Purchasing Amazon Web Services (AWS) reserved instances (RIs) can be a smart cost-saving option versus buying on-demand instances. When you have sustained usage of specific instance types in a specific region and availability zone over a year or more, the cost of a reserved instance can be up to 65 percent lower than on-demand prices. AWS recently slashed both the upfront and hourly rates on 1-year and 3-year reserved instances, which we detailed in a blog post along with cost comparison charts of AWS vs. Google pricing.
However, anyone who has attempted to decide when and whether to buy reserved instances for a variety of instance types in different availability zones knows how difficult the process can be. In addition, your reserved instances purchases should be based on your forecasted future usage, not just your past usage.
RightScale Cloud Analytics helps you visualize, analyze, and optimize cloud costs across public and private clouds as well as virtualized environments such as VMware.
In this blog and the accompanying video, I will show you how Cloud Analytics enables you to optimize cloud costs by:
- Determining the optimal number of RIs based on your forecasts.
- Running what-if scenarios to calculate upfront and ongoing costs for RI purchases.
Determine the Number of RIs You Need
To accurately forecast the number of RIs to purchase, start by reviewing your past usage. With Cloud Analytics, you can use the Analyze feature to filter down by Instance Types, Data Center (AZ), and Platform (OS) to see how many of each instance you have been running over the past few weeks and months.
In the screenshot below, I am analyzing the last three months of usage and cost as well as the minimum number of instances running on a daily basis. I can see that the minimum number of instances I am running is nine. If this workload will continue at the same level over the next year, I can purchase nine RIs to cover that minimum level.
If you are using AWS Trusted Advisor, you may get recommendations on the number of RIs to purchase. You can check the validity of the recommendations by using Cloud Analytics.
Create What-If Scenarios to Calculate RI Costs and Savings
Next, I can build a future-looking scenario based on historical usage. This will allow me to adjust the forecast and make RI purchase decisions based on planned usage and costs. You can add as many different RIs options as you like to a scenario (Light, Medium, Heavy, 1 Year, and 3 Year) and see the effects of the reservations on your future costs. I have selected "Heavy RI" in the screenshot below:
From the Analyze page, I can use the Create button to evaluate my options through Scenario Builder. If needed, I can add growth patterns to adjust my usage forecast. I can then add RIs to the scenario and immediately see my upfront costs and my monthly costs. I can also see that my total costs will decline to $93,000 over the next year due to the purchase of 1-year RIs as compared to a run rate of around $213,000 in subsequent years where there are no RIs.
RightScale Cloud Analytics helps you create multiple what-if scenarios to forecast cloud usage and costs. You can identify how many reserved instances to purchase; evaluate the upfront and ongoing costs; and determine the total cost savings. With Cloud Analytics, you can optimize your AWS bill and realize significant savings on your cloud spend.
If you want to start analyzing the impact of RIs on your cloud spend right now, get a free trial of RightScale Cloud Analytics.