Apr 162008

Google App Engine

I'm fascinated by Google's recent announcement of Google App Engine.

Quoting Google,

Google App Engine enables developers to build web applications on the same scalable systems that power our own applications.
google app engine
No assembly required: Google App Engine exposes a fully-integrated development environment.

It's easy to scale: Google App Engine makes it easy to design scalable applications that grow from one to millions of users without infrastructure headaches.

It's free to get started: Every Google App Engine application will have enough CPU, bandwidth, and storage to serve around 5 million monthly pageviews.

Right now the only dev language is Python (snakes on a plane!), but others are promised.

The API currently provides storage (vs BigTable), authentication (via Google Accounts), templating (via Django), emailing, form handling, and URL fetching. Everything runs in the cloud, sandboxed, no writing to local files.

Here's Google's Brett Slakin demoing a guestbook in 10 minutes.

And here's the resulting guestbook app: http://shoutout.appspot.com/.

Relevance to online retailers?

I don't think AppEngine is going to change or revolutionize online commerce as a platform.

The hard problem in e-commerce isn't scaling the technology, the hard problem is marketing, getting the customers.

But I do think AppEngine (and EC2, etc) could revolutionize the speed which new Web 2.0 communication apps are built.

In the next year or so, that may have implications for how online marketers communicate with customers and prospects.

My thoughts after reading the AppEngine docs was:

"Wow. With this tool, a smart small agile development team of three people could beta something like Twitter in a few days, or maybe Meebo in a few weeks, and release it, and if the world liked it, they could scale their app up to the planet, GoogleSizing it, with zero capex, zero sysadmin and zero hardware hassles, and simply pay for cycles/bandwidth/storage as it scaled, as it proved itself viable via revenue or pageviews or buzz or whatever. You'd build a layer between the code and the API to facilitate moving the app later to dedicated servers as needed if concerned about autonomy, but AppEngine slashes the VC needs of a startup even more -- now the costs are really just people, no machines -- and handles the scaling problem pretty far along the way."

(I also thought, "Dang -- I wish I had more depth in Python!")

The apps available now are demos and toys, but more will come.

I guess the entrepreneur / developer considering AppEngine need to answer the "how much do I trust Google to park my app on it", but I think that's less of an issue for the freshly minted startup.

My prediction: cloud app platforms will increase the velocity of experimentation and deployment of new crazy marketing and communication ideas.

Overheard at Shop.orgSocialMedia2010:

What, you're not advertising to teens via TwitzaZooMeeBling yet?!? You're so far behind the times. They're 90 days old but they've already hit 12 million registered users! And the only hardware they own are the founder's laptops!


5 Responses to "Google App Engine"
friarminor says:
'Snakes on a plane' just gave me the guffaws. :) And how about BigTable - which is the 'scale' ingredient at the cost of lock in? Still, clouds are pretty neat in that they will allow plenty of app deploys (both ugly and cutting edge. Expect more big players to jump in the next few months. Oh, et us not forget that there are small players in the game, too like Morph eXchange which pre dates what Goog might actually do when the time comes for them to support Rails. Best. alain
Not Meebo, at least for the moment. I wouldn't be surprised if they do put in some sort of push messaging system eventually (they could use Google Talk infrastructure) but for the moment it's very conventional-website oriented. Otherwise, though, I agree. I suspect that it'll do a lot of damage to the webhosting industry, but quite frankly that industry _deserves_ it. friarmnor: Although BigTable might be hard to replicate entirely, there are existing less scalable (but still scalable) object database systems out there which should be able to provide the functionality that GAE's one does.
Robert: Meebo today -- point well taken. Robert: why do you see this challenging the hosting (shared? dedicated? colo'd?) industry? I miss the immediate connection... IMHO, it is a different set of biz and IT skills to start a biz born in the cloud, vs. traditional hosting customer...
I'd be very surprised if small contract developers don't start using GAE; it's a lot less trouble than the usual suspects. In addition, I'd say Google will shortly let people start selling GAE apps; you could, for instance, write a CRM and sell it to companies, with Google taking a cut. As to hosting, I think it'll primarily hit shared.


Check out what others are saying...
[...] 17, 2008 by honewatson This is the best post I’ve seen on App Engine so far. “Wow. With this tool, a smart small agile development team [...]

Leave A Comment