Does anyone else have a bad taste in their mouth from Google App Engine and as such is wary of using Google services over Amazon?
It is probably irrational since App Engine has come a long way but I remember how Google generally treated their customers early on who were trying to adopt to their platform.
The bigger complaint I've got with App Engine, which I don't have with this, is that it's complete lock-in.
If I write my app against Heroku, I can pick up and move to AWS, Rackspace, Azure, or even a local cloud such as vSphere, VMware Fusion, or Hyper-V, quite easily. That's because the PostgreSQL/Unicorn/Redis/whatever that's on Heroku is the same as what I can run on any of those other services. Even if I use Microsoft's technologies, it's generally not that hard to move form a pure Windows stack to a Mono/PostgreSQL one these days.
That's not true for App Engine. Their database engine exists nowhere else. Their general web stack exists nowhere else. Yes, I know about Appscale, but it hasn't been updated in forever, and empirically, I don't think that HBase's performance is a real replacement for Google's data storage. As a result, if Google radically alters their pricing--as they've done before--then you just have to shut up and take it.
I don't want that kind of lock-in, no matter how awesome the product is.
Frankly no - I see that event as the watershed when the appengine community suddenly matured. Thousands of people who were using appengine as free hosting disappeared, leaving those who actually understand the value proposition appengine offers.
Appengine is NOT cheap hosting. It IS an massively-scalable, massively-redundant zero-admin development platform - and maybe I'm nuts but I expect to pay for that. I was glad Google set realistic prices and put appengine on a commercial footing, because I don't want it cancelled next time Google has a bad quarter.
Just to put this in perspective, we have four appengine apps that handle about half a million users between them. They generate a 6-figure annual revenue, and we pay Google around $50 a month. Yep, you read that right - $50. We've looked at AWS, but we'd pay a lot more - and that's without factoring in the cost of a good sysadmin to support it.
If you need a company website, appengine is probably not for you. If you're putting on a royal wedding and expect billions of hits for a short period, it's perfect. If you love tweaking your own firewall rules and tuning Postgresql for ultimate performance, you don't want appengine. If you're a development shop that wants a stable, scalable platform with the overhead of a sysadmin, welcome to your new home.
I think you mean 'without the overhead of a sysadmin'... but thank you, I've been looking for a curt explanation as to when App Engine is a good choice, and you've explained it better than I ever could.
Thanks for shedding some light on this. I've been wondering about what happened to those devs who complained when Google bumped up their appengine prices. The complaints lasted for a while but then they seemed to stop. Perhaps that supports your "maturation" observation.
You mean the completely arbitrary several thousand percent price hike with still absolutely no customer service and multi-week turn around times when you can report issues?
My main complain about App Engine is the almost complete lack of support (and care) for both free and paying customers. Maybe it's only my experience, but Google has never been good at dealing with customers (and I'm not referring here to their free services).
From what I see, hear and unbiased own experience - lack of support and automated enforcements at GAE are absolute deal breakers. With numerous incurred problems with the service, the only resort was to go to the Google Groups and beg for help and attention. It was totally unacceptable. But it is totally on par with other Google services, like Adwords, where our account was suspended and later re-instated after several months of communicating through Adwords person with mighty "Google's Engineers", who refused to re-activate account because some of the domains in our paused campaigns we used in 2008 have expired and after been re-purchased by unrelated people were now serving parked pages (by Google!). So they wanted us to make those domains complaint to Adwords terms again. It was a totaly absurd situation. Being reliant on Google's traffic is already too daunting, so I thought why the hell I am bringing myself willingly to GAE, to experience their erratic behaviour, ignorance of paying customers and to feel insecure everyday not knowing what else they might re-programm in their heads to fall victim to their algorithms.... No, thanks Google. With Amazon services now it really feels like a hell and heaven.
>> With numerous incurred problems with the service, the only resort was to go to the Google Groups and beg for help and attention
Same for Amazon unless you pay them for the premium support which started at $48/month for 12 hours response time. (in contrast, I don't need to pay extra for my Linode's support and they usually response within a few minutes.)
If you are interested, visit AWS forum and search "reboot help" [1]
Supposedly the App Engine price hike was a one-time event when Google switched company-wide from made-up below-cost prices to profitable prices. Since GCE started after that, presumably its prices were rational from the beginning and won't need to be raised.
The competition Google is putting up in this space reminds me of a Ben Horowitz quote from Startup School (paraphrased):
You have to have a 10x better product to beat established competition.
Compute Engine has a long way to go to be 10x better than AWS. For most (including me) it's not just about storage and CPU units, it's things like VPC, SES, IAM, etc.
It seems Google's sole contribution to the PAAS wars is to keep Amazon's pricing in check. Does anyone here actually use Compute Engine? With Google's customer support story, I couldn't touch it with a barge pole.
I was an early convert to App Engine, ran screaming from that mess after 2 years of waiting to see the light, there was no light.
That doesn't provide causation; you don't wake up on Tuesday, see a price change from Google on Monday, and change the keynote you are giving on Wednesday to rush through a price drop. If anything, it makes more sense to think that Google got word of the things Amazon was announcing at their conference a few weeks ago, and after some careful thought decided to jump Amazon's announcement so that they wouldn't look expensive.
I am not certain how to read the comment other than "it isn't possible for it to be reactionary as it happened first", which I believe I corrected (but maybe, as I demonstrated a stronger conjecture in addition, that doesn't count?).
According to the re:Invent keynote Amazon has dropped aws prices 20+ times over it's 6.5 year lifespan. Google doesn't need to tap into it's records to know Amazon is playing the high volume\low margin game.
It is probably irrational since App Engine has come a long way but I remember how Google generally treated their customers early on who were trying to adopt to their platform.