Profile your workloads first — you’ll net millions of dollars in savings, as well as strategic advantages such as agility and speed to market.
Migration to the cloud is a journey that presents new challenges. Along the way, you need to consider more details than you first thought, and there are many paths available to move applications, but not all of them are right for your migration.
The first step for every migration is to define the types of workloads you need to migrate. This does not mean each application, but the patterns of processing that the applications comprise. You must determine what those patterns are, then place existing workloads in each pattern.
For example, let’s say these three patterns of workloads exist in your enterprise:
- User interface-intensive
Let’s say you have 100 applications. It’s simply a matter of placing each application in each workload pattern. That way, from a planning and conceptual perspective, you’re dealing with only three types of workloads, not 100 applications.
By using workload patterns, you can place the applications and data in the proper cloud configuration selected to service that type of workload pattern. For example, database-intensive workloads need a cloud and cloud configuration that provides fast storage services, whereas compute-intensive workloads need faster server processors.
My example is of course a simplification of the process. There are typically dozens of workload patterns, and applications typically number 1,000 or more. However many patterns and applications you have, understanding those workload patterns lets you to better deal with the complexity of matching the right cloud platform and configuration to the right applications.
Migrating applications and data to the cloud is perhaps one of the most complex but important tasks that IT will take on this decade. Although you’ve ported workloads to new platforms in the past — such as to distribute computing systems, client/server, and the Web — this time you are making many leaps of faith, because the hardware is typically owned by other people, your data is maintained offsite, and neither is in your direct control.
If you do it right, there are millions of dollars savings to be had, as well as strategic advantages like agility and speed to market. Get the workload profiles right, and you’ll nail the cloud as well.
By David Linthicum