A Gmail outage yesterday that lasted for a few hours – and more than that for some users – was caused by a data center rolling blackout, according to Google officials.
By submitting your personal information, you agree that TechTarget and its partners may contact you regarding relevant content, products and special offers.
It all started in one of the company's European data centers where Google Inc. was performing routine maintenance, and things went downhill from there.
"Unexpected side effects of some new code that tries to keep data geographically close to its owner caused another data center in Europe to become overloaded," wrote Acacio Cruz, Google Gmail site reliability manager at the Gmail blog, " That caused cascading problems from one data center to another. It took us about an hour to get it all back under control."
Cruz added that the "bugs have been found and fixed, and we're in the process of pushing out changes."Microsoft building low-power, sleepy data center servers
At TechFest 2009 on Tuesday, Microsoft unveiled a development project for a data center server that uses less power and can go to sleep when unused.
Code-named Marlowe, the server uses Intel Atom chips, which were initially designed for inexpensive laptops. The Intel Atom uses about one-tenth the power as Intel's Xeon server processor, but can perform only about one-fourth the work. So Microsoft is studying whether it can use a bunch of Atom chips to replace a Xeon.
At the same time, the company is developing software that can learn – and then presumably predict – when there are lulls in its data centers. At those times, it wants to automatically put servers to sleep so they don't consume as much power. Then, when the action is about to speed up again (again, based on predictive models), the servers wake back up.