Optimizing Your VMware Cloud (VMC) on AWS Post-Migration
In our VMC first two blogs we looked at
- The overview of the planning process for VMC on AWS
- Discovering your on-premises IT infrastructure and doing basic analytics (i.e. lift-and-shift)
In this post we talk about how HyperCloud Analytics provides two mechanisms to right-size and create opportunities for savings – Idle VM Savings and Right Size Savings.
With other public clouds, HyperCloud Analytics runs Optimization Scenarios both pre-migration and post-migration. This can be done because the public cloud allows for small, incremental changes to resource allocation. For example, in AWS EC2, you can purchase or reallocate resources on the fly with mechanisms in place (e.g. Convertible Reserved Instances, true-up, Reserved Instance marketplace) to either pay for the incremental cost, or recover over-provisioned resources.
However, with VMC-AWS, the opportunity for post-migration optimization isn’t as great – because there is no mechanism to recoup the cost of over-provisioned resources at anything other than a host-level. Each host is a AWS EC2 i3.metal instance – 36 cores, 512GB memory and 15TB of NVME. In practical terms, this means the unit of growth (or shrinkage) is scores of VMs at a time. This means its critical to get the sizing right up front, as the opportunity for course correction isn’t as great as it is on other IaaS clouds.
Idle VM Savings
This is exactly what it sounds like – VMs running with a very high idle rate
Right Size Savings
HyperCloud Analytics will examine the performance history pulled from vCenter, calculate the average and minimum CPU and memory utilization over time to calculate the actual requirements for this VM. These are compared against a known database of OS sizing best practices and estimate the right-sized CPU for this particular VM. The end result is finally converted to a core-count and memory allocation based on the specifications of the target host.
Two points are worth noting about the right-sizing exercise:
- When you run an assessment, you specify a period (in months) of performance data to analyze – these form the basis of the right-sizing
- Only VMs have a history and are running will be right-sized
The right-sized algorithm does not upsize your VM – if you have a VM that is resource starved on-prem, it will continue to be resource-starved in the cloud as well
As part of the assessment, HyperCloud Analytics will calculate the savings that can be gotten from Idle VM and Right Size savings and re-calculate the cost based on these savings.
Comparing to other options
Once your environment is profiled, it is just as easy for to benchmark it against other IaaS clouds to give you a complete view of the TCO in the cloud. HyperCloud Analytics allows you to run migration analysis against all the major cloud players
The below is an example report from running an assessment on VMC
It doesn’t stop here – you can get even more granular
HyperCloud Analytics can get even more granular by allowing you to dive into application specifics in the Application Modelling and start modeling individual components of your estate. This allows you build a more detailed migration plan that allows you to pick the lowest hanging fruit first and built on early successes by choosing the applications and environments that produce the best results.
Try these capabilities for a 30-day no-risk, no-strings-attached trial of HyperCloud Analytics. It is available in the AWS Marketplace as well as from our website at http://hypercloud.hypergrid.com