Google awards $420K in App Engine credits to science


Lest anyone think Google App Engine is just a platform for running web and mobile apps, think again. Last spring, Google invited researchers to compete for a handful of grants worth $60,000 in App Engine credits, and Tuesday the company announced the seven winning projects. They vary wildly —  with winning ideas ranging from seismic networks to personalized DNA to improving App Engine itself.

Although the research-based projects are cool — building a new type of seismic network, using machine learning to characterize cities, and analyzing climate change using Google Maps —  I think the other four are noteworthy in their attempt to build useful services atop App Engine. One wants to power an open source alternative to Mathematica and Matlab, while another wants to analyze individuals’ DNA to highlight the traits that make them unique. Another wants to let anyone create computer vision applications using a drag-and-drop interface and a webcam (I tried; it’s not too easy yet), while the remaining project wants to create a service for estimating the performance and cost of App Engine apps.

MIT Media Lab's Vision Blocks app

MIT Media Lab’s Vision Blocks app

These types of projects and services actually aren’t that uncommon in the cloud, but they’re usually housed on Infrastructure-as-a-Service clouds rather than higher-level, supposedly developer-friendly Platform-as-a-Service offerings such as App Engine. In fact, Amazon Web Services is home to many research projects and high-performance computing applications, and Google itself just awarded 100 million computing core hours to researchers on its massive infrastructure.

After exhausting their $60,000 in credits, though, one has to wonder if any of these projects will continue to run on App Engine — or any other PaaS for that matter. When you’re talking about cost-conscious efforts such as academic research, open source projects and free software, the best bet in cloud computing would appear to be good, old-fashioned IaaS clouds like AWS and Google’s own Compute Engine. Cloud providers continually drive down prices on infrastructure-level services (while boosting performance), while maintaining higher margins on higher-level, automatically scaling PaaS offerings and other managed services.

Still, there is promise in what Google is trying to do with the App Engine Research Awards. PaaS providers are all about making cloud computing easier to use by eliminating the operational aspects — essentially opening it up to anyone who can code. But if Google, Heroku, Microsoft and others are serious about luring academicians and other non-commercial users to their PaaS products to do some serious computing, they’ll probably have to bring the costs in line with those users’ ever-shrinking always tenuous budgets.


Comments have been disabled for this post