From a business continuity model, mostly the same. Different tokens, different backend underneat the ES componentry. Does it matter? Depends how critical Clous ES is to your business.
The current issue with many cloud providers today is that they had designed all these fantastic features but then completely left it up to the customers to figure out new use cases on their own. They forgot that the majority of customers don't have much clue about the cloud let alone those advanced features, and cloud providers are sitting here complaining that they are running out of patience waiting for customers to get on board. Designing new features alone is not enough, they must also design the actual use cases to convince customers because most of them don't have any clue of how to actually turn those advanced cloud services into something more useful than just hosting. It's not a capability issue but rather it's a use case problem.
Google* isnt really an enterprise focused company. Getting talent and management to work on it must be tough, and internally I've heard it is sort of looked down on. They have a pretty poor partner network as well. I think if they beefed that up theyd do well.
Cloud is not about developers anymore: is not about tech. Why? Because enterprises understand that the foundation for their apps exists, if not their Cloud reps will listen and develop it for them as long as you are big big contract. Now the power relies all along in decision makers, Google reputation is not good there. Getting Fortune 500, Government contracts is very very very unlikely as in reality from Cx0 perspective...AWS/Microsoft have been there always...new things in Infrastructure...probably not unless you really don't want to give any business to AWS (Walmart/Target) and experimenting the new enterprise Google...no thanks
I have thought at times that HN was a place of rational, deep discussion. I cant understand where these types of endless emotional rantings come from. Everyone is going with a bigger hyperbole how "Google is killing GCP, they r so dum".
I can understand their ambitious plans, and naturally you decrease funding for a project if it keeps hemorrhaging money. It says the Google C-suite discussed the idea of leaving the market, as any rational person would do, but dismissed it. I think the way article describes these things is putting a whole of color on this.
But so what, Google thinks it can beat AWS and Azure? Well, it's good they have big goals. Such a heavy bet on their cloud business seems quite sensible, and while they mention cutting spending if they don't achieve those goals, I don't see it as a big of problem as people seem to think. I mean other cloud providers don't even have that amount of money to put into their cloud businesses' and no one is lamenting about that.
To me this whole debate should be about how impractical and out of touch with customers Google has become since.. I can't even recall. Maybe it's the algorithm-based engineer interviews but something's wrong with Google's internal workings. Google Analytics' dashboard is still terrible, I don't like even how Google Docs work, Youtube is I guess ok but they blundered the live-streaming et cetera. I've never used GCP but it always seemed a bit too over-designed and flashy and, well, not practical. On the other hand AWS's console, even though it is at times unintuitive it feels for some reason quite nice and practical. It's weird and subjective, I know.
Could it be that when I had my first touch with AWS I had been given a quite positive view of the whole service vs. I have quite negative prior feelings about other Google services that cloud my judgement. Who knows.
Yet still, perhaps Googlers at their fancy Mountain View office have become a bit too detached from the regular Joes working for their gray corporate masters who want nothing to do with their hipster dashboards and just want something simple that works. Maybe pride comes before a fall, like with Microsoft. Luckily for MSFT Nadella started to change their culture.
As someone who literally have a decision to make tomorrow for a large move from Azure/Hosted to consolidate GCP I may have to send a few email to postpone that meeting... This is really bad for GCP
Why would they even care about ‘winning’ the cloud? To what end? Google is officially and advertising company, and the large majority of their revenue depends on it.
This strategy is working so well for them, I just don’t see why they would even care about grinding out a little extra market share in the cloud hosting space. Why? Maybe just to boost their ego?
Is it really worth it to them? Whatever comparatively small amount of extra revenue they squeeze out will still be dwarfed by the ad revenue. I imagine they will continue doing what they do.
One of the things that google does is basically trade compute and storage (whether its a google search or gmail storage) for advertising revenue.
In a world where the public cloud is dominated by a handful of providers its likely (if not already the case) that AWS will become by far the biggest buyer (with the best economics) of compute and storage.
