My primary qualm with cloud computing is one of its most fatal (and fundamental) failings, reliance on a third party ‘super’ server. Whilst there is no such thing as a ‘super server’, it is in reality, a large collection of servers within a farm, I don’t think the systems that power a variety of services from email, storage, social networking sites to whole online Operating Systems can be thought of simply as ‘vanilla’ servers. To view them in such a light dangerously understates their potential infrastructural importance.
Despite my reservations, recently I have been using more and more cloud computing services and, you know what… I get what all the fuss is about. It’s easy, simple and for each service you use, it can potentially be one less infrastructural concern to think about. Great, but it (and potentially your data) is also wholly in control of a third party and their fault tolerance infrastructure. Thats not to mention issues of Internet connectivity, traffic shaping and net neutrality.
It was bound to happen sooner or later, now I have a perfect example to hypocritically point to, The LinkUp. I only glanced briefly at the details, but it appears the company hosted data for hundreds of thousands of paying customers. During a migration, they hit a snag and, whilst over half the data is still safe after the service resumed, it is reported that many thousands could have lost everything. What if this happens to Amazon’s S3 service or Google Mail? Cloud services are unlikely to go away and I will continue to use some of them, but I do urge anyone who has a heavy reliance so such systems to seek a failsafe or separate backup form, it all boils down to the old computing concept… common sense.