How large RSS reader works (netvibes, Google reader...)

I wonder how web applications like Google Reader, Blogline, techronati works, and what technics they follow to parse millions of RSS feeds using cron job at one time?