So everyone’s abuzz about the Google Application Engine. Is it a me-too play on Amazon Web Services? Is this gonna get ugly? (Bloggers love it when it gets ugly!) Here’s the low-down according solely to moi:
Amazon is playing the services game; Google is playing the application environment game. Put more simply: one is doing SOA and the other is doing web application hosting with extras (like authentication.) Then there’s the coupling issue: Amazon loosely-coupled (with some acceptable exceptions regarding S3); Google fully-integrated.
They’re apples and oranges, really. Though I don’t expect that to stop the comparisons and competitive hype. One could certainly take business from the other. There could potentially be plenty of ugliness. Somehow I doubt the Amazon WS team is quaking in its boots just yet.
In the enterprise, Amazon will win hands down. You can use Amazon WS without your business logic or operational data ever leaving your own data center. You can replace SQS with another queue service if you don’t like Amazon’s (assuming you’ve written your interfaces properly!) In the Web 2.0 space, it depends on a few things: 1) marketing, 2) price, 3) reliability, 4) stage in growth of web app–I believe at some point any successful web app will outgrow a sandbox.
As for the Python thing: well, it might have been a smart move. Python developers are quite vocal. If enough of them try and like GAE, the buzz generated might be deafening (see 1, above.) On the downside, if Python developers give it the thumbs down, GAE could have an even steeper curve to climb.
Me, I prefer apples to oranges. I’m just a loosely-coupled kinda gal. I like systems that are big, open, and distributed. Away with yer stinkin’ sandbox! And I happen to believe that SOA can be more than a buzzword (enterprise middleware suckage notwithstanding.)