When looking to move your application to the public cloud, it is important to plan for an optimized and cloud cost efficient configuration. Maybe you already have an application running in the public cloud, but are considering a hybrid cloud configuration that includes migrating certain components to an on-premises environment in order to reduce expenditure or maintain performance. In this blog post, we will compare VMware and AWS deployment costs using an online service that has thousands of users as an example case study. The information we provide should help you take the next steps in planning and making an informed decision about your current cloud needs.
The AWS Environment
Our example application is comprised of a system with three front-end servers, three middle tiers to support Java APIs, and master-slave databases. As shown in the table below, we use an AWS m3.medium instance for the front-end, m3.large for the back-end and r3.large for the database. See the table below to learn more about the specific vCPU and memory features of each. Overall, the application has nine vCPUs, 29 ECUs and a 56GB memory.
The VMware Environment
According to an article on Pythian, a vCPU in an AWS environment only represents half of a physical core. Assuming that is correct, with VMware, you only need about half of the cores (around 4.5 cores) that are needed with AWS. What we can see from this example is that with VMWare, we get eight virtual machines with a total of 12 cores. Ultimately, just under half of those will be required. We used the VMware TCO Comparison Calculator for this calculation and seeing as it keeps the total cost of ownership (TCO) in mind, all expenses are covered in its price – from the cost of running data center operations all the way to the cost of cooling the system.
Parameters for the VMWare Calculator
VMware and AWS Comparison Results
From our calculations, VMware works out to be slightly cheaper, which isn’t a surprise. Expenses over a two year commitment with VMware totalled $17,466. Amazon, on the other hand, only displays annual figures for half of a machine. In order to compare the two, Amazon’s figures must be quartered. The end result for VMware is $4,366, and Amazon averages out to around $7,000. It would be wrong to say they are the same when it comes to pricing, although they are definitely comparable.
When it comes to pricing models, a hybrid cloud works well when you leave your steady compute usage on-premises and leverage the public cloud for elastic, on-demand workloads (e.g., front-end servers). So let’s assume that you only want to move your database to on-premises and have all other machines deployed in AWS. In this case, if the database is situated in a different location than the API servers, data transfer costs could have a significant impact on the overall cloud configuration costs and should be taken into consideration. We’ll take a look at advanced hybrid configuration models in a future article though.
Comparing VMware and AWS is not apples to apples, even with their similarities. You may be able to relate to the example above if your user demand and stack configuration are relatively static, in which case there is no need to invest in extending your on-premises environment. However, if your user demand and stack configurations are dynamic and your online service needs to scale according to demand, you can efficiently horizontally scale within AWS using autoscaling.
Additionally, you need to consider upfront costs and the need for CAPEX. While Amazon incorporates these fees into OPEX, VMware does not. In order to compare on a level playing field, further research specific to your use case is required.
Start Your Multi-Cloud Comparison Today
Start your 20-day free trial today to easily identify greater cloud efficiency opportunities and view your entire AWS deployment in one place. Gain full transparency, accountability, and control.