by Hovhannes Avoyan
| Dec 03, 2009
When you think about it, the booming growth of unstructured data, app and web servers and virtual machines at many companies today is also quite counter-productive to what most companies are seeking in these troubled economic times: capital preservation and operating expense reduction.
So what’s the solution? It lies in the cloud, according to a recent https://cloudhosting.ulitzer.com/node/1200132 article I read. The story offers guidance – four key steps – for IT professionals to move their datacenters to the cloud.
First, companies should identify their most critical business applications as well as the largest consumers of datacenter IT resources. These should then become the primary targets for transforming the datacenter. “Rule of thumb is 30% of applications typically consume or create need for 70% or greater of the datacenter infrastructure,” according to the piece. Next, decompose these apps one at a time or in groups of similar types and then “measure and map the workload” across the IT supply chain. Taking this approach to change, firms can do twice the work on half the infrastructure.
Next, you should “institute a discipline” to measure and monitor consumption and performance and that also takes into account the IT supply chain dependencies of every application. The point here is that companies need to understand what an end-to-end application view and dependency looks like before they can virtualize.
The third step involves standardizing the management strategy of your IT supply chain across the datacenter from the top down. The advice: use building blocks of runtime management and service delivery with holistic virtualization. The building blocks include:
- dynamic runtime management of workloads (bringing demand to supply as it happens),
- a combination of datacenter footprints “that provide optimized physics,” from energy draw, to heat dissipation, to quantity and types of cables required to connect, communicate and run workloads. (Unified networks, compute, storage, apps and software that match workloads supported by the enterprise creates a simpler, leaner platform.)
- lifecycle management of provisioning, re-purposing and re-provisioning standard builds on top of the unified footprints. And the idea here is that your business gets the ability to flex its infrastructure in order to keep pace with fluctuating demand – at the same time taking advantage of optimized footprints and matching workload on a dynamic basis.
The fourth step is to turn these building blocks into a new way of delivering instantaneous support of IT – as business requires it. Mix in approval processes and standard operating procedures. And IT should constantly analyze data in order to “proactively predict, tune and adjust the infrastructure.
While this article at times might sound a bit complex (and perhaps daunting), I think it provides good advice on taking a step-by-step approach to enabling data centers for the cloud and saving your company money.
