David,
> It's interesting that the solution livejournal have arrived at is quite
> similar in ways to the way google is set up.
Yes, although again, they're using memcached as pseudo-clustering software,
and as a result are limited to what fits in RAM (RAM on 27 machines, but it's
still RAM). And due to limitations on memcached, the whole thing blows
whenever a server goes out (the memcached project is working on this). But
any LJ user could tell you that it's a low-availability system.
However, memcached (and for us, pg_memcached) is an excellent way to improve
horizontal scalability by taking disposable data (like session information)
out of the database and putting it in protected RAM. On some websites,
adding memcached can result is as much as a 60% decrease in database traffic.
--
--Josh
Josh Berkus
Aglio Database Solutions
San Francisco