Skyblog, Friendster, twitter, yahoo mail..... how to design database for huge no. of users and rows

[Leo Park]

I read the case studies and got to know that skyblog and friendster are running with MySQL.

As a college student majoring in computer science, I always wonder how these websites manage their huge database.

For example, how Skyblog would store the blog entries of their users? Obviously, it won't store all the blog entries in a single table - the table will be huge and performance will be much degraded. Would it be that the DB design being that each table stores blog entries for, say, 1000 users? But, even so, the table will accumulate to containing billions of rows. What would they do? Using merge or partition?

Wikipedia is using MySQL too. As far as I observe, each article may have up to thousands of revision. How do they maintain the huge table of revisions?

Just interested in the design and implementation of these huge database. Any advice?

Thanks a lot.