Capacity Planning for your Virtual Data Center and Cloud – Part 2
In this 2nd part of my blog on capacity planning, we shall look at the steps to implementing Capacity Management for your Virtual Data Center and Cloud.
Many IT organizations have used virtualization and cloud to paint a visionary picture of agile, on-demand and cost-efficient IT that will meet changing business requirements for their business users. While this is possible, many IT organizations by now should have realized that at the heart of this capability is capacity management. Put simply, this is what allows IT to always have adequate compute, storage and network resources at a minimum possible cost to the company. Traditionally, application owners would size for usage peaks plus buffers while IT administrators would pile some more buffers and contingencies on top of these. We need to move away from this practice of “Just in Case” capacity planning – which is not optimal and results in underutilization and wastage – and move towards a “Just In Time and Just Enough” model for capacity planning where the right amount of resources are available at the right time and, most importantly, the right cost.
In today’s highly virtualized or “cloudified” IT environment, running IT has a lot in common with running a factory. For example, both need to ensure that their stock inventory is minimized to reduce costs and yet have enough available surplus to meet the production plan. They also need to be able to quickly reshuffle and redeploy resources and inventory should there be changes to production demands. In a factory, this role is titled ‘Production Planner’. In IT or ITIL framework, we refer the role as ‘Capacity Manager’.
One of the first steps to implementing Capacity Management will require assigning a Capacity Manager (sometimes it is also referred to as Capacity Planner) who will oversee the capacity planning and management of virtual/cloud resources. Some of the key activities the Capacity Manager will perform are:
- Getting forecast figures from the business units to understand project pipelines and future demands on virtual/cloud resources
- Baseline current capacity and BAU demands; and develop a hardware and software procurement plan and budget based on forecasts
- Set capacity thresholds and monitor usage to identify wastage or to initiate hardware/software acquisition to add capacity when usage hits the upper thresholds
- Regularly review forecasts and actual demand. Readjust procurement plan and budgets as business requirements change
- Through monitoring, reclaim or balance workloads and resources as appropriate
The next thing we must establish is the capacity management process, which addresses the activities listed above. A typical high level process flow is shown below:
As part of the process deployment, policies, rules, metrics and KPIs need to be defined so that appropriate monitors and reports can be implemented.
We will continue our discussion in my next blog!