Wednesday, September 2, 2009

Gmail Outage Affects Millions


If you tried checking your Gmail at some point yesterday afternoon, you may have felt some frustration. The same frustration shared by millions of other Gmail users when Google experienced a service outage for a few hours during routine maintenance.

Google had taken some of its Gmail web servers offline to perform some routine maintenance and upgrades Tuesday morning. While Google performs this kind of maintenance quite often, they had underestimated the amount of traffic that the remaining request routers would be left to handle. The web servers that handle Gmail’s web-based interface where severely overloaded, unable to direct users’ requests to Gmail’s web servers.

Gmail users that send and receive their email via IMAP/POP access experienced no interruption, since that service doesn’t require communication with the same request routers as Gmail’s web-based interface. Those who have Gmail Offline enabled through Google Apps were able to at least connect with their inbox, but sending or receiving email was disabled. All the same, tens of millions of Gmail users were unable to access their email at all.

It wasn’t just personal email accounts that panicked about their crippled services; over 1.75 million businesses are dependent on Gmail as part of the Google Apps platform. Google uses its Google Apps service as a web-based alternative to compete with Microsoft’s various office software. Google insists that web-based "cloud computing" is far more secure than locally installed applications.

Cloud computing is also one of the leading concepts used to promote netbooks. However, outages like the one experienced on Tuesday suggest a critical danger in the dependence of other entities to support virtualized utilities.

As you’d imagine, one industry that’s hurt more by Gmail outages more than any other is the spam industry. Companies out to sell us Viagra, weight-loss pills, and fake lottery winnings couldn’t reach a large percentage of their market.

Google claims that the outage only lasted for about two hours, though many users experienced and outage for several hours on Tuesday, with reports of some outages on Monday as well. Once they’d recognized the problem, they immediately brought as many available request routers as possible online to spread out the massive demand on traffic.

Also hit with a taxing boost in traffic were services like Twitter and Facebook, who also suffered short outages on Tuesday. As Gmail users discovered their email was down, many fled to other sites to tweet their complaints. Other popular social networking sites were used as secondary means of sending messages to their peers.

In the wake of this disaster, Google has promised to learn from this experience so that it doesn’t happen again. Obviously, the first step is to insure that plenty of extra request routers are in place in order to handle potential traffic during maintenance. Additionally, they’ve set up their network so that if all request routers are overloaded, they’ll merely slow down instead of refusing requests altogether.

After much apology, Google assured its users that they’ll be hard at work over the next couple weeks to implement the necessary system improvements so that outages like this one are at most, rare. While most users may be forgiving, but businesses considering Google Apps to handle important tasks may be more hesitant to subscribe.

No comments:

Post a Comment

ShareThis