How to Determine Cloud TCO
March 7, 2013 – The cloud should not be viewed as a panacea for infrastructure ills or struggling IT, but there are ways to assess workloads ahead of time to find the most from your cloud investment, according to a report released today from Saugatuck Technology.
In “Understanding Cloud Infrastructure Costs: Navigating for Savings,” lead author and Saugatuck analyst Charlie Burns gives a framework for general expectations with various cloud offerings and workloads, with caution that a switch to cloud infrastructure does not automatically result in savings. From a high-level view, all true cost reductions and savings in a switch to the cloud are dependent on:
- Workload characteristics – the size of compute resources required for desired performance can determine options between public or private/dedicated clouds, for instance.
- Operational characteristics – pertaining to specific workloads and including characteristics such as average utilization of infrastructure resources (higher utilization results in a bigger cloud bill) and the need for high availability.
- In-house environment – efficiency and optimization of existing IT environments “establishes the baseline costs” for determining cloud savings.
“An important area of consideration when evaluating the TCO of any cloud IT infrastructure is how it compares to the efficiency and optimization of relevant, traditional IT infrastructure,” Burns said in the report. “While significant savings are achievable, we caution companies to deeply understand their workloads – as the savings very much depend on the pre-existing level of resource utilization and levels of optimization,” said Burns.
Historically, IT departments have found frustration in their cloud fact-finding ventures for two main reasons: configuration and pricing are inconsistent between providers, and workloads are not homogeneous. To see through much of that uncertainty caused by vendor messaging and the chatter over emerging tech, IT leaders must take a granular-level look at their workloads.
From Saugatuck’s view in the report, most of those workloads can be divided into these workload categories: Web/transactional services; database services; network or data transmission services; application services; and system management support. To make general comparisons of the costs of workloads in the cloud, Saugatuck assessed those workload categories along with a few assumptions on the variety of technical components, load deviations and existing efficiencies. Among those workloads and variables, the firm reported the following savings compared with those same workloads run in-house: businesses running smaller-sized workloads can regularly save 6 to 70 percent in a public cloud, a wide range that trended toward the higher percentage of savings, according to Saugatuck CEO and Founder Bill McNee; businesses running medium-sized workloads can save 10 to 35 percent using a hybrid cloud; and businesses running larger-sized workloads can save 5 to 20 percent using a private cloud.
After making any type of cloud deployment, a business should compare actual costs with projections, with results used as the basis for subsequent vendor evaluations and IT workload plans.
To register for a copy of the report, click here.