Obviously there are many other barriers to competing with google, but if amazon could provision the amount of storage and compute needed for YouTube and search in the headroom of its public cloud (and even worse if AWS customers could do this for a cheaper cost than google itself) this would be a negative thing for google.
They would have to bring something far more compelling compared to AWS if they are to gain much market share. Either in terms features or notable cost difference. Technical people with a general mistrust of them go into planning with GCP as a hesitant second or third choice. The traditional IT departments moving to Azure aren't going to consider the other vendors much when there's such a direct migration. Combined with headcount and capital cost reduction so clearly mapped out for them.
Have they considered making their pricing easier to understand, or did they collectively decide that the AWS model of "Good luck understanding your bill!" is a good one?
Maybe they could consider applying more reasonable pricing?
E.g. Hetzner offers an 8-core 128GB Xeon SkyLake machine with unlimited 1gbps traffic for around $120/month while Google offers n1-highmem-8 with 8 cores and 52GB and no traffic for $240/month.
1gbps unlimited traffic on Hetzner means 324 TB/month, which would cost 26000$ on GCP at the lowest 0.08$/GB rate.
So I would consider cutting the instance prices by half, and dividing the bandwidth prices by 1000, and then they will be competitive.
GCP TPU options beat the pants off of everyone. I love bigquery.
Azure has by far the best storage options. The bucket storage is much cheaper. The azure file storage is best of class. I don't like it's bigquery/athena alternatives
AWS is the incumbent, I love the spot market, but more and more we have just been running low end workloads there. Athena is the best of class of analytics tools if your data is already in a bucket.
They already are making the same mistakes as other vendors. They think the managed openstack/k8s services are something that big people use. The reality is tools like kops or aks-engine support a much more stable cluster.
Wow, this seriously highlights Google's lack of understanding regarding enterprise customers. Enterprise purchase decisions I've been a part of place stability and the long term partnership way above cutting edge tech or price. It takes about 2 seconds to extrapolate all of the internal investments you'd have to scrap in 4 years if Google decides to defund GCP.
Only way Google perhaps can take a lead in Cloud is if they invent a new paradigm - quantum cloud or something (just to illustrate the point). That is, change the game and not change themselves for the current game.
Google just can't compete with hustlers like Amazon, Alibaba and to some extent Microsoft (ex-Oracle leader not withstanding). Cloud business is commodity operational business.
Just spitballing: what if they rented compute based on sandboxed access to the huge trove of behavioral and activity data they have on their users (which amounts to everyone who uses the web)?
It would be the ultimate differentiator. The only potential competitor would be Facebook, who doesn’t offer cloud services.
The United States has no data protection laws to speak of (over this kind of data). Renting/selling access to run queries over this sort of data is entirely legal. All it
would take is another half dozen years of our society continuing to move in the direction it has already on privacy.
It may just be de jure legal though; I bet it would be shown to be de facto illegal once people started mining data from federal judges and legislators...
Has this become just a commodity war, or are there major sustainably exclusive technology/IP advantages that define the competition (e.g. processor, hypervisor, database, CDN, OS license)? It seems like third-party suppliers are chipping away at leads in each of these. Is UX or lock-in a significant differentiatior? What is it that defines this business?
I'm doubtful if cloud would be as profitable business in 2023 as it is now. Race to the bottom is already in place and only bound to accelerate further. This is like Microsoft determines to win the phone market in 2015 when iPhone peaked up the steam. By 2015, the smartphone market is changed dramatically.
Knowing that there is a risk of being de-platformed and loosing all your client data, and not being able to offer your service, etc, is a deterrent for most people, Google has a very difficult job here to change and then communicate the way they will deal with users will not be like gmail or YouTube.
This reminds me of Google+. I remember one year where they got very serious about G+, tied bonuses to the growth, and otherwise tried to take on Facebook. G+ is shuttered now, but at least I have a cool green check mark next to my name I'm my YouTube comments... :-)
Well, it's pretty obvious Google wont be able to beat either AWS or Azure in the next few years... How can information like this leak or even be entertained internally? Google is famous for terminating, even successful products. Best to stay far away from GCP!
