11:15 AM on Friday, October 26, 2012
Location: Lavery Hall 340
Speaker: Gabe Westmaas, RackSpace
With the rapid growth of cloud control infrastructures, scaling has become an important and difficult problem to solve. The need to balance the ability to handle millions of requests per day and control tens of thousands of physical machines and millions of virtual machines withthe needs of fast, accurate feedback poses a difficult challenge. Rackspace and the OpenStack* project are utilizing data sharding, asynchronous service interaction, and event processing, to iterate towards a system that provides the best of both worlds. This case study will describe an early proprietary version of the system architecture used by Rackspace called Slicehost and the weaknesses in that system, and the architecture of the OpenStack software that is replacing it.
*OpenStack, founded in 2010, is an open source cloud operating system that controls large pools of compute, storage, and networking resources throughout a datacenter.
Gabe Westmaas manages teams at Rackspace that write software for the OpenStack nova and glance projects. He has been a member of the OpenStack community since January 2011, and worked on both Rackspace Email and Apps and Rackspace Cloud Servers in his 4 years at Rackspace. Specifically, he has worked on an email routing and spam protection system responsible for delivering over 100 million messages a day and the full Cloud Servers stack, from API down to the hypervisor layer.