Wednesday, March 24, 2010

Real World Probability - failure rates and social networking

People sometimes have difficulty understanding the relevance of Mathematics in today's world.

One way of overcoming this is to take the time to appreciate (and talk about) real world examples.

This is only a short post just to give you a pointer to something interesting related to - something lots of people just leaving the education system will have had exposure to.

Social networking runs atop of massive datastores, these datastores are at the leading edge of current computing theory, with regard to latency and failure rates. Mathematics is at the heart of understanding the probability of failure.

Facebook is making use of a datastore named Cassandra and the Gossip protocol - what these are is not really so important (from a Mathematics perspective). What is interesting Mathematically is the probability theory involved in the operation of the Failure Detector.

There is a great presentation by Avinash Lakshman and Prashant Malik, which you should be able to locate by websearching using this link.

The mathematics begins around page 17 of 24 titled "Properties of the Failure Detector"

( You don't have to really understand the content - just skipread the slides and ignore much of the jargon. )

The key point of this article, is hopefully, that Mathematics is being put to use today so that you can get your Facebook quicker :)

(Note: Twitter and Buzz are my social networking choices, but that does not stop me lauding Facebook and their use of Mathematics to achieve their business goals)