The volume of factors involved in premium capacity utilization makes it difficult to forecast the amount of premium capacity (and thus cost) required. This complexity is particularly acute for large deployments with diverse use cases to support. Additionally, for organizations relatively new to Power BI, the level and growth of user adoption, as well as the requirements for future projects, can be unclear. Nonetheless, to provide an initial estimate of the cost of deploying Power BI with premium capacities, Microsoft has developed the Power BI Premium calculator. This online forecasting tool provides free estimates of the combined monthly cost of Power BI Pro licenses and Power BI Premium capacity.
In the following screenshot of the Power BI Premium calculator, the user has specified 2,500 total Power BI users and used the two slider bars to distribute those users among the following profiles—Pro Users, Frequent Users, and Occasional Users:
In the example, 170 Pro Users results in 170 Pro licenses costing a total of $1,700 per month (170 * $10/mo). More importantly, the calculator estimates that two P1 nodes (8 v-cores) of Premium capacity will be required for the organization at a cost of $9,990 per month (2 * $4,995/mo). Therefore, an organization could use $11,690 per month as an initial and high-level estimate of their Power BI deployment. The Power BI Premium calculator is accessible at the following URL http://bit.ly/2eKil1I.
Thankfully the P1 month-to-month SKU eliminates the need for organizations to make large financial commitments in advance of actual usage. The 8 v-cores provided by a Premium P1 capacity (month-to-month) can be acquired in the Office 365 admin center as shown in the following screenshot:
In a fiscally conservative approach, an organization could create a single P1 capacity and test this capacity against different workloads until both the usage metrics described in the previous chapter and actual load test experiences suggest additional capacity is needed. Load testing of premium capacity typically consists of multiple users simulating normal user behaviors in the Power BI service such as clicking a slicer, pausing to view the results of a selection, and then making another filter selection. Additional options and guidance for load testing premium capacity are included in the Power BI Premium Capacity Planning and Deployment Whitepaper available at the following URL http://bit.ly/2Hu57DK.
Once it's determined that additional capacity is needed, an organization can simply provision another instance of a P1 (month-to-month) thus obtaining 16 total v-cores. In the following image from the Purchase services page of the Office 365 admin center, the quantity of an existing P1 (month-to-month) SKU can be revised:
Selecting the ellipsis (...) at the bottom of the Premium SKU exposes the Change license quantity option. Once the number of premium instances has been increased and submitted, the additional v-cores will be available in the Capacity settings page of the Power BI admin portal. Creating and managing premium capacities, including the assignment of app workspaces to premium capacities, was described in Chapter 12, Administering Power BI for an Organization.
With both v-core pooling and single-click scale up capabilities available in the Power BI admin portal it's very easy to reallocate premium capacity. In this example with 16 total v-cores purchased, an existing P1 capacity could be scaled up to a P2 capacity. Alternatively, a new and isolated P1 capacity could be created thus leaving two P1 capacities to handle the distinct workloads generated by different app workspaces.
The three-step process of purchasing additional v-cores, scaling up existing capacities or configuring new capacities, and testing workloads against premium capacities by assigning workspaces to those capacities will be repeated throughout the deployment life cycle. In the event that usage or resource needs decline, the monthly price for a P1 month-to-month instance can be avoided thus reducing the amount of v-cores available.
The flexibility of premium capacity and other cloud services such as Azure Analysis Services naturally aligns with the frequently changing needs of BI projects. The following sections describe both AAS and SSAS as a primary tool to complement Power BI Premium and to support large-scale enterprise deployments of Power BI.