Author(s): Colin Lacey, Posted 01/31/11
Readying your data center to handle the changes wrought by the consumerization of IT represents a rare opportunity to rethink how your company chooses, uses, and pays for information technology.
I’ve written before that data center readiness is not just a matter of technology, but rather, one of process. Technologies — or should I say “computing models,” such as cloud computing — make possible new IT processes that can transform an organization for the better.
When we help clients capitalize on consumerization by implementing a private cloud infrastructure, we find the biggest payback coming in the use of the provisioning, automation, orchestration, and charge-back models that support deployment of and to the cloud.
At the heart of the cloud model is a provisioning portal, where employees can go to request IT services from a predefined service catalog. The approval and deployment of these services can be completely automated, orchestrating all the necessary steps to connect and customize the infrastructure as needed.
Using our own experience at Unisys as an example, in a traditional (non-cloud) setting, a request for a dedicated instance of an operating system and application for test development typically took about 10 days. This was largely because the provisioning was a manual process, with multiple checkpoints for people to approve the request.
Once we moved this environment to a private cloud, a request for this same IT service could be routed for automated approval and deployment. Automating the provisioning of the same OS and test development environments has cut the delivery time to less than 30 minutes.
We automate the physical provisioning of the servers, from connecting a piece of hardware to the network, to connecting the storage LANs, to starting to image it, to putting an operating system on it, then loading the applications on it, and then mapping it to the appropriate network domains so it looks like it’s part of the right pool.
By automating that process you also remove human error. When something goes wrong in the process, it’s usually not a technology problem. It’s a human error: something has been configured incorrectly.
Automated provisioning also gives clients the chargeback tooling to see exactly how the infrastructure is being used, which department is using what, how often it’s being accessed. It provides the ability to bill back departments for their real IT usage. The data are equally useful in planning for future requirements and for future budgeting cycles. This enables greater visibility into infrastructure utilization, and aids future planning.
In addition to usage, chargeback helps you identify non-usage — that is, infrastructure that’s underutilized. As part of our own automated process, for example, we actually provide a lease on a virtual machine. If you request a virtual machine, you can request it for a maximum of [n] weeks or months.
Let’s say you request the virtual machine for three months. You might be done after two months. But at three months, you’re going to get a message reminding you that the virtual machine is going out of service. You then have the option to kill it, to extend it, or archive it in case you want to access it again later.
Archiving the virtual machine works like hibernating a laptop. The active image is captured, stored, and can be reinstated at any time. These are just some of the non-obvious, yet compelling benefits of embracing the cloud computing model.
Consumerization is not just about the anytime, anywhere access to applications and information. It’s also about the radical simplification and dramatic cost reductions that are possible when your cloud is empowered by automation, orchestration, and chargeback.
The statements posted on this blog are those of the writer alone, and do not necessarily reflect the views of Unisys.