RightScale Blog

Cloud Management Blog
RightScale Cloud Analytics Free Trial
Cloud Management Blog

RightScale PlanForCloud Analyzes $1 Billion in Cloud Cost Forecasts

It’s a paradox: Forecasting enterprise cloud costs is a hard task to do well, yet understanding all the costs up front is one of the main requirements of a successful cloud initiative. Last year, we launched PlanForCloud to help enterprises adopt cloud computing and ensure success in their early cloud projects. Today, we are excited to announce that PlanForCloud has helped forecast more than $1 billion in cloud spend — $1,016,619,975, to be exact.

From the large and growing number of organizations that have been planning their deployments with PlanForCloud, we’ve been able to analyze data from more than 9,500 deployments and uncover a number of trends on planned cloud usage.

Growth in Cloud Cost Planning

The growing trend toward cloud adoption is mirrored by a strong increase in the forecasts run using PlanForCloud. In June, users forecast more than $155 million of cloud spend. That’s more than $5 million of cloud spend running through the simulation engine every day.

Breakdown of Cloud Spend

Many enterprises ask where they should be concentrating their cost optimization efforts. Data from PlanForCloud shows that the majority of cloud costs are associated with compute resources, accounting for 70 percent of the total bill. Storage is the second largest cost, accounting for 18 percent of the bill. Data transfer costs averaged 6 percent, and other costs, such as support and read/write transaction charges, account for the remaining 6 percent.

Adoption of AWS Reserved Instances

Amazon Web Services (AWS) was the first public cloud provider to offer multiple purchase options for cloud resources. For example, a Reserved Instance is an AWS-specific purchase option in which a user commits to using a specific resource type and pays up front to get a discounted hourly rate. Organizations that use Reserved Instances can save up to 65 percent compared to On-Demand prices. (Microsoft Azure also offers similar purchase options, although we do not yet support these in PlanForCloud.)

Looking at the period of commitment on AWS, we see that more than half of AWS forecasted usage involved Reserved Instances, while 46 percent involved On-Demand instances. Organizations that do not or cannot commit to using a resource for a period of time are potentially missing out on savings on their cloud costs.

It is interesting to note that even with the potential savings of up to 65 percent on using three-year Reserved Instances (vs. 45 percent using one-year Reserved Instances), only 20 percent of organizations are willing to commit to that timeframe. This may indicate that many early cloud projects have less than a three-year commitment, making shorter commitments on Reserved Instances a preferred option.

When purchasing Reserved Instances, users also have to commit to a utilization level. The higher the expected utilization, the greater the upfront cost and the lower the hourly cost of running those instances.

The majority of users who commit to Reserved Instances go all-in and commit to the highest level — heavy utilization. With this level of utilization, regardless of how long an instance actually runs for, it will be charged as if it is running 24/7 for the duration of the commitment (one or three years).

Usage of Instance Types 

The instance types that organizations consider give an indication of the types of workloads and application sizes that enterprises plan to run on the public cloud.

PlanForCloud forecasts show a relatively even distribution between Small, Medium, Large, and Extra Large instances, which together make up 83 percent of all instance types. While a concentration on the mid-range instances is expected, there may be opportunities for companies to consider the use of larger or smaller instance sizes for some workloads.

Note: We calculated instance type mappings from the different public cloud providers based on RightScale Compute Unit (RCU) categories. RCUs abstract compute usage across public and private clouds to provide a rate that is roughly equivalent, across different instance types on different clouds. So, for example, an AWS m1.small instance and a Rackspace 1GB server are categorized as Small.

It will be interesting to see how these figures change over the next year as more enterprises move applications to the cloud.

I hope that these insights help you with your cloud strategy. Try out PlanForCloud to forecast your future cloud costs, and get in touch if you have any feedback or feature requests.

Post a comment