January 12, 2017
Cloud Smart Series: How to Model Future Cloud Costs
Cloud has been around long enough now that everyone can easily recite its many benefits…agility, scalability, self-service, and of course the fact that you only have to pay for what you use. However, there’s also the flip-side of those benefits that businesses tend to worry about, things like security, reliability, and the predictability of costs in a pay-per-use model.
While having a solid forecasting capability is one way to plan for future costs, there are other scenarios that are more difficult to accurately predict.
- Moving services to a different cloud provider. One of the great advantages of public cloud is the avoidance of vendor lock-in. In fact, a recent survey by RightScale found that, on average, businesses leverage 3 public clouds to test or run their apps. For some companies, selecting their public cloud of choice is influenced primarily by features and functionality, but for others, it’s all about saving money. But how do you know just how much you stand to gain or lose by moving your workloads to a different provider?
- Upsizing or downsizing workloads. It’s hard to predict just how much to provision for a given workload. Do you need an A1 instance or an A2? How about S3 storage or EBS? What about the impact of moving your apps from QA to production? Being able to predict the impact of future moves based on how you’re using your cloud resources today greatly improves your ability to both plan for future cost increases, as well as opportunistically reap the benefits of lower costs. But combing through your bills and spreadsheets and trying to model your future costs against the pages and pages of services and rates on AWS and Azure is a huge endeavor…and frankly, not a fun one.
- Changing service regions. Public cloud providers offer cloud services in multiple regions or availability zones around the globe. Being able to select a region or zone closer to your customers helps reduce latency and addresses requirements for data redundancy, failover, or even data sovereignty. But different regions also have different price tags associated with them so it’s important to consider the cost impact of different regions when provisioning services.
At Cloud Cruiser, we know that the most accurate cost predictions are based on our customers’ historical usage patterns. Rather than guessing how a particular decision will impact your bottom line, we show you what your costs would be for your chosen scenario using our latest What-If™ cost modeling feature. All you have to do is choose the service(s) you wish to model, the historical time period you’d like to base it on, and the provider, service type, and/or region that you’re interested in. Create one scenario or a hundred – it’s up to you! So let’s see how it’s done.
1. Click on WHAT-IF on the left navigation.
2. Click on the ‘+’ sign in the upper right corner to create a new scenario.
3. Provide a name for your new scenario.
4. The default What-If scenario displays qualified services from AWS and Azure and is prepopulated with fields that provide information relevant to your decision-making process, such as Provider Name, Resource Type, OS, Region, and Cost. Using the report authoring pane, you can add any additional fields you need to provide additional context for your scenario.
5. Select the services that you would like to use for your scenario. For example, you might want to select all of the services you have in Dev or QA if you are creating a scenario to model the cost of moving these services to larger VMs for Production. You can use the Filter option in the report authoring pane if you have a lot of services to sort through, or you can sort on any of the columns by selecting the down arrow.
You can also select the timeframe that you would like to base your cost model on. The timeframe selected should be a good representation of the expected future usage. If, for example, your consumption patterns are cyclical (eg. heavy month-end processing or seasonal spikes), you’ll want to select a timeframe that captures those patterns for the most accurate model.
In this example, we are selecting all of the AWS m4.large services running in production. Click on the services you’d like to select or select a group of services by using ctrl-alt-shift and highlighting the desired services.
6. Click on EDIT SELECTED button to start.
7. A window will pop up with all of the most current rates for AWS and Azure. By default, the displayed rates will be for the same filters currently assigned to the services you are modeling. In this example, those filters are for the US West region and Linux operating system. The counter in the blue circle will display the number of qualifying services.
8. You can filter the types of services you want to consider by selecting a different region or operating system from the pull-down menus or use the Search bar to type in a custom search.
When you scroll down the list of rates, you’ll see the service(s) that you are already using highlighted in grey with a count of the number of each type of this service you have.
In this scenario, we are going to select an m4.xlarge service to see what the impact of this move will be. Click on the rate card to select it and click Save to keep your selection.
NOTE: If you are using AWS Reserved Instances (RI), you will see the rate that you are actually paying for the service, which will be significantly lower than Amazon’s published rates. Keep this in mind when comparing an RI against a non-RI rate.
9. You can now see that moving the selected production VMs to a larger size will increase monthly costs by about $66 for an overall monthly cost for the selected VMs of $1,477.84.
10. You can export your What-If scenario to a CSV in order to share with others or perform additional analysis in Excel. When you’re done, click the disc icon to save your What-If scenario for future reference within the Cloud Cruiser app. And remember, you can create as many cost models as you like and view them at any time from the main What-If landing page.
Being able to accurately predict the impact of a future state based on your own historical usage patterns is an invaluable tool to have in your cost management arsenal. And the value to your business?
- Fact-based decisions to support your cloud investments
- Accurate cost projections for more reliable budgets
- Cost savings for underutilized or non-critical services
- Cloud Adoption LIfecycle
- Cloud Analytics
- Cloud Financial Management
- Customer Success
- Finance Corner
- Hybrid IT Service Delivery
- Penny's Finance Corner
- Product Announcements
- Professor Andy
- Tips & Tricks from the Field