Cloud computing sounds like a good idea on the surface. You store your data and run both your email and web applications on someone else's servers over the Internet, accessing everything from wherever you happen to be.
Companies using cloud computing services save money on servers and running costs, and if they need more space, they can just rent it. What could be simpler?
Not everyone is convinced, however. Some prominent names in the computing industry have begun speaking out against the concept, calling it "complete gibberish" and even "stupidity". They say that cloud computing is nothing more than fashionable marketing hype. The problem is, the recent history of cloud computing suggests that they might well be right.
Cloud computing providers have a responsibility to keep their services running at all costs, because they become a single point of failure for a huge number of users when there's an outage.
Unfortunately, such failures have begun to happen. One big example is Amazon's well-known S3 service, which has experienced two significant outages this year alone, affecting some of the popular sites that rely on it.
During the first outage on 20 July, people noticed that images on the Twitter blogging service had disappeared, as had the photo-sharing site SmugMug. The online storage service Jungle Disk, among others, also stopped working. The reason for all this disruption was that Amazon's S3 service was unavailable for eight hours. And an earlier S3 failure in February saw similarly widespread problems occur.
SmugMug was surprisingly pragmatic about the knock-on effect that the S3 failures had to its business. The company stated: "Every component SmugMug has ever used – whether it's networking providers, data centre providers, software, servers, storage or even people – has let us down at one point or another. It's the name of the game, and our job is to handle these problems and outages as best we can."
The Cloud Community website maintains an up-to-date list of outages suffered by various cloud computing sites on this wiki. The site shows that in 2007, there was just one major service failure, which happened to Amazon. By September 2008, however, there had been a total of 12. Of these, three were classed as 'high' severity and another three were 'critical'.
Though most were service failures, a couple were caused by security breaches and one was a malicious exploit of the Google Docs service that caused the service to go offline for patching. Although attacks by malicious hackers on cloud services are still thankfully rare, one log entry on the Cloud Community site may give users pause for thought for an entirely different – but equally frightening – reason.
The Linkup is dead
When cloud computing provider The Linkup ceased trading on 8 August, its CEO Steve Iverson could only say that 55 per cent of its users data was safe. "We know that there were definitely a lot of customer problems, and when we looked at individual accounts, some people didn't have any files and some people had all their files," he said.
The Linkup was originally part of an online storage company called Streamload, which split into MediaMax (subsequently renamed The Linkup) and a second company called Nirvanix in 2007.
The Linkup intended to transfer all of its user accounts from its old MediaMax servers, but this routine operation didn't go according to plan, causing the whole business to fold. The official company blog broke the bad news to The Linkup's users.
"After 8 August," reads the blog, "your account will not be accessible, all of your personal information will be deleted... and your files will be deleted. Please download any files that are in your account that you wish to keep before Friday 8 August at 5.00pm PDT."