Facebook has been building its own hardware for about 5 years. Starting with compute servers and moving to storage, racks, data centers, networking and more. Here we describe the path from 2011 to now. We will talk about some of the interesting challenges going from medium scale to larger and larger scale. Not only have we developed tools to help us execute our hardware validation tests, but we have also greatly improved our ability to ensure proper hardware provisioning, monitoring, and remediation at large scale. (55 mins)
Link to slides: http://files.opencompute.org/oc/public.php?service=files&t=6d5f67b88fe7d3f421415c702c037206