As someone who has been responsible for large cloud budgets, I wonder if Google is aware that their "support strategy" toward both content creators/contractors and customers across the entire brand has completely ruled out GCP for me.
GCP might have great support. They may not have automated bots suspending accounts. They might actually follow the DMCA counterclaim process instead of some weird process on top of it.
I'll never know: I restrict Google products to casual, throw-away contexts. I've imagined myself trying to explain to my boss or a client "They won't respond" or "They say they can't provide any information." Just typing that freaks me out.
Anyway, if you want to be a huge corporation living under one brand umbrella, you've got to take the good with the bad. Google is for entertainment stuff and toys to me.
I was involved a bidding war between GCP and AWS where each was offering a healthy six figures in free credits to move our business to their cloud. GCP was offering nearly 3x as many but ultimately I didn’t want to move to a service where there was a chance offerings were going to get the axe with no warning. Their reputation has indeed hurt them — at least in one instance.
A fair percentage of companies exploring my SaaS company as potential clients are doing so because Google doesn't provide any support / insight for practically any issue they may have.
As someone working for one of its competitor, I'm glad that waking up at night and solving a customer's problem (and/or our problem) is helping in a tangible way.
I work for a company which helps huge enterprises move to the cloud, and we found a lot of places opting for Azure because their experience with GCP support was really bad. These are people paying for the top tier of support too.
I was part of a focus group a year ago. The general consensus was Google might be great, but everyone felt better about using AWS for real world projects. MS Azure was a close second.
Not to heap it on with anecdotes here, but I am also part of this group. I was working with the platform team as part of a marketing tech org that sees half the us pop in monthly visits. We were deciding between signing a long term deal with AWS or GCP. We wanted to use k8s at the time too and even realized that GCP was far superior in this space versus ECS / EKS. However support, and product lifespan was the deciding factors.
Similar factors in deciding between Office 365 or GSuite. Different products, same Google story.
I've worked at a couple "mid-sized" financial companies (i.e., not "tech" companies, but with a healthy need for various computer products and services) throughout my career where migration plans from one technology or provider to another were measured in years, and would sometimes get extended to the order of decades for larger migrations (e.g. moving core business functionality off a legacy mainframe system). A year or two notice that you'll need to migrate something important is really not feasible for a lot of businesses who would otherwise be very interested in outsourcing to a cloud provider.
A year is an eternity for a start up. But it is a really short time to a large company with complex integrations, a massive book of work and budgets being trimmed every week.
Google has a history of either cutting products (even successful ones) with poor offerings of migration (i.e. this from 2 days ago: https://news.ycombinator.com/item?id=24548876 where they aren't even allowing bulk export of accounts), or cancelling your entire google account for issues in a single product with only automated systems there to respond to you (i.e. https://news.ycombinator.com/item?id=4013799)
A year isn't much for a non-startup business. Google has also increased prices in the past (maps, kubernetes) while AWS has afaik never increased prices on anything.
I mean AWS does the same thing - mine crypto using credits to speedrun getting your account terminated. The only difference with GCP is that they have a reputation for this.
They likely meant how google axes whole products like aws with no recourse for customers. Amazon probably would not do that to a product like aws in a million years.
Regular credits as well. I'm speaking more about when an attacker mines monero on a compromised server, and I wasn't talking about products being axed (but I now realize the OP was).
Ex-GCP here. Unless you shell out at least $15,000 / mo [1] for a Technical Account Manager (TAM) you are a nobody to Google Cloud. Hence all the bots suspending accounts and scripted processes. If you are a small business, bite the bullet and use AWS.
We have a TAM, but it doesn't buy us much. They use their TAMs as a sales lever to get you to use all of the vendor-lock in features. They'll help you architect your platform so that it won't run anywhere else and charge you for the privilege. You can keep telling them all you want about how you operate multi-cloud and how you won't budge on that and they'll just keep trying to run higher up your flag pole.
For actual issues needing real support, like when anything breaks, they're useless.
We have a TAM as well and they've not show any value. In fact, every issue we've had we've had to escalate through non traditional means (IE personal relationships). Google absolutely does not care.
We've ultimately decided that we're moving out of Google strictly to AZURE / AWS, and ironically enough Oracle.
The job title “account manager” says it all. That is a 100% sales role. So don’t expect deep technical expertise. Any assistance rendered will be to grow the vendor footprint.
