Saturday, December 4, 2010

White Paper: Analytics for Internal Cloud Management

This paper describes using a high-level approach the resource provision model that will be needed in a cloud computing environment. Some snippets from the paper are quoted beneath.
For many organizations, cloud is more of a business change than a technical change, as it alters the way departments interact and the way costs are allocated. Cloud computing enables IT departments to disintermediate themselves from the day-to-day process of providing access to applications, software platforms and IT infrastructure. Instead it allows them to focus on aligning supply and demand, and efficiently provisioning infrastructure in a way that bridges the gap between capex-oriented procurement and opex-oriented consumption. From an IT consumer perspective, it puts control back in their hands by allowing them to respond to their business needs more quickly, while isolating them from the arcane business of buying, installing and managing IT infrastructure.
The addition of specific automation, analytics and cost allocation models is key to making a cloud a cloud, as it enables the self-service model while at the same time assuring infrastructure is managed as efficiently as possible.
Rather than using old-school capacity management approaches that focus on trend and threshold models, what is required is a new focus on workload placement and resource allocations. In this new model, workload placements dictate the optimal use of capacity, much like in the game of Tetris™.
In this cloud computing environment it is important to do strategic optimization, which is a proactive, long-term placement of the resources based on supply and demand. There exist two possible consequences of the shift to this form of resource allocation:
  • Over-provisioning: which leads to resources with a low level of utilization, so this means we are wasting money.
  • Under-provisioning: which leads to operational risks, so the capacity cannot meet the demand on some occasions.
This is illustrated by the following graph:

The corresponding white paper can be downloaded here and from the original site.

No comments:

Post a Comment