“Cloud is cheaper than on-prem!” You’ve heard that marketing blast from cloud providers constantly this past year. But now the game has changed, with the availability of VMware Cloud on AWS announced this week at VMworld. VMware’s strategic move for partnership with Amazon enables you to cut your consumption costs on AWS by 40% through their new offering. How do we know this? We follow the data.
Dive Into the CloudPhysics IT Industry Data Lake
Cloudphysics is a pioneering solution in the SaaS-based operational analytics market, serving the pervasive VMware installed base with the data needed to drive nextgen infrastructure design and operations. More than 13,000 organizations have connected to the CloudPhysics platform since inception. As a consequence, we have filled a massive, industry-wide data lake containing the statistically meaningful distribution of vSphere resource configurations, utilization, and performance.
VMware Announced Pricing—But You Don’t Know Your Expected Costs
For market research as part of our Data as a Service (DaaS) offering, CloudPhysics data scientists periodically run interesting queries on our anonymized global dataset, looking for unique insights and data trends. So when VMware approached CloudPhysics to help with cost analysis for their new offering on AWS and its go-to-market, the team pulled together a comparison between AWS and VMware Cloud on AWS.
Because both are time-based consumption models, the pricing announcements this week require solid data science to determine your actual expected costs. We had built cost models and analysis for our VMworld session this week. This is what we found by analyzing cluster scenarios across several thousand organizations, comparing AWS (EC2) instances with VMware Cloud on AWS:
- VMware Cloud on AWS is cheaper than AWS in 52.81% of clusters after CloudPhysics optimizations
- VMware Cloud on AWS is at least 20% cheaper than AWS in 17.98% of clusters after CloudPhysics optimizations
- VMware Cloud on AWS is at least 25% cheaper than AWS in 10.11% of clusters after CloudPhysics optimizations
- The most any customer saves is 39.90% by using VMware Cloud on AWS over AWS
So VMware Cloud on AWS offers you the ability to maximize the utilization of physical servers presented by AWS. You can achieve optimizations on that hardware that AWS itself cannot (or will not) deliver on the hardware AIM instances share. VMware Cloud on AWS changes the game by expanding the unit of consumption to the cluster instead of individual instances. The result: you can control and maximize utilization as never before by oversubscribing on resources and leveraging the elastic physical capabilities provided by VMware Cloud on AWS. You can’t reach this max utilization on premises. Why? Because of the need to have capacity always online to handle scheduled peak loads and unpredictable events. Further, you can’t do this in a colo or other public clouds because there is no equivalent of “elastic physical” to dynamically add/remove physical capacity that is fully integrated with the SDDC stack.
Know Before You Go
However, there’s a catch in realizing the cost structure that VMware Cloud on AWS offers: you have to know what you’re doing. That is, you must leverage operational analytics to find your optimal use of the infrastructure in AWS. If you don’t, you’ll end up with sticker shock—in a cost structure more expensive than on-prem or AWS! Therefore a best practice (and job-saving) approach requires that you:
- understand your workloads in all 4 resource dimensions (cpu, memory, network, storage)
- understand the cost basis of those workloads over time to compare your on-premises costs to VMware Cloud on AWS and other public cloud alternatives, like instance matching
- use CloudPhysics’ analytics to apply workload profiling/resource consumption to the bin-packing problem of optimizing your VMware Cloud on AWS clusters. This is the domain of data science
ABC—Always Be Costing
It’s also important to recognize that this is not a one-time analysis. Your business, product vendors, and cloud providers (including VMWare with its AWS offering) change constantly. So continuous cost and resource analysis is needed to identify and re-optimize your cost structure when:
- workload performance and resource demands change
- new on-prem resources or cloud offerings are made available
- new workloads are created and need to be provisioned
- new pricing becomes available (for on-prem hardware and cloud offerings)
“Should I go to the cloud?” is no longer a pressing, elusive question to answer. The CloudPhysics platform can answer that question now and ongoing, adding new dimensions to your job of designing and managing IT from a financial perspective. Go get it here or visit us this week at VMworld Booth 1006.