Hyperscale operations are not massive replicas of enterprise data centers, with a heterogeneous mix of server brands, hundreds of applications and a large staff to implement change and monitor the system. The average Facebook sysadmin handles 20,000 servers, thanks to standardization, automation and a failure-tolerant IT infrastructure.
Check out these photos from inside Facebook's data halls, looking at Open Compute servers and storage, cooling systems that Facebook designed, and more.
Meredith Courtemanche asks:
Should enterprise data centers model their operations on hyperscale examples? Why or why not?
0 ResponsesJoin the Discussion