In existing cloud environments it is not possible to mix, on the same server at the same time, workloads that use part of a processor core, or that use cores on a best-effort basis, with workloads that must both be assigned to a single core and have that core dedicated to their use (i.e., nothing else runs on the core). To address these challenges and inefficiencies, techniques are presented herein that support a division of resources in a way that they can then be appropriately assigned to workloads. One logical pool of cores may be assigned for workloads requiring shared resources and another pool may be assigned for workloads requiring dedicated resources. The boundary between those pools may shift dynamically as, for example, additional resources are required.

Creative Commons License

Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.