Hi all: Surely, most of us would have tried twitter atleast once. It is a cool 140 character (aka tweet) micro-blog facility which then members can follow and be updated.
Interestingly enough this Ruby+MySQL implementation is plagued by outages given the usage load among other issues. i was just wondering if anybody would like to take a crack at the problem from an algorithm perspective. "There are M users and each user follows N other users. Each user's timeline is rendered with latest 20 tweets from all the friends." What is a good way to model this problem ? How to scale the solution when N is very large say 10 million ? All suggestions are welcome. thanks Saifi.
