More Outages in Forecast for Cloud Computing - Systems Management News On The Web

CHANNELS

ON THE WEB

PRINT EDITION

BZ MEDIA

AS OF 8/20/2008 9:37AM EST

More Outages in Forecast for Cloud Computing
By Jeff Feinman

July 25, 2008 — With the outages occurring in Amazon’s S3 cloud storage service and Apple’s MobileMe device, analysts say growing pains for cloud computing will lead to further problems.

James Staten, a principal analyst with Forrester Research, said in these early stages of the cloud that outages are expected. He said he fully expects more outages this year and next year for the S3 service, but if Amazon is doing the proper things to improve service, the outage lengths will decrease from the eight-hour incident on Sunday.

S3 users should make sure that there is a backup in place, whether it is in another cloud storage service or in replicating data to their own data center, Staten said. However, cloud outages, compared to ones that the average enterprise experience within its own data center, are pretty minor.

Staten downplayed Amazon and Apple’s outages, saying that an eight-hour outage isn’t something to worry about, in Amazon’s case. Additionally, S3 didn’t report any data loss, only connectivity outage. He also attributed Apple’s MobileMe e-mail issues to congestion in the system.

“That said, all the [companies] that want to offer cloud service have to keep in mind that an outage is a whole lot different at this level,” Staten said, because of connectivity loss. “A cloud outage gets you front page news.”

John Rath, an independent consultant, agreed that because cloud computing is still early in the game, there will be outages. He also said that a cloud should consist of multiple data centers, so that if one goes down, there will be minimal effect.

“I kind of thought that’s what S3 was, and some of the other Amazon services,” Rath said, “but I’ve since learned that they’re not that distributed, and there’s still a single point of failure with the log-in piece. You’re logging in through a central place, and if that goes down, you’re not able to get to your resources.”

Amazon’s other cloud service, the Elastic Compute Cloud Web service, has also experienced outages, but Staten said they have diminished in frequency and length, “which shows they’re getting better at this.”

The key for Amazon and Apple to improve service in their cloud systems is monitoring and triage processes to fix the outage. “Did the company know what the outage was right away, or did they lose hours trying to diagnose the problem?” Staten asked. “There’s no universal answer here. Sometimes you’re going to find out what the outage was caused by and you’ll address it right away, but other times, you may never find it, and then suddenly you’re back in business.”

Spotlight on Recent Problems
Staten said the growing-pain period for cloud computing will take several years, but it will happen “player by player.” Amazon may have a head start on newer cloud providers in ironing out initial problems because they were the first in the market. At least through 2010, people should take the position that clouds are immature, so they should have proper backup in place, according to Staten.

Along with Amazon's S3 cloud service disruption on Sunday, service disruptions were also reported with Amazon’s Simple Queue Service, which lets developers put data on components and applications. The social networking site Twitter was also down. The S3 disruption was first reported at 9:06 a.m. PDT on Sunday, and Amazon described the problem as “an issue with the communication between several Amazon S3 internal components.”

Service in Europe was fully restored at 3:23 p.m. PDT, and U.S. service was back by 5:12 p.m. PDT, according to Amazon. The company experienced a similar outage in February of this year, which the company attributed to elevated numbers of authenticated requests.

Apple’s MobileMe, which automatically pushes new e-mail, contacts and calendar events to a person’s Mac or iPhone, has been experiencing e-mail problems for several days. The issue began on July 18, and the company said it was due to a mail server. The issue is effecting only about 1 percent of MobileMe members, the company claimed, and customers are unable to access e-mail using e-mail client software such as Mail on a Mac or Microsoft Outlook on a PC. As of today, the problem remains unresolved.

Related Search Term(s): backup & recovery, cloud computing, e-mail, Amazon, Apple

EMAIL THIS ARTICLE

SEND FEEDBACK