1) Understand what TCO really means
Before embarking on your cloud transformation journey, one of the first steps you should undertake to calculate the overall TCO is to look at it holistically and not just focus on its benefits from a technology perspective.
Too many times, we have seen customers rely solely on tools to make this decision. Unfortunately, most TCO tools will provide a good firsthand understanding of the benefits (or lack of it, based on your use case) only from a technology standpoint.
However, they miss providing the big picture inclusive of the non-technology aspects. Hence, it is essential to identify key people and process-based benefits upfront.
2) Not all benefits are quantifiable
Elaborating on the point stated above, understand that not all benefits that contribute to TCO will be measurable. Cloud providers provide far better reliability, ease of procurement, usage, and a variety of deployment models than physical data centers. Adapting your product to a fast-paced digital landscape, failing fast and moving on, which is often an essential ingredient to building innovative products, is often much easier to achieve on the cloud than with traditional setups.
Note, however, that it is equally important to account for challenges that an organization will encounter in its cloud transformation journey. This greatly depends on the present level of cloud maturity that an organization is at, but is an important consideration all the same.
3) Scrutinize the existing environment
A detailed evaluation of the on-premises environment is a vital step to arrive at the right TCO. Compute, storage, networking are the foundations, but there are other factors that need to be considered. Suitability of existing workloads to work on the cloud, legal and regulatory issues, privacy concerns, existing licensing considerations, non-supported platforms on the cloud, integration with other systems, applications or physical hardware, data transfer rates, latency-sensitive applications or need for high, sustained I/O, etc. are all valid concerns. Whereas there may be ways to address or mitigate most of this on the cloud, your architecture, deployment model, or even choice of cloud provider may be dictated by these constraints and, in turn, will have an impact on the TCO.
4) Understand that TCO is a journey
Cloud is a moving target and various cloud-computing providers are constantly coming up with innovative services and costs. Inferred instance discounts, sustained use discounts, hybrid fleet instances, service price drop, changes in billing strategies (for instance, moving to per-second billing) are a few among the numerous changes cloud providers are coming up with to stay competitive. Look out for these and plan optimizations wherever possible, which in turn will affect your overall TCO.
A related aspect is comparing competing services in different deployment models such as PAAS, IAAS, and serverless. Whereas one may seem cheaper than others on face value, you need to identify which model will be cost-effective for your setup in the longer run. Besides, don’t let cost be the only deciding factor. Managed service providers might help address issues around maintenance, patching, backing, availability that are common while using IAAS. Some other services may introduce lock-in risks, which may be acceptable to some organizations but not to others.
5) Make “Smart” TCO Calculations
When using tools to calculate the overall TCO, it is crucial to consider the practical usage patterns. This is especially important because one of the key drivers for improving TCO on the cloud is to embrace and work around the elasticity that the cloud platform provides. Often, with on-premises hardware, costs are calculated assuming consistent usage patterns. Not because the usage patterns are themselves linear but because it is difficult to provision or de-provision hardware – a classic over-provisioning issue. On the cloud, however, understanding of usage patterns over a period, understanding peak loads versus non-peak loads, patterns used to adapt to these changing conditions and, most importantly, calculating TCO based on these inputs and not assuming linear usage patterns is important.
Most of the tools may not provide this out of the box so it is up to you to use them effectively.
A related issue is the smart classification of workloads, identifying what needs to be running when, and accounting for it while calculating TCO.
To conclude, it is important to understand that TCO on the cloud is not a one-time effort. Moreover, it greatly depends on how well workloads are optimized to work on the cloud. Often this can be a paradigm shift for teams that are used to designing products a certain way. However, it is well worth investing in this shift!
About the author
Viraj Nadkarni is a Solutions Architect within the Cloud business unit at Persistent. His focus is on helping customers with their cloud transformation journey.