That matches my experience actually having worked on the GCP support team. The TAMs just deferred anything technical to the support team... unnecessarily escalating things and being pushy even though we were already investigating the support request with highest priority.
AWS is run like a gigantic social experiment in "how half-assed and crusty can we make it and still have people pay?"
The reason we put up with them is because A. they're the new IBM, nobody gets fired for choosing AWS, B. you can reliably reach a human who will at least give you a straight answer when you start to suspect that the AI-powered auto-scaling is actually marketing fluff sprayed over a double/halve cron job that runs at the top of the hour.
>B. you can reliably reach a human who will at least give you a straight answer when you start to suspect that the AI-powered auto-scaling is actually marketing fluff sprayed over a double/halve cron job that runs at the top of the hour.
You say this as if it's an afterthought as opposed to an incredibly important and massive advantage AWS has over GCP.
^ My impression as well, after spending nearly $200K of my clients' money on AWS. I use GCP for my own work (+ a bunch of on-prem hardware) FWIW, but "the new IBM" phenomenon and "half assedness" is very real and palpable in the case of AWS.
Off-topic, but this might be a good opportunity to address this:
Do people find "?" as being aggressive? I think questions generally, tend to be aggressive (which is why deflecting back at your opponent is common in online arguments), and I'm having a hard time seeing how a question mark does more than just attempt to annoy the other person.
I'm curious what others think, I may be too sensitive.
Yep, that's how I intended it, effectively the same as "What?" with the implied offense level being no larger or smaller than the offense inherent in expressing surprise at someone else's opinion. That level is not zero, but it falls well within the bounds of civil discourse, especially given the lack of elaboration in the original position and the fact that I invested seventeen times its length on explaining my own -- only to not receive a followup from the original poster.
Yeah, next time I'll just type out the question. However, given that it's starting to look like OP ghosted the conversation, I'm increasingly comfortable with the originally-unintended rude undertones of "?"
Well it usually has been when I've personally used it.
There is the disdain of not explaining exactly what you are questioning with the implicit assumption you are questioning everything said. It's pretty much a stand-in for "WTF are you on about? I don't even know how to ask a sensible question about what you said." when I've used it.
In this case though? It's a single statement so maybe a bit abrupt but I read it as a request for clarification. A fairly strong statement was made (essentially "don't use this company") with no backup or explanation. So an equal lack of effort was made proportionally in response.
This is great for AWS compared to GCP, but what about Azure? If anyone was "the new IBM" I would have expected it to be Microsoft (but I agree, it's not).
Only if you forget to factor potential support problems into your economic analysis. What would your systems being down for a week because Google capriciously shut you down cost you?
True, but due to the large extent of the services provided/required, it's very hard to know the unexpected costs, even if one has experience.
For example, one will hardly think of the cost of the disk speed (IOPS, in AWS), before moving to AWS. Then, they will suddenly have to deal with it (note that IOPS will be mostly opportunity costs, in case one doesn't choose provisioned IOPS of larger capacity).
The cheapest provider in existence runs on renewables (Hetzner Online). There isn't too much of a difference, especially for EU datacenters as they have a high excess of renewables and the energy market works in a way where you pay for renewables but just as everyone else you'll be using the base capacity generating dirty sources.
> One TAM unit provides on
average one business day’s
worth of effort per week
Just to have an ear on deck about Google's issues?
It sounds like the platform is bad enough that we're in the ballpark of Bald Tony the TAM coming in with a bat and saying "nice place ya got here in this rough neighborhood... pity if something bad happened to it!"
I had just this scenario yesterday where Google ran out of capacity of NVMe-based N1 hosts in a zone. I was forced to upgrade a large number of nodes to N2 hosts, which cost more (despite their claims about them being cheaper -- the sustained use discounts are lower) and are still in Beta.
I just wanted info before I started about whether we should expect to be able to even build those hosts that day and they wouldn't give us _anything_. We're not exactly a small company and the amount we're spending with them is large.
Their answer was basically either reserve capacity or screw you. I had to relay that to my CTO.
Google tried with GAE, which is much closer to the Borg model (“I don't care about VMs, just run this code somewhere for me, and make it scale, make it automatically have access to a database”), but the world wasn't ready for it.
The world wants VMs and virtual networks and firewalls and clickable web interfaces and to run big legacy enterprise apps that MUST NEVER DIE and to fuck around with Terraform and to SSH onto boxes. This sucks, though, and Google SREs know better than to use such terrible abstractions unless forced to. The internal software development ecosystem is eons ahead of anything you can get with any cloud provider, but that sort of alien technology doesn't sell.
