Posted 11/02/09 at 08:32:25 AM by Paul Lilly

Google's support forum has filled with messages from Gmail users who say they've been receiving 502 error messages all weekend. Many complained that the 502 blues left them without email service for 30 hours or longer.

But what those affected found most frustratingly was Google's slow response to the problem. For most of the weekend, Google kept mum about the situation and didn't reference the hiccup on its 'Apps Status Dashboard,' nor did the company respond to support requests, according to the complaints.

"I've been reporting [the outage] since yesterday evening but all's been quiet from Google," one user wrote. "The worst part is, no one I know who has Gmail is experiencing the problem. This is ridiculous."

Google did finally acknowledge the problem on Sunday afternoon, and was apparently able to resolve the issue by late evening. The company hasn't yet disclosed what caused the glitch in the first place, but did say "less than 0.001 percent of Gmail users" had been affected.

Read More

Posted 10/12/09 at 08:25:47 PM by Ryan Whitwam

It turns out that Sidekick users aren’t the only ones losing data this week. Several thousand Facebook users that have been unable to access their accounts since last week are finally seeing their profiles return. The only catch? The data from recent profile updates has been lost.

The outage was the result of, “a technical issue with a single database,” according to Facebook. Luckily, only a small fraction of Facebook’s users were affected. Unluckily, Facebook has so many users that the ‘small fraction’ works out to about 150,000. When the profiles were restored, Facebook presented users with a message that read in part, “We have done our best to restore your account to its most recent state, but some data and settings may not be current.” This could include pictures, status updates, and friend list changes, just to name a few.

Overall, the data loss was relatively minor. No profiles have gone missing entirely. So at least on that front they’re ahead of Danger/Microsoft. Most of the complaints arising from the incident seem to revolve around Facebook’s customer service. They said very little about the situation until just recently. A Facebook rep indicated that the company wanted to get the specifics figured out before providing potentially incorrect information. If you rely on Facebook, how much downtime is acceptable? Would you consider keeping important contact info or pictures on it?

Read More

Posted 10/12/09 at 09:07:27 AM by Paul Lilly

T-Mobile Sidekick owners last week were reminded about the risks of relying too heavily on cloud computing when a massive failure at Microsoft's Danger subsidiary left the handheld owners without access to the Web or their address book for several days.

If that weren't bad enough, after some of the data had been recovered, Microsoft on Saturday told customers that any other missing data may be permanently lost. While it's not uncommon for outages to occur, the potential for permanently lost data casts a dark cloud over, er, cloud computing.

For Microsoft, the timing couldn't be any worse. Next month, the software giant will launch Windows Azure, an operating system in the cloud. Microsoft is quick to point out that the Azure service is built with redundancy in mind and is able to withstand failures in single or multiple nodes, but will that be enough to convince users to put their confidence in the cloud?

Read More

Posted 03/19/09 at 03:05:33 PM by Mark Edward Soper

The promise of hosted application "cloud computing" platforms is the ability to work anywhere, anytime. Unfortunately for Microsoft, the latest storm to obscure the promise of hosted applications hit its Windows Azure development platform last weekend. It was unavailable for 22 hours on March 13 and 14, eWeek reports.

It should be noted that Windows Azure, introduced at last October's Professional Developer's Conference (PDC), is still in its test phase. It's due to become generally available before the 2009 PDC in November, according to eWeek. Although it's still in testing, an essentially day-long outage isn't good news for Azure.

Is Azure the only cloud computing provider to have had problems from time to time? How reliable should cloud computing be? For your chance to sound off, join us after the jump.

Read More

Posted 10/21/08 at 03:08:05 PM by Pulkit Chandna

Last week’s Gmail outage, which lasted for about 28 hours, has once again highlighted a major shortcoming of cloud computing and web-based services. The incidence exemplifies cloud computing skeptics’ greatest concern that unheralded disruptions in cloud computing services might cost businesses’ and individuals dearly.

Some Gmail users – including paying Google Apps subscribers - couldn’t access their accounts between 16 and 17 October. Incensed users expressed their indignation across the internet, while Mark, a Google Apps adviser, provided regular updates on the status of the issue, as long as it lasted.

“We know how important Gmail is to our users, so we take issues like this very seriously, and we apologize for the inconvenience,” Mark wrote in a Google Groups post.

Earlier this year, Amazon’s Simple Storage Service remained unavailable for 8 hours. That particular episode had also spawned similar questions regarding cloud computing. Companies will have to come out with ways to keep outages to a negligible count.

Read More