Google may be a search giant, but when it comes to application development it remains a 98-lb weakling. People who write applications for business users (i.e., paying customers) want transparency, consistency and a clear roadmap when it comes to their tools. And as cool as some of Google’s offerings are, there is no assurance that what’s available today, will be around next week.
Secondly, Google’s cloud, as massive as it is, is seen as something of a roach motel for applications: you can check them in, but not necessarily check them out should you opt for another deployment choice. Developers say once they write for GAE, the application is locked in.
For legacy applications — and lets face it — there are many, it’s much easier to update them and put them on an Amazon-like Infrastructure-as-a-Service than to rewrite them for GAE.
“With Amazon, you can just upload a generic VM image of the OS running your database and whatever application you have written in whatever language. It’s not that much different from running it on physical hardware,” said a Boston-area programmer who specializes in web commerce applications. That’s not what he sees in the Google world.
And then there’s the perception that Google tools, no matter how slick, are for kids, not big-boy developers. “Google App Engine is for making toy apps and science experiments,” said Carl Brooks, cloud analyst for Tier1 Research. In June, Google killed off the Google App Engine for Business but said it was rolling those enterprise-friendly features into the broader GAE offering. It doesn’t look like everyone got that memo.
Still, Google keeps churning out new tools. This week it made its BigQuery data analysis tool — which promises to crunch terabytes of data quick — available to more users.
For developers who already use GAE and generate a lot of data, BigQuery could be useful. Still, Google’s tool strategy is “throw it against the wall and see if it sticks,” Brooks said. That is not a model that business-oriented developers can rely on long term.
Counterpoint: Google tools better for business than you’d think
Michael Cizmar, president of MC+A, a Chicago-based development shop, said risk-averse IT departments put the kibosh on Google tool use for reasons both valid and not.
Businesses have people trained on J2EE or .NET so they want to use them, which is understandable. But there can be good reasons to offload at least some development to Google’s toolset, in his view.
He cited one example when Google Fusion Table, a large-scale spreadsheet tool, would have been perfect for a customer application, but management didn’t give it a shot. Fusion Table takes user supplied data, automatically geo-codes it and represents it on a map. “IT pooh-poohed it … so that project took six to nine months rather than one or two,” he said.
The idea that Google represents a one-way trip for applications is overblown, Cizmar said. Provided the programmer develops to the Java Persistence API, the resulting application should be somewhat portable although he also acknowledged that GAE itself does not support all of the API so it is unclear how easy moving an app off GAE would really be.
Cizmar acknowledged that Google needs to get serious about showing why its tools are a good alternative for business developers.
If Google wants to win the hearts and minds of these developers, it has to get more systematic and clear about roadmaps, has to be more circumspect about changing things without notice, and be precise about how apps that run on Google infrastructure now will or won’t run on other infrastructure in the future. The problem is, it’s not all that clear that Google itself is serious about this market. And if it isn’t, why should anyone else take the plunge?