People want to develop their tech on an open platform without vendor lock-in. If they pushed an open source platform on top of Kubernetes I know I would be more interested and comfortable using them. Right now I just don't see a strong benefit over others.
i run a bunch of services for some online social clubs and foss projects that i'm involved in. i've floated between a bunch of different service providers (Digital Ocean, Scaleway, Hetzner, Joyent Public Cloud, Vultr, and Linode to name a few) and now i'm working on setting up some colo infrastructure. i'll probably keep some of the cloud stuff around for failover and maybe some edge cache type stuff, but you don't really need cloud unless you have extremely variable loads on your system.
i won't say cloud isn't nice, i've been able to learn a lot by playing around with cloud servers. just good to remember that cloud isn't the only option for a lot of workloads.
Besides the notion of Google not investing in cloud seems rather ludicrous to me. They might as well formally announce that they're not competing in tech going forward & just count adtech dollars instead.
It would be funny if Google started using AWS instead of their own datacenters. While the compute capacity per space unit have kept growing exponentially, growth has stagnated as they already got 100% market reach.
I know you said that in jest, but in all seriousness, I think search and ads will definitely be around since they are the #1 player and making boatloads of money from search and ads.
Google seems to have a tendency to shut down products that meet the following criteria:
[1] Free products that could have millions or billions of users, but no revenue or not enough revenue to justify a team working on it
[2] Free or paid products in which Google is NOT #1
I am as shocked as everyone else that Google would consider shutting down GCP if it's not #2 by 2023. I would have thought they would treat enterprise business very differently from their consumer business ... but I guess not.
I am also extremely skeptical they'll be able to make it to #2 over the next 4 years. So I guess we might as well say goodbye to GCP at this point.
A business like Google Cloud would not be wound down, it would be sold. MS and Amazon woulc not be allowed to buy it for antitrust issues, so potential buyers might include Rackspace and Apple, among others.
Sorry Google, you need a time machine to fix this one.
No one trusts your products to stick around or has the confidence that you won't shut down their accounts overnight due to an automation thinking something is wrong.
I'll wait for more information before ditching Google Cloud, but this will certainly speed up getting a fall-back/migration plan together for what we've got on the service.
So this is a PR disaster for GCP. Already people are penciling out contingency plans because this inept leak. And Google's history of shuttering stuff really doesn't help.
Also they just hired a new guy to head up GCP from Oracle, I doubt someone like that would join if there was even a hint of there not being 100% commitment to GCP in the air.
Dream on, Goog. You're third in a three-horse race. You need a serious, and I mean serious value add to beat Amazon in a LOB where they bring their A game.
Having extensively used AWS, and having spent a considerable amount of time with GCP, I think AWS is the superior product for the vast majority of enterprises.
I’m a fan of BigQuery and by extension GCP. I find it simpler to use than AWS but AWS has more to offer I think. It’s good to have competition in the space.
GOOGLE: we want to to be in the top 2 or we will throw our toys and leave / defund.
EVERYONE ELSE BITTEN BY GOOGLE DISCONTINUING STUFF: errrr, why would I EVER choose you guys to host my stuff NOW unless you are ridiculously cheap and it would be super simple to switch to azure/aws.
heck if I was using them now, and saw this, I'd be making plans/contingencies to shift off them.
GCP generates $8B annually. It's a bit dramatic to think Google will give up on GCP. AWS has a loyal customer base, which is well deserved, but Google does have some unique products (like Spanner) that AWS may never have. Point being there will always be a place in the market for GCP.
IIRC Google killed off a lot of products that were profitable. They just decided it was not profitable enough by their own arbitrary metrics, not that it wasn’t profitable at all
I have sometimes wondered why they don't offer a separate, uncomplicated fixed price (and bounded resource) version - basically like a Heroku - and provide a migration path to their core offerings.
