An interesting article on Wired.com today, featuring Google computer scientists Jeff Dean and Sanjay Ghemawat:
Time and again, we hear the story of Xerox PARC, the Silicon Valley research lab that developed just about every major technology behind the PC revolution, from the graphical user interface and the laser printer to Ethernet networking and object-oriented programming. But because Google is so concerned with keeping its latest data center work hidden from competitors — and because engineers like Jeff Dean aren’t exactly self-promoters — the general public is largely unaware of Google’s impact on the very foundations of modern computing. Google is the Xerox PARC of the cloud computing age (more following the link…).
These Google technologies aren’t things you can hold in your hand — or even fit on your desk. They don’t run on a phone or a PC. They run across a worldwide network of data centers.
They include sweeping software platforms with names like the Google File System, MapReduce, and BigTable, creations that power massive online applications by splitting the work into tiny pieces and spreading them across thousands of machines, much like micro-tasks are parceled out across a massive ant colony. But they also include new-age computer servers, networking hardware, and data centers that Google designed to work in tandem with this software. The idea is to build warehouse-sized computing facilities that can think like a single machine. Just as an ant colony acts as one entity, so does a Google data center.
While Silicon Valley stood transfixed by social networks and touch screens, Google remade the stuff behind the scenes, and soon, as the other giants of the web ran into their own avalanche of online data, they followed Google’s lead. After reinventing Google’s search engine, GFS and MapReduce inspired Hadoop, a massive number-crunching platform that’s now one of the world’s most successful open source projects. BigTable helped launch the NoSQL movement, spawning an army of web-sized databases. And in so many ways, Google’s new approach to data center hardware sparked similar efforts from Facebook, Amazon, Microsoft, and others.
To be sure, Google’s ascendance builds on decades of contributions from dozens of equally unheralded computer scientists from many companies and research institutions, including PARC and Bell Labs. And like Google, Amazon was also a major influence on the foundations of the net — most notably through a research paper it published on a file system called Dynamo. But Google’s influence is far broader.
The difference between it and a Xerox PARC is that Google profited mightily from its creations before the rest of the world caught on. Tools like GFS and MapReduce put the company ahead of the competition, and now, it has largely discarded these tools, moving to a new breed of software and hardware. Once again, the rest of the world is struggling to catch up…
Check out the full story here.
(Contributed by Erwin Gianchandani, CCC Director)
