Came across this blog post by Dare Obasanjo, PM on the Windows Live Team, about our approach to data center management where he highlights a paper written by Michael Isard of Microsoft Research, called Autopilot: Automatic Data Center Management, which describes the technology that Windows Live and Live Search uses to manage their web server farms. 

The paper provides a high-level overview of AutoPilot, while Dare’s post goes into more of its practical implementation.  Both are definitely worth reading :-)