I’ve just been reading about business adoption of cloud computing and the risk vs cost equation. This piece goes along nicely with my previous piece about what workloads are appropriate for cloud computing. The problem with these issues isn’t cloud computing, it’s being stuck in old “common wisdom”.
The real issue here is that common wisdom simply can’t deal with the idea of inexpensive on demand infrastructure. Of course, before the rise of cloud computing, that’s the way it was. Trying something meant a large investment in equipment and facilities that could create a real catastrophe if it didn’t work. This leads to focusing on nothing but reducing the risk of failing because common wisdom trained everyone that the cost was going to high regardless, so that side of the equation could basically be ignored.
These days, the cost of infrastructure from a cloud computing provider has become microscopic compared to the past and failure means little more than wiping out a virtual machine and starting over again in minutes. Even with this change, extreme risk aversion is still the controlling factor. I’d even go so far as to say that much if not most of the opposition to cloud computing comes from this rather than any real issues.
In my previous post about cloud workloads, I noted similar behavior. Common wisdom of the past, based on expensive pre-cloud infrastructure, led to condensing as much work as possible on decreasing numbers of larger and larger servers,
It’s absolutely true that workloads like this, in their current form, aren’t a good candidate for direct transfer to a cloud. The problem is the first thought isn’t to reverse course and spread these workloads out, the first thought is that this is an excuse to keep the old static infrastructure in place, regardless of the cost, reliability, and agility benefits of the cloud .I read a statistic a while back that claimed that many companies were declaring their virtualization projects complete with only 30% of their workload virtualized. That’s pathetic.
The first article called it right, fail more often but fail small. I’ll follow that up with big workloads are bad, small workloads are good. It’s time to dump common wisdom that isn’t so wise anymore and embrace cloud computing.
Email or call me or visit the SwiftWater Telecom web site for cloud computing services.