Scaling Down to Keep Up

HELP DESK

TRADITIONAL SOFTWARE development tends to hew to the "waterfall" model, where a concept is developed, assessed, deployed and operated. But that's a time-consuming process that can be outpaced in the go-go online world, where code for Internet services is often written on the fly and fixed as needed as it is used by millions of users. Rapid deployment is great for consumers and providers alike, but it requires Internet services to keep huge technical support staffs. And that's an expensive proposition. Now a new lab at the University of California, Berkeley, aims to help budding inventors of technologies that use statistical machine learning, which could make such large support staffs unnecessary. The new Reliable, Adaptive and Distributed systems laboratory-or RAD Lab-will be funded with $7.5 million over five years with the money provided by three of the biggest names in information technology: Google, Microsoft and Sun Microsystems. Each company will give $500,000 a year to the lab. Moreover, the companies will provide expert consultants to advise the lab. Founding Director David Patter, a professor of electrical engineering and computer sciences, says the companies gain from seeing ideas at the earliest stages of development, "and they will help point out the real-world obstacles that must be overcome." If Berkeley's RAD Lab succeeds, the waterfall model of software development could become a mere trickle.-TG

Copyright American Society for Engineering Education Mar 2006
Provided by ProQuest Information and Learning Company. All rights Reserved