MeFi post: MapReduce: running large-scale computations in parallel

I read this paper a couple of years ago and implementing something that runs on servers is fairly simple. You just write a map function, and a reduce function (usually the simpler of the two), and that's it. I know that many google engineers often write MapReduce programs, so it's not just the large-scale problems and top engineers that use them.

It's interesting how many problems can be reduced to Map-Reduce form. Lots of real world problems as well as your standard... [more]
posted to MetaFilter by tasty at 12:30 AM on September 8, 2006