Google Drive, the cloud storage and applications suite used by millions at home and at work, has suffered three service interruptions this week, making it impossible at times for affected users to access their files and applications.
As logged in the official Google Apps Status site, the first incident happened on Monday, and was an outage that lasted about three hours and affected 33 percent of Google Drive user requests. Affected users got error messages, long load times and timeouts, according to an incident report posted on Wednesday.
This problem was triggered by a bug in the Google network's control software, which caused the system to shift traffic over to unaffected network connections and servers. Unfortunately, this traffic load-balancing increased latency in the servers and caused another glitch, this time in the software that manages Drive user connections and sessions, causing the access problems.
According to Google, it fixed the control software bug that caused the initial problem, and will change its load balancing policy to provide "greater isolation" between network services. The company is also fixing the second Drive-specific bug, making changes so Drive will be "far more resilient" to latency and errors, and to improve Drive's alert and monitoring system.
Drive suffered another outage on Tuesday that affected an undetermined number of users and lasted apparently two hours. On Thursday—the day after Google published the report about the first incident—Drive was rocked by another outage that took about 12 hours to patch. It's not clear what portion of Drive users were affected by this third incident, which Google first acknowledged at around 9:30 a.m. EST and declared resolved at around 8:40 p.m.
Google didn't immediately respond to a request for comment.
Google Drive is part of free, individual Google accounts, as well as a component of Google Apps, the suite for businesses, schools and government agencies that also includes Gmail, Calendar and Sites. Google Apps is free for schools and universities, but starts at $50 per user, per year, or $5 per user, per month, for businesses.
As Google Drive faltered this week, frazzled users took to discussion forums, Twitter and other social media forums to express their frustration and anger about losing access to their documents.
Last week, Microsoft struggled to contain a bug that affected Hotmail and Outlook.com webmail services for about 16 hours and its SkyDrive cloud storage service for several hours.
Even outages as brief as 10 minutes can have serious ramifications for users who are trying to access time-sensitive content, so organizations that rely on cloud-based services for things like storage, email and collaboration applications need to carefully vet providers, said industry analyst Michael Osterman from Osterman Research.
CIOs and IT managers need to evaluate providers’ level of redundancy, service level guarantees and security safeguards, for example, to make sure they match the requirements of their organization.
However, he cautioned against assuming that cloud-based services are more prone to outages than systems housed inside companies and managed by their IT department.
“It is important to note that on-premise, internally managed systems can be subject to the same types of problems and the same impacts on downtime. These kinds of internal outages are rarely reported outside the company, and so it can appear that cloud-based services are less reliable than internal systems and that is normally not the case,” he said via email.
Updated at 2:17 p.m. PT with comments from an analyst.