Stormy Weather Hits Amazon's Cloud

By Judy Mottl

July 22, 2008

Amazon's S3 service ran into some stormy weather on Sunday when technical glitches knocked the company's "cloud" storage offline for several hours. The mishap underscores what could become a growing problem for channel partners who move too rapidly into the increasingly popular world of cloud computing.

The company's service -- also known as "Simple Storage Service" -- offers storage hosted "in the cloud", on Amazon's servers. But the service went offline for three hours when, communication problems among S3's hosts prohibited the servers from interacting properly, spokespeople said. To restore service, Amazon was forced to take the entire system offline to restore communication aspects and then re-launch it several hours late

"These are sophisticated systems and it generally takes a while to get to root cause in such a situation," a company spokesperson said. "We will be providing our customers with more information when we've fully investigated the incident. We're proud of our operational performance but any downtime is unacceptable, and we won't be satisfied until it is perfect."

While its technicians were working to restore service, Amazon provided frequent updates on its Service Health Dashboard Web site, a change from how it handled matters during the service's earlier outage in February.

The provider came under fire from users for not providing more details during that several-hour crash, which it later attributed to overloaded authentication problems. It took Amazon two weeks to inform customers what exactly had happened.

In April, the company launched the Service Health Dashboard Web site, which Amazon said had been in development since before the February downtime, and did not come in response to user complaints.

Amazon S3 and other cloud options are becoming more important to the channel as they grow increasingly popular with both small companies and power users who need low-cost data backup and disaster recovery services. Yet, as one analyst said, such outages illustrate the technology is still in the early days of development.

"All providers are going through some growing pains," said Henry Baltazar, an analyst at the 451 Group."The timing was good on [Amazon's outage] since it was Sunday and not during the workweek."

As with any hosted service, downtime and meeting service level agreements (SLAs) are huge factors for survival in the increasingly competitive marketplace for cloud storage.

According to Amazon's SLA, the service will "store data durably, with 99.99 percent availability. There can be no single points of failure. All failures must be tolerated or repaired by the system without any downtime."

Whether Amazon's track record hurts its long-term chances remains to be seen. But it's clear that there's a growing list of companies ready to move in on its turf, including storage powerhouse EMC's Mozy service and newcomers such as Nirvanix.

Just last week, for instance, EMC debuted software for companies to manage both their online storage through Mozy as well as their local storage on EMC's Iomega hard disk drives.

In the future, according to one player in the field, the field may be further upset by cloud storage solutions tailored to specific customer needs.

"There is no need to have a super-mega-cloud to realize cloud storage benefits," Sajai Krishnan, CEO of Parascale, which makes software for companies to build their own in-house storage services.

"For example, regional and specialized cloud storage vendors could provide different kinds of clouds for different kinds of applications, such as fast streaming media versus 50-year archives, to enable better SLAs," he said.

(This article was adapted from Internetnews.com.)