Simplifying capacity management with unified v-cores
We’re excited to announce a platform update to simplify the way customers manage Power BI Premium capacities. Starting on December 11th 2022, we’ll be rolling out changes to unify the concepts of front-end and back-end virtual cores for capacity management. Unification of these concepts addresses the feedback we’ve heard from customers who sometimes misinterpret the concepts of customer-managed v-cores from Microsoft managed v-cores. This change will simplify the model used by customers to both manage capacities and rationalize usage metrics against capacity sizing decisions.
How do Premium capacity v-cores work today?
Power BI Premium has historically used v-cores as the unit of measure that customers use to reserve compute resources when purchasing a capacity. After a capacity is purchased, its resources can then be used to run a variety of Power BI workloads. Before this change, purchased v-cores in a reserved capacity would be split evenly between front-end cores used by Microsoft for service delivery and back-end cores used by customers to power workload execution. In this model, capacity Administrators only manage back-end v-cores which account for 50% of the purchased v-core capacity.
How are we simplifying our approach to capacities?
The Capacities Team is on a mission to deliver scalable cloud compute with the simplicity of SaaS experiences to our customers. Unification of front-end and backend v-cores for capacity management is another step in the platform evolution to simplify capacity management while offering nearly limitless scale and flexibility. With this platform update, capacity Administrators can purchase and manage resources using a single v-core pool for each capacity.
What does this change mean for my existing capacities?
In order to make this change as seamless as possible for customers, we are enacting changes to both the capacity platform and usage reporting experiences. The amount of effective “throughput” on each capacity SKU will remain the same before and after these changes. Updates will be made to scale usage reporting and adjust for the unification of v-cores.
To better describe the implications of the change we’ll examine a before and after example.
Capacity Administrators will see these platform changes reflected in the Capacity Metrics Application. Timepoint drill views in the Metrics App will now show:
- Updated values for available compute on each SKU
- Adjusted usage reporting metrics
The percentage of capacity utilization for a given workload will not change because of this update.
- Gen2 platform architecture and scalability improvements
- The benefits of our updated smoothing policies for interactive and background processes to simplify capacity management
- A deep dive on how workload evaluation occurs in capacities and its impact to auto-scale and throttling
Comments
Post a Comment