The majority of large organisations are migrating internal virtual infrastructure to the cloud because they believe it will reduce costs, according to a recent survey. The survey finds that only 17 per cent of organisations achieved their utilisation and ROI goals with virtualisation and yet, they intend to use similar planning and management approaches for their move to the cloud.

The survey interviewed 94 executives responsible for virtual and cloud infrastructure decisions at organisations with more than 25,000 employees. It revealed that many organisations are ill-prepared to make the move: 77 per cent of respondents plan to use cloud-vendor supplied tools or spreadsheets to plan the migration of workloads to the cloud and only 48 per cent plan to implement new solutions to manage cloud infrastructure.

While cloud operating models have the potential to reduce spend, it is more likely that infrastructure costs will increase if these initiatives are poorly planned and managed. Virtualisation provided many organisations with some quick hits in terms of cost savings on hardware, but the reality is that few have fully met their objectives for utilisation and ROI. Despite this, the majority of organisations are betting on the cloud without dramatically changing the approach to planning these environments.

Cloud operating models can naturally increase inefficient use of capacity and the amount of “excess” capacity an organisation has on hand in internal clouds by their very design:

  • Providing users with self-serve access to capacity can result in buffet-style over-indulgence as application owners request more capacity than they actually need to safe-guard against risk.
  • Pre-defined instance configurations and sized “buckets” of capacity may enable easier management, but they can also result in built-in excess capacity in allocations vs. customising allocations for each workload’s true requirement.
  • Increased responsiveness requires a supply of excess capacity to be held as a demand buffer for new workloads. Sizing this capacity requirement, however, is tricky and teams could end up with unnecessary idle capacity taking up room on the data centre floor.

Key findings from the survey reveal that organisations will face a direct conflict between high hopes for cost reduction and poor planning and management methods:

  • 39 per cent of respondents felt that virtualisation costs were higher than expected or delivered an uncertain ROI.
  • 70 per cent of respondents felt that moving to cloud infrastructure would decrease costs and 42 per cent cited cost reduction as the primary reason they would move systems off of internal virtualised infrastructure to the cloud.
  • Despite the hopes for cost reduction, a total of 77 per cent planned to take a very basic and biased approach to migration planning, using a cloud vendor-provided tool or spreadsheets to plan the migration of their workloads to the cloud.
  • 75 per cent planned workload movements using spreadsheets in currently virtual environments, which not only slows response times, but also takes a very simplistic approach to sizing and placement in internal cloud environments.

According to Gartner analyst Alessandro Perilli, in the June 9, 2011 research paper “The Big Mind Shift: Capacity Management for Virtual and Cloud Infrastructures”:

“Gartner defines “optimised” as a virtual infrastructure where the workload placement satisfies all of an organisation’s technical, business, and compliance constraints and the capacity is allocated to avoid resource wasting (i.e., rightsized),”

Perilli also recommends:

“The capacity management tool should allow for the definition of complex, multi-dimensional placement rules according to the technical, business, and compliance constraints inherent to each service that the infrastructure is hosting.”

Strategic workload placement is critical to achieving savings, particularly in internal clouds. Taking a manual approach to planning cloud migration, like many organisations have done with virtualisation is a recipe for inefficiency and reduced return on investment. There are simply too many factors to consider in placement and capacity sizing decisions to be able to do so efficiently and accurately using home grown tools.