As many organizations heavily adopted cloud services over the past few years, optimizing cloud costs is becoming a priority. The complexity of cloud pricing and pricing options leaves IT teams into uncertainty: is our cloud consumption really optimized? Or are we burning money because we don't identify the right-sizing opportunities?
As part of established FinOps practices, the right-sizing potential of all managed cloud resources has to be assessed continually, in order to save costs. Utilization metrics over time, in particular, are good insights to identify right-sizing opportunities.
Txture helps you succeed in your right-sizing initiatives: both during cloud migration planning coming from on-premises, and for the analysis of your existing cloud portfolio across cloud service providers (CSPs) like AWS, Azure or Google Cloud and others.
The Txture platform already helps you identify right-sizing opportunities based on utilization data (see for instance its functionality around configurable right-sizing strategies).
But it goes far beyond by factoring in typical performance scores (for instance SPEC CPU® 2017) of the different compute processors equipped to run e.g. cloud compute or managed cloud databases. Txture maintains benchmark scores across AMD, Intel and some ARM based chips. In the platform, when evaluating a cloud service, you have clear insights into what chip (or chip family) is used, as well as its performance score (See Figure 1).
Figure 1. AMD backed C3 compute product instance from Google Cloud, with information about the underlying CPU used in the platform.In the release version 43 of Txture, we have now introduced a cloud target architecture setting that combines our knowledge of CPU models backing compute related cloud resources and their individual performance data.
Activating the so-called "processor performance based right sizing" helps to identify and simulate change opportunities for saving cost and carbon emissions by upgrading to cloud products that are using more modern and efficient processors (see Figure 2).
With this new mode active, Txture no longer generates candidate cloud products solely based on resource requirements from your as-is estate (i.e. number of cores/threads, amount of RAM), but rather establishes candidate cloud products by considering performance comparison tables between all the different processor models out there in the market.
Figure 2. Target architecture preference for activating the new processor performance based right sizing mode.Figure 3 shows a part of a Txture generated Bill of Materials, both showing classic resource and utilization based right sizing, but also the selection of a target candidate VM based on the new processor performance based right sizing mode enabled.
Figure 3. Example single line item of a Txture generated BOM showing the effect of classical and the new performance based right sizing mode, including savings to be achieved.Beyond the newly added functionality we described in this article, we want to point out that you can further control the selection of compute based products with Txture's general processor preference (see Figure 4).
It allows you to tell Txture to look for more cost effective options in contrast to going fully performance oriented or otherwise deciding for a more balanced approach. This setting works in both scenarios where processor performance based right sizing is active or has been disabled.
Figure 4. Txture offers also a general preference to control how compute related cloud services are selected.Processor performance based right sizing is a great way to save on cost and carbon across you managed application portfolio. With Txture you can continually evaluate optimization and modernization potentials and keep track over time of how your portfolio performs.
Want to learn more about the Txture platform? Reach out to us, we'll be happy to help!