Info on Cricinfo
I don’t know how many of you out there are avid cricket fans and follow the progress of your favorite team ball-by-ball no matter how much pressure your delivery manager imposes on you or even during the toughest times of your work. Yes! Cricket fans like me by now would have guessed what I am talking all about. Cricinfo.com, one of the biggest sports websites in the world, logs in almost 10 million unique visitors on a monthly basis.
It all started when Simon King, an English graduate working at an American University was fed up with not knowing how his team was doing. What began as a network of fans over Chat instantaneously grew about to become the world’s largest cricket website.
Founded in 1993, the technology then was quite basic. It had one person recording the score and the other generating a flat file. With more demand from the ardent, 21st Century fans regarding the statistical analysis and more new rules introduced from the International Cricket Council, the then Cricinfo went topsy-turvy and didn’t have any other choice rather than scaling itself up. Although it doesn't seem like it, scoring cricket is fantastically sophisticated since there are an extraordinary number of things which can happen, such as a super sub or ball hitting the helmet kept behind the keeper and so on.
Zach Hitchcock, ex-employer New Zealand cricket team developed a Video Analysis software and started out his own venture with the help of Chris Cairns, New Zealand all-rounder, which latter became ‘Feedback Sports Ltd’. Based on his experience in the cricketing analysis, he was referred to Cricinfo. With Zach and Lupo Data Concepts, a BESPOKE application was developed that feeds real-time scoring information into the website.
Here is how it works. An operator sets up a match in the software by accessing the relevant match information in the database. The scorer sits in front of a television and uses a ‘point and click’ interface to describe the action as it happens. The software has an interface to score each ball of the cricket match live, in real-time. The scorer has the option to automatically generate real language descriptions of the game, which can be over-written should anything unexpected happen like super sub, etc. Once the information for a ball has been entered, the scorer sends the same and the software packages it up and in turn sends it to their Linux-based server which displays the data on the website.
Matt Pickering, Business Development Manager for Lupo Data Concepts, says
The technology choice for the project was crucial. Before we even began we could see a number of challenges we would need to address. Cricinfo has scorers in India, South Africa and England which meant we needed to deploy the software to all of those locations
Deploying real-time software simultaneously in multiple locations is not as easy as what you think. The whole application was developed in .NET Framework 2.0. The team chose to introduce Microsoft technology into a Linux environment. Cricinfo now has a Windows Server 2003 running in its data centre. This, combined with SQL Server 2000, provides the communications centre for the system. The scoring software itself employs Smart Client technology, built with Microsoft Visual Studio 2005.
Tom Gleeson, one of the owners of Cricinfo, says
We now have an option that allows users to query particular wickets or what type of balls Shane Warne bowled when Flintoff hit a four. You can really drill down and analyse the data. It’s relatively simple in terms of what the site is capable of but from a user point of view it adds a great deal of value to the site
Cricinfo now has a database that stores information about each and every match that is being played, such as who, where they’re playing, as well as historical data on every international game that has been played since 1970. Gone are days when web used to be mere pages. It’s all about streams, syndication, real-time entities, and some finest RSS technologies now.
December 8th, 2009 - 18:08
Informative post! As a student of technology, I would’ve liked to see a few more details like their architecture, stack, optimization etc. Do post them if you manage to get it.