AppEngine is priced very high and is definitely not a constant price. I'm using it right now, and it costs me $5/day to run an Elixir/Phoenix app for ~2000 active users. I'm leaving as soon as my free credit runs out.
I still lay claim that google is not hiring the talent that they think they are or claim to be. With such huge budgets and failed product after failed product one has to wonder what is the genesis of their failings. Having met countless arrogant but mediocre engineers that leave Google after four years, I will bring up the old algorithms only hiring nets bad employees with good memories.
I do think their engineers have likely slipped in quality over the years, but that's not the problem. The problem is their culture of product managers being rewarded for launching projects and nobody being rewarded for maintaining projects. They financially incentivize people to launch and kill projects.
It sounds like a leadership problem, because only leadership can change the culture like that.
Probably a lot to do with this. Incentives are probably the strongest driver of behavior in a company, much stronger than any stated strategy.
If people are richly rewarded for launching new products, they'll optimize their work for launching new products, even when that product doesn't make sense, or when more business value could be created by supporting/enhancing an existing product.
For another example, Microsoft's dysfunction and infighting in the Ballmer era can be tied to stack ranking. If your success depends on your peers doing worse than you, you're incentivized to not cooperate at best, snipe and sabotage at worst.
Google absolutely has to change their internal incentives if they want to sustain success into the future. In the case of GCP they're gonna have to start figuring out how they can shift their culture to wooing and keeping enterprise customers. This leak, their culture, and their history is gonna make that a tall order.
That, and the absolute 100% lack of customer support or engagement.
Unless you are a Fortune 100, they don't do support.
Want to move an SMB with 50+ employees to Chromebooks with Chrome Enterprise? You literally cannot pay Google to support you. You have to work with one of their third-party vendors, who all suck.
Perhaps with Brin and Page finally leaving, this is something that can change. But, I doubt it will, since they hired the leadership that's implemented their vision. The next chapter of Google is likely to be how Sundar Pichai leads them to the brink.
This is also true, but Google is a very engineering centric company, who they hire with regards to software end up running other parts of their business. Lots of really powerful people there came in and worked as engineers.
> I still lay claim that google is not hiring the talent that they think they are or claim to be.
Obviously I am biased. I don't think I'm particularly great or anything, but I find my colleagues to be by far the most capable I've ever worked with, and I did previously work at AWS, where engineers were also very smart.
> With such huge budgets and failed product after failed product one has to wonder what is the genesis of their failings.
Personally I think it's an issue with management structure rather than individuals. Frankly I don't think smart people matter all that much.. for the most part you only need sufficiently smart people to do their job, and good culture/management will make the team shine. I'd argue that engineers are individually significantly less important than product managers and managers.
That being said, I think the issue is that Google is just too engineering focused. This makes it a wonderful place to work, but I don't think it leads to great products most of the time. It's fine for products like search and ads, where you can literally measure if you are making improvements, but not so good for most products.
Engineers are more or less allowed to pick what they want to work on. Again, I enjoy this flexibility, but I think it also leads to a lack of focus, especially when product managers probably have a much better understanding of what customers want or need.
Of course, Google is a massive company so I don't know how reflective my own experience is of the larger picture! This is just my personal feeling.
It's a little strange to me that otherwise experienced people feel the need to "study" for Google interviews.
To me, this seems less like they're selecting for good engineers and more like they're just selecting for people who really want Google on their resume.
If they're going aggressively for market share this could only mean one thing: price war. They've been running (and cost-optimizing) cloud services for much longer than Amazon or Microsoft, so they could actually make it _really_ painful for the other two players, and they know it. Most of all, they could undercut them on egress pricing. I'm all for it. The current pricing in all three clouds is nuts.
One of the great downsides of Google having such a massive hit with Search is that they don’t have the kind of willpower that it takes to absorb defeat after defeat on the path to victory. Look at how many failed fucking pen devices Microsoft has shipped, since Windows CE. They don’t cut funding, they just keep going and going and going. Google doesn’t do this.
