CloudDatacentreSoftware

Google outage was due to “overload”

Peter Judge has been involved with tech B2B publishing in the UK for many years, working at Ziff-Davis, ZDNet, IDG and Reed. His main interests are networking security, mobility and cloud

A two hour blackout of Google mail was blamed on “data centre overload” and new code bugs, by the company’s official blog.

Google’s Gmail service was unavailable for two hours on Tuesday, due to routine maintenance that went wrong in the company’s European data centres. The company has issued an explanation on its blog for the outage, which is expected to damage the credibility of cloud computing.

“This morning, there was a routine maintenance event in one of our European data centres,” said Acacio Cruz, Gmail site reliability manager. During maintenance, accounts are moved to another centre, but in this instance that caused problems because of “unexpected side effects of some new code”. Google has software that tries to keep data geographically close to its owner, and this put another European data centre into spasms, which then cascaded from one centre to another.

The software that caused the problem has been fixed, and the changes are being pushed out, said Cruz: “We know how painful an outage like this is — we run Google on Gmail, so outages like this affect us the same way they affect you.”

Although Cruz promises that problems like this are always investigated so they do not recur, the Google outage is being described as a body blow for the concept of cloud services. The event may already cost Google money, because the company has offered free credits to users who pay $50 a year for the Premium Edition.

This offer, following the Google Apps service level agreement, presumably also applies to businesses which have used Google Apps to replace Microsoft Office, such as Guardian Media Group, publisher of the Guardian and Observer newspapers and guardian.co.uk sites, which announced a move across earlier this month.

Google’s outage may be high profile, but it follows other less-wiedly reported incidents such as a half hour outage in cloud pioneer Salesforce.com in January.