Planning Large-Scale Cloud Migration? Take a Lesson from a Global Leader and Build a Cloud Factory

It is no surprise that the vast majority of global organizations are looking to incorporate cloud models into their IT strategy in 2011.  There are lots of studies out there that point to why clouds are so appealing. The top reasons uncovered by our own survey included cost savings, greater flexibility for infrastructure provisioning, easing management pain through standardization increased responsiveness, and shorter provisioning times.

As enterprises dig deeper into cloud operating models, they realize that achieving IT nirvana and actually migrating to the cloud is tougher than it first seemed.  The easy part is setting up a self-service portal to enable users to request capacity.  Much more difficult to tackle are establishing processes for managing the workflow of inbound requests, figuring out what existing infrastructure should migrate over, how to handle that process and do it in a reasonable timeframe while ensuring that the aforementioned goals are met. Add in the growing list of infrastructure options to sort through, business and operational policies, security concerns, service levels and internal politics about cloud infrastructure and a migration initiative can turn into a veritable rat’s nest.

Top performing organizations have recognized that poor decisions lead to inefficient, risk-laden environments that fall short of everyone’s expectations. Often they have experimented with spreadsheets or vendor supplied tools to determine sizing and other basic processes to determine configurations to enable standardization. Once they grasp the complexity of the task, realize the challenge in front of them and just how time intensive and error-prone manual approaches can be, they start searching for a better way.

Key to success is investing in a process that supports a methodical, consistent way of analyzing workloads , applying pre-defined, documented policies to decisions, and placing workloads appropriately – whether in a cloud or somewhere better suited.  This kind of process not only enables consistency in decision making, but applies a discipline that is required to ensure that decisions are based on all the critical criteria and that a “cloud first” policy is applied only where and when it makes sense. One of the organizations we have worked aptly called this process “The Cloud Factory”. Truly fitting considering the fact that they were processing 100,000 workloads!  A good “Cloud Factory” process involves:

  1. Qualifying candidates for suitability for virtualization and cloud environments against admission policies, sifting out those that don’t belong.
  2. Sizing qualified workloads, comparing them against the available instance catalog(s) and choosing the best environment and instance options based on fit, service level requirements, and cost.
  3. Mapping candidates to optimal OS and SW stacks to enable best fit and standardization.
  4. Analyzing workloads into target environments.
  5. Dealing with exceptions when workloads don’t fit or aren’t suitable to place into clouds.
  6. Monitoring internal environments leveraging a fuel gauge to understand how much room remains for additional workloads and ensuring that over time, instance containers continue to be appropriately sized and workloads appropriately placed.

Consistent application of qualification criteria, utilization analysis, operational and business policies can be challenging and is why many organizations skipped some steps when planning virtualization. They simply weren’t aware of a better way to do it and the challenges of compiling the necessary data, conducting the analysis and coming up with an answer with any kind of accuracy and in a reasonable timeframe were simply too great to overcome.  Applying purpose-built analytics to the problem streamlines these processes while standardizing analysis approaches for smarter, faster decisions in the end.

As our friend James Staten of Forrester says: “Success with your internal cloud won’t come simply because you build it.”  Only by methodically analyzing the workload demands against the resource supply, and meticulously managing the placements of cloud instances and the resources allocated to them, can clouds strategies pay off, and ultimately provide a level of agility that will truly transform the way IT resources are managed.

This entry was posted in Cloud and Virtualization Management and tagged . Bookmark the permalink.

Comments are closed.