The Hackers: The Brains Behind Tomorrow's Tech

John Wilkes spent a year negotiating his move to Google, and when he finally agreed to join the company, he still didn’t know what he’d be working on.
Wilkes was recruited by a Googler named Bill Coughran — who had joined the web giant after years as a researcher at the famed Bell Labs — and he knew he’d be working alongside people a lot like Coughran. He was joining a team stacked with top computer science researchers, many poached from the leading corporate labs of the pre-internet age, including Bell Labs and Xerox PARC and HP Labs, where Wilkes himself had worked since the early ’80s. But no one would tell him what his Google research would look like.
Wilkes was lured by the prospect of teaming with these big thinkers — and he knew they wouldn’t be at Google unless the work was intellectually satisfying — but he also knew there was an added importance to their research. “One of the things that made Google interesting is that they wouldn’t say what I’d be working on,” Wilkes says, with the impish grin that so often punctuates his view of the world.
At Google, research isn’t what it used to be. When Google recruits the best and brightest computer science researchers, it doesn’t set them up inside some sort of secluded lab where people noodle on ideas that may or may not see the light of day. The web grows too quickly for that. Google pushes these minds to the front lines of the net, where they seek to remake the world’s technology as soon as possible. John Wilkes wasn’t told what he’d be working on because he was joining the team that builds the fundamental hardware and the software that underpins Google’s entire online empire — the stuff that Google views as its most vital of trade secrets.
Like other seasoned computer science researchers before him — including Jeff Dean and Sanjay Ghemawat — Wilkes joined Google in large part because the company gave him the opportunity to change its immediate future, and perhaps even the future of the web as a whole. Over the past decade, in moving these top minds to the front lines, Google itself has evolved into what you might call a research lab for the internet at large. Time and again, the company pushes the edge of the envelope, finding ways of more quickly and more efficiently juggling the billions of requests that hit its web services with each passing second, and more often than not, the rest of the net follows suit.
Most famously, a sweeping number-crunching platform called MapReduce — developed at Google more than a decade ago — is now mimicked by just about all of the web’s biggest names, thanks to an open source project dubbed Hadoop. But this is just one of many examples, ranging from data centers that snap together like LEGO bricks to world-spanning databases.
The rub is that when these Google minds go to work on such projects, there’s very little give-and-take with the larger research community. This is often the case with corporate research projects, but the secrecy only heightens at Google. The Google research model, Wilkes says, “increases the value of the ideas, and provokes more need to be circumspect, compared to when I was working on ideas that weren’t going to make it into a product.”
Even when you talk to academics who know Wilkes — and others working on the guts of the Googlenet — there comes a point where they stop and say that Google won’t tell them what it’s working on. But Wilkes is pushing to change that — at least in some ways. This can help advance the net as a whole, he says, but also benefit Google.
According to some who know him, Wilkes can be intimidating, if only because his body of work is so impressive — and because he’s a Cambridge University graduate who speaks with the sort of educated English accent Americans are so often intimated by. But they’ll also tell you that he is every bit the mentor, someone who is kind yet firm and uncompromising in his criticism. “He’s a very hands-on type,” says Andy Kominski, a University of California at Berkeley graduate student who interned at Google under Wilkes, “which is only positive for a Ph.D. student looking to do research.”
Wilkes clearly thrives on such collaboration — something you can see even on Saturday afternoons, when he blows glass with fellow students at San Francisco State University, not far from Google’s Mountain View, California, headquarters — and he’s intent on expanding the way Google collaborates.
The best evidence of this is that he’ll actually tell you what he’s working on. He’s overseeing the creation of a mind-boggling new system that will orchestrate each and every computing task that runs across Google’s worldwide network of data centers, perhaps the largest single operation on the net. This system is called Omega, and it’s meant to replace an existing tool known as Borg, which has helped drive Google’s empire for nearly a decade.
Over that decade, Borg was the best kept of secrets — outside of Google, you never even heard the name — but Omega is another matter. “I came and worked on Omega,” Wilkes says, “and I chose to do things differently.”
 
By Cade Metz 
Source and read more:
 

0 yorum: