Open source grid computing takes off

This has been fun to watch. The Hadoop team at Yahoo! is moving quickly to push the technology to reach its potential. They’ve now adopted it on one of the most important applications in the entire business, Yahoo! Search.

From the the Hadoop Blog:

The Webmap build starts with every Web page crawled by Yahoo! and produces a database of all known Web pages and sites on the internet and a vast array of data about every page and site. This derived data feeds the Machine Learned Ranking algorithms at the heart of Yahoo! Search.

Some Webmap size data:

  • Number of links between pages in the index: roughly 1 trillion links
  • Size of output: over 300 TB, compressed!
  • Number of cores used to run a single Map-Reduce job: over 10,000
  • Raw disk used in the production cluster: over 5 Petabytes

I’m still trying to figure out what all this means, to be honest, but Jeremy Zawodny helps to break it down. In this interview, he gets some answers from Arnab Bhattacharjee (manager of the Yahoo! Webmap Team) and Sameer Paranjpye (manager of our Hadoop development):

The Hadoop project is opening up a really interesting discussion around computing scale. A few years ago I never would have imagined that the open source world would be contributing software solutions like this to the market. I don’t know why I had that perception, really. Perhaps all the positioning by enterprise software companies to discredit open source software started to sink in.

As Jeremy said, “It’s not just an experiment or research project. There’s real money on the line.

For more background on what’s going on here, check out this article by Mark Chu-Carroll “Databases are hammers; MapReduce is a screwdriver”.

This story is going to get bigger, I’m certain.