Guest post from Tony Lucas of XCalibre Communications
Excuse the pun of yet another ‘2.0’ expression, but it does actually fit quite well in with the overall point of this post.
Over the past weeks/months, I’ve had conversations with more Web Application developers than I can count and time and time again, the same problem has crept up. Most of these developers have pretty well formed ideas on how to scale their application from an internal coding point of view, but vary from having some plans about scaling their application involving large hardware expenses, to getting grey hairs from not knowing how to solve the problem.
The problem really comes down to the fact, that many ‘Web 2.0’ developers are exactly that, developers not system administrators or datacentre technicians or network admins and with the cost base for building web apps becoming leaner and leaner as the market gets more competitive (with barriers to entry being lowered), this is not something people can usually afford to have onboard, until their application becomes a success (or is funded!).
This is where utility computing comes in, by having a pay as you go platform for hosting your web application, this then enables you to completely forget about scalability concerns, building the platform out, coping with sudden growth, expected or otherwise, and dealing with peak demand.
It also increases redundancy and high availability, as you are hosted using fully redundant systems, but at a very cost effective price, compared to the current solution of buying the cheapest systems that you can, without any redundancy.
Utility Computing is still in its relative infancy, but there is already a handful of companies offering a range of services, in both the UK (Flexiscale) and the US (Amazon EC2 & S3), and is a market that is going to grow rapidly in the coming years, with fast growing business expected to gain the most from it.
It also has a nice side benefit that a typical Utility Computing based application has between 50-80% less Carbon output than a typical hosting solution (based on our own internal analysis).
I’d be very interested to hear other people’s opinions on this.