Although the T-Mobile Sidekick data disaster caused much gnashing of teeth among the device’s users this week, the incident might not represent as dark a blotch on the concept of cloud computing as it may seem.
Sidekick users lost all their data, which is stored on the infrastructure of Danger, the company that developed the Sidekick and is now owned by Microsoft.
Microsoft later claimed it managed to recover most of the lost data, and it announced plans to restore it soon. Meanwhile, at least two users have filed lawsuits over the issue, and the entire episode has raised questions about the reliability of the cloud.
Clouding Up the Issue
Sidekick users can’t store their data on the devices — it is saved on the infrastructure of Danger. So when users lost all their data, many reports appeared to blame the problem on the Danger cloud.
Coming to the public’s attention shortly after recent reports about hackers attacking Web-based email accounts at Yahoo, Google and Microsoft, as well as social networking sites such as Twitter and Facebook, the Sidekick fiasco worked to strengthen doubts about the reliability of cloud computing.
However, Sidekick users’ data was not stored in a cloud. Cloud storage consists of a combination of virtual and physical servers that create a pool of resources. Servers are called into use when there is demand and taken offline when demand goes down.
Off-site storage, on the other hand, is a fixed resource that is always available, and that’s how Danger’s backup and storage system was structured.
“I think the event is being presented as an issue with cloud computing or with Microsoft technology, but it really does not seem much related to what Microsoft is doing in the cloud,” Al Hilwa, program director of application development software at IDC, told TechNewsWorld.
“Danger has a separate infrastructure based on Java and running on open source products — Linux and MySQL. It runs independently from Microsoft, which is 100 percent on the Windows platform and SQL Server.”
Danger Is Their Business
Microsoft is involved because it bought Danger last year for a reported US$500 million. T-Mobile, which carries the Sidekick and initially offered customers $20 in credit as compensation, has also been caught up in the issue.
At least two lawsuits have been filed against Microsoft and T-Mobile over the data loss.
Information that supposedly came from a Microsoft insider has kicked up a rumor in the blogosphere that places blame on Microsoft management. A Microsoft executive told technicians upgrading Danger’s storage area network (SAN) that it wasn’t necessary to back up the data before proceeding with the upgrade, the rumor goes.
After asking Microsoft for comment, TechNewsWorld was referred by company representative Brianna Pinder to the latest statement on the issue published on Microsoft’s Web site.
That statement, signed by Roz Ho, corporate vice president for premium mobile experiences (PMX) at Microsoft, stated that most of the data had been recovered and that Redmond would begin restoring the data as soon as possible after validating the data and its restoration plan.
“We will then continue to work around the clock to restore data to all affected users, including calendar, notes, tasks, photographs and high scores, as quickly as possible,” Ho said.” The data loss affected “a minority” of Sidekick users, according to Ho. The next update on the timing of data restoration will be delivered by Saturday, she said.
The data outage was due to a system failure that created data loss in the core database and the backup, Ho said. Microsoft rebuilt the system “component by component,” recovering data along the way, she explained.
Microsoft has made changes to improve the overall stability of the Sidekick service and initiated a more resilient backup process to ensure the integrity of database backups, Ho said.
Independence Is Dangerous
It’s not uncommon for companies to allow their acquisitions to run their own infrastructures, Rob Enderle, principal analyst at the Enderle Group, told TechNewsWorld.
“Danger’s platform was unique, and the migration would have to be done without disrupting the current user base,” Enderle explained. Danger may have stopped doing full backups before Microsoft acquired it and perhaps did not reinstitute the practice.
That could have caused the delay in restoring data. “When the problem occurred, the redundant backup was not initially available, and Microsoft had to rebuild the database,” Enderle said.
“Generally, unless the storage medium has been destroyed, there are a number of ways, all expensive, to recover from something like this, and it appears they were successful,” he explained.
“The problem was the result of bad practices meeting bad circumstances,” Charles King, principal analyst at Pund-IT, told TechNewsWorld, “but it does call into question what consumers can and should expect from a hosted services provider.”
Microsoft had probably not fully audited Danger’s IT procedures, according to Enderle.
“That’s not an uncommon mistake in an acquisition, but one that can be avoided if the acquisition team doing the review is experienced,” he said. “Mistakes can happen, and this was a big one.”