Kubernetes/GKE is a good compromise between something that Google sells but Google also wants to use. That, and some of the GCP versions of internal tools (Spanner, BigTable, BigQuery, ...).
I'm an independent consultant and have also been involved in large public and hybrid cloud budget decisions between the 3 big players. GCP will never get my backing as long as Youtube continues on it's current path. That is purely from a Google/Alphabet ethics point of view. Don't get me wrong, the other two players are not certainly not perfect either, but they're a farcry better than GCP in my opinion.
Yeah, Amazon aren't exactly good guys either - there's too much to mention really, they're at least as bad as Alphabet.
Alphabet doesn't to my knowledge use poverty wages combined with prison like surveillance and massive union busting to prevent employees to be treated as thrash.
If you're choosing cloud service providers on the basis of corporate ethics, Azure is your best choice at the moment.
The optimists would say that's a testament to how much Msft's behavior has improved over the past 30 years; the pessimists would say its an indictment of just how bad Amzn and Goog really are.
The company I work for is currently waiting for a reply from google related to an increase in quotas for google sign-in.
The feature has not worked for more than 8 months now, customers complaining and all. We have done everything on our hands (all documentation sent, demo app showing the feature, video explaining all usages, etc..) and it's as good as shouting to the abyss.
We were recently evaluating alternative cloud providers (our main one is AWS) and we discarded GCP because of this.
GCP has some nice data and ML tooling I'd like to try but the potential risk just isn't worth it. Inefficiencies and extra costs that are constant I can adjust, budget for and deal with.
I think the important thing is to have a mitigation strategy. Google was very smart in how they commoditized cloud offerings by inventing Kubernetes. You can explode your manifests into GCP, you can explode your manifests into AWS, and it mostly works the same. Obviously you can really screw yourself by not having offsite backups (backup your GCP database into AWS, or vice-versa), or by using their proprietary stuff (Cloud Spanner, AWS Aurora). But if you're careful, you're in control, and the cloud provider is just a commodity that you pick based on price and nothing else. As an application developer, that's a great place to be.
(Similarly, I use Gmail for my e-mail, but I control the MX records. So if I get kicked off for whatever reason, I can be back up and running in no time.)
YouTube creators, however, do not have this luxury. People watch YouTube videos because they happened to be on YouTube and YouTube suggested the video, not because the end user was looking for your videos specifically. Social networks and video sharing sites are appealing to creators because there is the chance that The Algorithm (I hate that term) will award them with eyeballs that they can then sell to advertisers for very little effort. That is something you can't just migrate off of, because there are no competitors. You might convince your Patreon supporters to follow you, but even that is uncertain. I have watched many streamers move off of Twitch because they got a better deal from a competitor, only to watch their viewership drop to nothing until they were able to move back to Twitch. If one of these video sharing sites shuts you down, your career is over. Google kills your GCP account, you can still be a software engineer.
So my TL;DR is that cloud computing is very different from being an independent creator. Computers are a dime a dozen, and it doesn't matter which ones you use. But video sharing platforms give you a free audience, and that is something that is very hard to build yourself.
Non-technical users want easy solutions, and they need to be invested enough in your content to pass whatever barrier you put there. Projects like Floatplane[1] only capture a tiny segment of your audience.
Personally, I'd try a simple app that OAuths from YouTube/Google accounts and collects an email, so you control content announcements, even in case YouTube fails. It also has the advantage of being independent from other social media platforms.
> Similarly, I use Gmail for my e-mail, but I control the MX records. So if I get kicked off for whatever reason, I can be back up and running in no time.
I do the same, and also back up my email regularly with getmail so that it will be easy to transfer, either if Google arbitrarily locks my account, or if they just shut down Gmail for good.
I wish there was a good solution to similarly back up Google Photos -- only partial metadata is exposed via the API.
GCP might have great support. They may not have automated bots suspending accounts. They might actually follow the DMCA counterclaim process instead of some weird process on top of it.
I'll never know: I restrict Google products to casual, throw-away contexts. I've imagined myself trying to explain to my boss or a client "They won't respond" or "They say they can't provide any information." Just typing that freaks me out.
Anyway, if you want to be a huge corporation living under one brand umbrella, you've got to take the good with the bad. Google is for entertainment stuff and toys to me.