Google has always withheld the number of servers running in its data centers, but at a recent ACM workshop presentation, Google engineer Jeff Dean showed that the company is in preparations to manage as many as 10 million servers in the future. One of the keynote speakers on large-scale computing systems, Dean discussed some of the technical details of the company’s unfathomable infrastructure, which is spread across dozens of data centers around the world.
In his presentation, Dean also discussed Spanner, a new storage and computing system which will seek to automate management of Google services across multiple data centers. Dean says Spanner will be designed for a future scale of 1 million to 10 million machines, the goal being “automatic, dynamic world-wide placement of data & computation to minimize latency or cost.”
Over the long-term, cost management like this could address both regional differences in bandwidth costs as well as power costs. The ability to seamlessly shift workloads between data centers creates energy management possibilities, even including a strategy which takes advantage of lower costs for power and cooling during the overnight hours.
Automation like this could also allow Google to plan more energy-efficient facilities like the one it erected in Belgium. When the weather gets too warm to operate servers safely, Google plans to shut down equipment as needed in Belgium, shifting the computing load to other data centers. Courtesy of datacenterknowledge.com