Turbonomic Blog

Migrating On-Prem to Cloud: Part 2

Posted by Dallas DeMarr on Jul 11, 2017 8:22:42 AM

In the previous article, we spoke about the complexities and costs of migrating workload from an on-prem environment to the public cloud, and the negatives of an allocation-based migration model versus a demand-driven migration model that makes use of the actual utilization of your infrastructure. In this article, we’ll see how easy Turbonomic makes it for you to plan and accomplish your migration.

In the New UI, create a plan by selecting “Plan” from the menu on the left and clicking on the special plan type named “Migrate to Public Cloud”.

turbonomic-migrating-on-prem-1

Now, we can select the Virtual Machines to migrate to the cloud. In this case, I’m selecting a few clusters from a small test environment:

turbonomic-migrating-on-prem-2

Then, I can select what cloud providers I wish to migrate to (notice the ability to select more than one, though I only have AWS added on this instance):

turbonomic-migrating-on-prem-3

Turbonomic also allows you to specify regions or groups of regions to migrate to:

turbonomic-migrating-on-prem-4

Upon running the plan, we’re presented with (among other items) a Resource Summary:

turbonomic-migrating-on-prem-5From this one panel, I can see that I am able to go from 57 to 37 on-prem active hosts! Twenty machines that can be powered down, or used for other facets of the business… and I can also reduce the number of storage devices in this environment, while keeping the same number of Virtual Machines – and my cloud cost will be $9,901.

At the click of a button, a good portion of my migration math is done for me. However, the value of this plan doesn’t end here. For example, let us look at the Cloud Cost Comparison chart:

turbonomic-migrating-on-prem-6

Based on this chart, we can get a lot of information at a quick glance. The first column is a “what-if” migration, assessing costs based on allocation, and the second is the cost of a migration performed through Turbonomic. The remaining two columns show the difference between the two, numerically and as a percentage.

The first two lines show us how many virtual machines were under- and over-sized before migrating to the cloud- each oversized VM, of course, was potentially multiple thousands of dollars in wasted spend yearly. The third line, Average VM Cost, is of particular interest. We can see from the fourth column (ignoring the actual dollar value of this particular environment) that utilizing Turbonomic to drive this migration reduced costs by 25%! What a world it would be, in which 25% of our annual spend could be reinvested into other areas! The fourth column really shines in that rather than raw value, it can be used to show proportionate savings, and allow for projective cost-benefit analysis.

In this case, we’ve no existing cloud compute, so that row is all zeroes. We can see from the row below, however, that a demand-driven Turbonomic migration will save 27% over an allocation-based migration. In case it hasn’t been said plainly enough, using Turbonomic to both plan your initial migration to the cloud and to maintain your workload once there.  Let’s take a closer look at one of the actions:

turbonomic-migrating-on-prem-7

This action is to scale down this virtual machine from 2GB of VMem to 1GB before moving the VM to the cloud. As you can see, this VM is barely using any of the resources allocated to it. Using this current information along with historical data, Turbonomic has determined that it is safe to reduce the VMem, and in so doing, reduce the cost to you. Upon moving to the cloud the VM will continue to be monitored, and just like an on-prem machine, will resize to template as necessary.

After the initial migration, Turbonomic will continue to maintain the desired state in an autonomic fashion, driving your environment to be responsive no matter where your environment lives. At the end of the day, Turbonomic will do what it has always done – stand guard as a perpetually-vigilant Warden, constantly assuring the performance of your applications and scaling your infrastructure to accomplish this, be it in the cloud, on-prem, across multiple clouds… all the while ensuring that you won’t be spending more than you need in order to achieve this goal, and live within any compliance constraints that you and your business may have.

Topics: Cloud

Subscribe Here!

Recent Posts