"Google brass" are basically Alec Baldwin's soulless asshole character from GLENGARRY GLEN ROSS:
"We're adding a little something to this month's sales contest. As you all know, first prize is a Cadillac Eldorado. Anybody wanna see second prize? Second prize's a set of steak knives. Third prize is you're fired."
I actually don't think that's what happened at all. Best I can tell, this was a multi-year, aggressive, OKR-like goal. You're not expected to 100% achieve your OKRs at Google. 100% on everything indicates your OKRs are not aggressive enough and you're sandbagging. You're merely saying: "I think this is worth doing and I'll commit serious effort to achieving this". I don't know why people are concluding that if world domination doesn't happen by 2023 Google will set its cloud datacenters on fire.
Or, to be perfectly honest, I do know: it's to drive outrage and clicks.
Google is not going shutter any business which brings in more than $1B in profit per year. It never has, and it never will.
Disclosure: ex-Googler, GCP user, hold no position in company stock.
It really depends on how you split the numbers. It's kind of like saying SQL and Excel are the most popular programming languages... yeah, but they've captured a far different market than most people would consider to be the "programming" field.
A service that only exists in China to serve the Chinese audience may be bigger than worldwide services that serve worldwide audiences, but it's kind of irrelevant if you're not a Chinese company doing business exclusively in China.
From what I have seen Alibaba cloud is #3. Don't forget that China is a huge market, and it is very hard for Western companies to be successful there. Alibaba has China government backing and that relationship is allowing them a lot of access in China
For example, AWS China and Azure Operated by 21Vianet (official name of Azure China) are operated independently by Chinese companies, and have incomplete integrations with the rest of their services, because of how hard Chinese govt compliance is. And GCP only has a location in Hong Kong, which means Chinese users can mostly access it but it's unreliable https://gitlab.com/gitlab-com/migration/issues/649
Different articles and sources give different numbers, so don't take this as gospel, but:
IBM's *-as-a-service ARR is larger than Google Cloud's.
Oracle claims theirs is too, but my understanding is that they roll a lot of their licensing costs into the same reporting, so it's hard to tell for sure.
For my personal projects, GCP is definitely my preference although I also like AWS and enjoyed using Azure while I had a Microsoft BizSpark grant. GCP is very easy to use.
For huge customers GCP provides good support, but I think that the GCP team really have to work harder to support small companies (I base this on HN comments, no personal experience).
The engineers from Amazon, Google and Microsoft should get together and make sure they're all using different metrics for success, so they can all be first at the same time. Or work out a schedule for who will be first which year.
I have a hard time believing articles from sources with no track record making bold claims. Seriously, "theinformation.com" and hardly anyone is skeptical?
I do not understand the hyperbole around this report. Nowhere does it say that Google plans to shutter GCP if it doesn't reach #1 by 2023.
> said people with knowledge of the matter.
As in, people who may not have even been at said meeting which took place nearly two years ago, under a different VP and now a different CEO, and saw some meeting notes.
> The group’s leaders told staffers that if Google couldn’t reach a certain size with its computing and storage business—two of the most commonly used cloud services—the cloud unit might never become profitable, the person said. To reach such scale, they said, Google would need to be in a top two position in the market.
So it's not about being #1, it's about being profitable.
In other words, this was a conversation about literally every product at an executive level ever.
You set goals, you decide what happens when those goals are met. Furthermore, there is no concrete assertion here that what happens if this deadline is passed. "at risk of losing funding" can literally mean anything from changing FTE allocation to capital expenses to a million other things that get discussed at an executive level.
And finally:
> A Google spokesperson declined to comment prior to the publication of this story, but after it appeared released the following statement: "Reports of these conversations from 2018 are simply not accurate."
There may be some truth to these accounts--reality is often between the lines--but this is extremely soft on details and actual first person accounts.