Google to drop 'scheduled maintenance' clause

Google has declared it will drop a clause excusing it from paying customer service credits on  "scheduled downtime" of its Google Apps service level agreement.

The software giant said it will now pursue 99.99 percent uptime for services on Google Apps and has amended its service level agreement (SLA) accordingly.

Google's previous threshold before it begins issuing customers service credits was 99.9 percent.

The company did not outline whether it plans to change the schedule of reimbursements offered to customers in the event of downtime. At present, Google offers three days compensation when a customer's availability falls below 99.9 percent and  seven days if it falls below 99 percent. Below 95 percent a customer should expect 15 days compensation.

But Google did say that all downtime will be counted and applied towards a customer's SLA.

"Unlike most providers, we don't plan for our users to be down, even when we're upgrading our services or maintaining our systems," Matthew Glotzbach, Google enterprise product management director said on the company's blog.

"Going forward, all downtime will be counted and applied towards the customer's SLA. We are the first major cloud provider to eliminate maintenance windows from their service level agreement."

Google claimed that last year Gmail was available to both business and consumers for 99.984 per cent, putting its email close to but not quite on par with the reliability levels of phone lines.

Fellow enterprise cloud provider, Salesforce.com, was struck down by a short outage last week.

Glotzbach took the opportunity to compare Google's effort with Microsoft's rival platform - Business Productivity Online (BPOS) - noting that it has recorded 33 planned downtimes.

Microsoft customers also suffered several unplanned outages over a week last year due to a network upgrade gone wrong, forcing an apology from the company.

Google customers suffered an outage after the search giant's backup power generator failed in 2008, knocking out 25 percent of its servers and disrupting services for two hours.

"Despite our best efforts, we will have outages in the future," Glotzbach said.