Google Engine: Competitor or knock-off?

by Mathew on April 7, 2008 · 21 comments

Call it a clash of competing clouds. It seems that Google is launching an application-hosting service that appears to be going head-to-head with Amazon’s trio of distributed computing services — the EC2 computing network, the S3 storage service and the SimpleDB database offering, all of which provide a kind of back-end in a box for companies that want to scale quickly. So is Google’s “App Engine,” which the company described at one of its invitation-only campfire events on Monday night, a real competitor for Amazon, or just a me-too knock-off?

Aaron Brazell of Technosailor — former technology guru for b5media — says the Google announcement is “much to do about nothing.” Among other things, Aaron says that Python, the only programming language that Google’s service currently supports, is not trivial to learn or to implement (several commenters on the TechCrunch post also seemed to think that restricting it to Python was a big negative as well). Aaron’s other beef with Google’s initiative is that it seems like an “Amazon S3 me-too” kind of product. “There is no innovation here,” he says.

To be fair, however, at least some of what Aaron is skeptical about — including privacy concerns, and the wisdom of hosting applications on remote systems run by some other company — arguably apply to both Amazon’s and Google’s suite of services. To me, the bigger question is whether companies will be drawn to Google as a host for their distributed services over someone like Amazon. I think they might. And if the Python limitation is only temporary (as Google suggested it is) then that could open up the doors even further for developers. Brady Forrest of O’Reilly says that he likes the approach Google is taking.

So now we’ve got the Google File System going up against S3, and BigTable going up against SimpleDB, and EC2 going up against Google’s server stack (no cool name for that, apparently). Is this the Muhammad Ali vs. Joe Frazier fight of the tech world? Hulk Hogan vs. King Kong Bundy? Or is it Paris Hilton vs. Nicole Ritchie? Update: SmugMug CEO Don MacAskill (whose service uses Amazon S3 a lot) has a take on Google’s App Engine — he sees it as interesting, but not much of a competitor — and he’s also worried about lock-in.

  • http://smoothspan.wordpress.com/ Bob Warfield

    Aaron has missed the point here. App Engine is actually much easier than Amazon Web Services. AWS is pure API's. If you wanted to run your app in Python, you'd have to figure out how to set it up on the raw virtual Linux machine. Google kicks things a level up and let's you start in a relatively easy scripting language. Once they add Ruby on Rails and PHP, that's an awful lot of the known universe.

    More on my blog:

    http://smoothspan.wordpress.com/2008/04/08/earl

    and:

    http://smoothspan.wordpress.com/2008/04/08/well

    Cheers,

    BW

  • http://technosailor.com Aaron Brazell

    Thanks for the comment, Mat. My beef about privacy is second tier to my beef about selling the world a bill of goods on scalable infrastructure built on Google. It's the same bill of goods that Amazon sold the world with S3.

    Look, these service have their place. If you talk to the WordPress.com guys, arguably one of the most dynamic environments in existence, they will tell you that they use S3 for cold cache. That's about the extent of the benefit I think you'll see with GFS. Python is not trivial to learn, nor easy to use, but I'm betting there will be other libraries. However, that doesn't take the case in point out of the weakness in cloud computing. It is impossible, at this time, to have rich, and responsive apps built entirely on cloud. At b5, we have discovered what many others have also discovered – different environments require different optimizations. Serving images from a cold cache is one thing. They are binary and relatively static. Serving data rich dynamic scripting or data storage is a completely different ball of wax. On scale. and that's what they are touting.

    I'll give Google the benefit of the doubt for a bit. Maybe they can make it work. My guess is that they won't though.

  • http://www.digitalspace.ca/ Mark

    I'd agree with it being classified as a knock-off for now, but I think the plan is definitely to compete in the future. Python is only temporary, and to dismiss the service based on language alone is a bit premature in my opinion. The interesting part will be what the next language they'll support is, and how soon that comes.

  • http://technosailor.com Aaron Brazell

    Bob-

    Google said it was easy to use. Easy for whom? Python programmers? Great… it's easy to use.

    Chances are if you're a python programmer, you have your own environment. Why do you need GFS?

    This thing was marketed from the beginning of the announcement as being for people to have a low barrier to entry. It then progressed to low barrier of entry if you know what you're doing. Well, damn, does that really change the status quo?

    That's what I'm getting at. I didn't miss the point and in fact, I'm dead on if you want to approach this from the perspective of an entrepreneur and not a developer.

  • http://www.mathewingram.com/work mathewi

    Thanks for the comment, Bob.

  • http://www.mathewingram.com/work mathewi

    That's a good point, Aaron. Most of the applications I've read about
    that use S3 extensively, like SmugMug, use it primarily for limited
    image-serving. Not that many dynamic scripted apps or services I can
    think of.

