Skip to main content
Documentation - HPE Usage and Analytics

Optimizing your cloud cost with What-If modeling

  1. Last updated
  2. Save as PDF

This topic explains how to optimize your cloud cost using the HPE Consumption Analytics platform.

Using What-If to model changes

The HPE Consumption Analytics portal's What-If feature allows you to model changes to your VMs and see the impact to your cost, comparing the current state of your workload against alternate states. You can create a What-If model for each decision you want to explore, such as downsizing over-provisioned VMs, moving to a different region, or even switching cloud providers.  What-if compares your actual results to what they would have been for the same period, assuming different options.

To use What-If, you choose a period of actual usage to date, choose a specific set of VMs to model, change something about those VMs (region, size, etc.) to create an alternate state, then compare what your usage actually cost you to what it would cost you in that alternate state.

What-If supports Amazon Web Services and Microsoft Azure, but not HPE GreenLake or Google Cloud Platform.

You can model the cost of your workload in a public cloud even if you’re not yet using that cloud provider. The HPE Consumption Analytics portal regularly retrieves rates from both AWS and Azure, which means your model is always using up-to-date rates.

If you want to share a What-If model with others you can export it to a CSV file.

To use the What-If feature, you must have the Manage What-if permissions capability. See Permissions.

Understanding the What-If results

Before you create a What-If model, it's helpful to understand an example.  Let's look at the following:
WhatIf_Overview.png

1

The baseline period of usage is shown in the upper left of the page:

  • Baseline Period: The period of usage. The cost for your alternate scenario is calculated against this baseline. 
  • End Date: This is the final day of the baseline period. For example, if you set the End Date to April 6 of the current year and your Baseline Period to Previous 30 Days, the period will be March 8, 2017 through April 6, 2017.

2

An overview of your model's changes is shown in the upper right of the page:

  • VMs Changed: How many VMs you changed in the Scenario.
  • Cost Difference: The cost decrease or increase from the Baseline to the Scenario.
  • New Cost Total: The total cost for all your resources for the Scenario.

3

Resource panel: Shows the VMs from all of your AWS and Azure public cloud collections. Note this does not include Google Cloud Platform. The default columns shown are needed for What-If modeling, including the resource name and, if needed, an ID that uniquely identifies the VM.

Note you can customize the table fields in the Tools panel. Toggle Tools with the pencil icon.

4

Baseline: Shows the specific service the VM is an instance of and the VMs actual cost for the baseline period.

5

Scenario: Shows the service you want to model. The baseline represents what is and the scenario represents what that would become.

Creating What-If models

Since What-If is based on the HPE Consumption Analytics platform’s tabular reports, you have most report authoring tools available to create a meaningful What-If model (for more information, see Reports.)  For example, if you wanted to downsize over-provisioned VMs, you might start by sorting the table by CPU utilization percentage, to bring the least-used VMs to the top. You might also filter out VMs that are irrelevant to the current model, add contextual columns to identify the VMs that support a particular application or project, and group on a column to work with a segment of VMs at a time.

To create a What-If model

  1. In the left menu, click What-If. To create a new model, click the + icon in the upper right of the screen, or if modifying an existing model, click the model's name.
  2. Set the Baseline Period for comparison purposes. This is the period of actual usage against which the cost for your alternate scenario will be calculated.
  3. Select the desired End Date from the calendar dropdown.
  4. Select one or more VMs that you want to model in your What-If scenario, and click Edit Selected. The Select New Service window appears.

 WhatIf_SelectNewService.png

  1. Select a cloud provider: Amazon Web Services or Microsoft Azure. The available regions, operating systems, and services appear.
  2. Select the region and operating system.
  3. Select the specific service you want, or use the search to narrow the list. You can select any service offered by Amazon Web Services or Microsoft Azure. The service catalog marks the baseline service in gray so you can compare it against others.
  4. Click Save.

The scenario updates to show:

  • Your selection
  • Its cost for the baseline usage quantity
  • The difference versus the baseline cost

When you’ve finished creating a scenario that represents your alternate state you can save the model for future reference or update.

  • The values in the scenario cost column can change in a saved model, because HPE Consumption Analytics platform keeps them up-to-date with changing public cloud rates.
  • If you have an Azure collection, What-If uses your discounted EA rates to calculate the cost for your scenario, not the general pay-as-you-go (PAYG) rates. This makes your model more accurate.  If you don't have an Azure collection, What-If uses the PAYG rates.
  • The cost for any AWS VM in your scenario does not take into account any reserved instances (RIs) you own. This is because What-If models changes to a specific VM, and cannot know whether one of your RIs will be applied to a given VM vs. some other VM. Therefore, depending on your RI ownership, your cost for AWS VMs may appear high, since the baseline reflects the cost you actually paid, lowered by RIs.

Video overview

The following video is a high-level introduction to using What-If: