The Private Cloud hoopla

Another fine post from Tom Bittman If You Build a Private Cloud, Will Anyone Come?. He helps cut through the noise around private clouds with a statement that could be applied to much more than just cloud computing

We’ve got to get our IT people to stop thinking about products and technologies and even architectures first, and instead to focus on understanding their service requirements first.

Okay I’m guilty of over exuberance myself at times about technology for the sake of it (recent SSD post as an example) but Tom’s spot on – we need to think requirements and then think solutions. There is a little too much “oh I need a cloud. Yes a public on and a private one"…..why? “because Johnny the CIO at company X has one”.

Cloud computing holds great promise and gets a lot of attention for potential cost savings. At our Worldwide Partner Conference recently we talked about potential cost savings of up to 20% over 3 years.

Does that mean every company and every workload should move to the cloud now? Of course not. To paraphrase Tom, grab a piece of paper and a pencil and go find a quiet place. Sit down and list out what the workloads you have are, what are their SLA’s, do you have projects that spin up quick and down quicker that may suit a cloud environment? All relatively easy to do and it gives a much more grounded view of the cloud world. Stuff like HR systems and finance are not going to move to the cloud overnight but your online store may do – or part of it. When done with that, check out Tom’s post titled Virtualization Unlocks Cloud Computing as it may well be you can get all the perceived benefits of cloud through internal adoption more widely of virtualisation. Note virtualisation and cloud computing are brothers in arms but they’re not one and the same thing.

As for a private cloud, unless the organisation is setup for internal charging and SLA’s against specific workloads I’m still struggling to see where this is different from outsourcing.

Comments

• Anonymous
  August 12, 2009
  Spot on.  The driver for building an internal cloud is going to be  that one big project, building the cloud platform for it to sit on then positioning things next to it. No company in their right mind is going to build a cloud in the hope that the project will come - those that do will build an under resourced one for test that will fall on it's lack of resources. Internal services companies (SLA driven, internal charges) have the most hope of getting the tech off the ground, and the most to gain.  They'll see economies of scale etc. Most however aren't commercially focused enough to be able to reach escape velocity as each service is a separate entity and therefore it's easy to charge back.  Creating a platform requires some long term vision, beyond project scope and very strong architects who can lead their businesses thinking.