  • http://www.fridgebuzz.com Vanessa Williams

    Amazon is very far ahead, it'll take some effort to catch up and Python-only will slow their momentum down. So will the fully-integrated thing: all or nothing. Though it doesn't sound like a big deal I think this is a major differentiator between Amazon and Google's offerings. I still say Amazon gets it more than Google does. Independent services that can be composed is the right way to go. Google's just playing the lock-in game. (Or else their architecture isn't as good.)

  • http://www.mathewingram.com/work mathewi

    I thought the same thing, Vanessa — Google seems to be aiming much
    more at a lock-in type approach, whereas Amazon seems more open.

  • Pingback: Amazon vs. Google: Services vs. Application Hosting Plus at fridgebuzz

  • http://www.fridgebuzz.com Vanessa Williams

    One is playing the services game, the other is playing the application environment game. In other words, one is playing SOA and the other web hosting. In the enterprise, Amazon will rule over Google. In the Web 2.0 space… it will largely depend upon four things 1) marketing, 2) price, 3) reliability, 4) stage of growth of app–I believe at a certain size any app will outgrow a closed sandbox.

    On the other hand, if enough vocal Python developers get on board and like it, the buzz for GAE may become deafening.

  • http://www.geekzone.co.nz/foobar/4846 foobar

    You have to have some legitimate business concerns about growing your new application on Google, though. Service level guarantees? API lock-in with a potential competitor? I have written more about that here: http://www.geekzone.co.nz/foobar/4846

  • http://aaronwhite.tumblr.com aaronwhite

    My thoughts here: http://aaronwhite.tumblr.com/post/31133767

    Summary: It's exciting + slimy. It puts Amazon in an excellent place because whatever simplicity the Google stack achieves, can be executed on top of AWS, but the a-la carte, custom linux instances AWS provides is NOT possible on GAE.

  • http://www.digitalspace.ca/ Mark

    Comments by Leah Culver (of Pownce) to Robert Scoble after the announcement echo that statement – it's a valid argument.

  • http://www.fridgebuzz.com Vanessa Williams

    Sorry for replying to myself, but there ain't no trackbacks and I thought this summary might interest: http://www.fridgebuzz.com/2008/04/08/amazon-vs-

  • http://www.mathewingram.com/work mathewi

    Thanks for the link, Vanessa — here's hoping Disqus launches support
    for trackbacks soon :-)

  • Cam

    All this hubbub about Python is unbelievable. Google simply used what they know, and knowing Python isn't a bad thing at all. Python is easy to learn. The language is compact and well considered, and compared to any of php, ruby or perl, it's remarkably readable. Is it always the right choice? Of course not.

    As for Aaron's comment that “for the growing number of non-technical entrepreneurs, python is neither easy to use and the demonstration does not demonstrate easy to scale”, I don't think Google are aiming at non-technical entrepreneurs (actually, I'd love to see a non-technical entrepreneur create an EC2 AMI.), but rather a subset of applications (and developers) that interest Google. AWS redux this ain't.

  • Daniel Gibbons

    I'd be interested in your take on the economics of both AWS and Google's offering. Someone whose opinion and knowledge I trust told me earlier this year that AWS was considered internally to be a nightmare, contributing a tiny fraction of Amazon's revenue yet representing a disproportionately large proportion of the company's infrastructure costs.

  • http://www.mathewingram.com/work mathewi

    There's no question that AWS has to be a fairly gigantic cost center
    at the moment, since Amazon said in its last quarterly report that the
    bandwidth it uses is greater than all the rest of the company's
    businesses put together — and the revenue is lumped into a section
    that came to about $131-million, which is peanuts.

    I assume the business model for this and for Google's competing
    service is the old “hook 'em with free (or cheap) and hope they
    upgrade” model.

  • Daniel Gibbons

    I'd be interested in your take on the economics of both AWS and Google's offering. Someone whose opinion and knowledge I trust told me earlier this year that AWS was considered internally to be a nightmare, contributing a tiny fraction of Amazon's revenue yet representing a disproportionately large proportion of the company's infrastructure costs.

  • http://www.mathewingram.com/work mathewi

    There's no question that AWS has to be a fairly gigantic cost center
    at the moment, since Amazon said in its last quarterly report that the
    bandwidth it uses is greater than all the rest of the company's
    businesses put together — and the revenue is lumped into a section
    that came to about $131-million, which is peanuts.

    I assume the business model for this and for Google's competing
    service is the old “hook 'em with free (or cheap) and hope they
    upgrade” model.

  • Pingback: News: Google announces Amazon Web Services competitor | TechWinter

Older post:

Newer post: