Multi-tenant edge cloud system power management
In one embodiment, an apparatus of an edge computing system includes memory that includes instructions and processing circuitry coupled to the memory. The processing circuitry implements the instructions to process a request to execute at least a portion of a workflow on pooled computing resources, the workflow being associated with a particular tenant, determine an amount of power to be allocated to particular resources of the pooled computing resources for execution of the portion of the workflow based on a power budget associated with the tenant and a current power cost, and control allocation of the determined amount of power to the particular resources of the pooled computing resources during execution of the portion of the workflow.
1. An apparatus of an edge computing system, comprising:
memory comprising instructions;
processing circuitry coupled to the memory to implement the instructions to:
process a request to execute at least a portion of a workflow on pooled computing resources, the workflow being associated with a particular tenant;
determine an amount of power to be allocated to particular resources of the pooled computing resources for execution of the portion of the workflow based on a power budget associated with the tenant and a current monetary cost of power, wherein determining the amount of power to be allocated comprises determining a power consumption associated with execution of the portion of the workflow based on a power consumption model for the pooled computing resources, the power budget specifying a monetary cost per unit time and the power consumption model specifying an estimated power consumption per unit data per unit time; and
control allocation of the determined amount of power to the particular resources of the pooled computing resources during execution of the portion of the workflow.
2. The apparatus of claim 1 , wherein the portion of the workflow comprises one or more of a network function and a service, and the power budget is specified based on at least one of a per-tenant, per-service, per-network function, per-multiple service, or per-multiple network function basis.
3. The apparatus of claim 1 , wherein controlling allocation of the determined amount of power comprises controlling allocation of power on at least one of a per-DIMM and a per-disk basis.
4. The apparatus of claim 1 , wherein controlling allocation of the determined amount of power comprises controlling allocation of power on a per-computing resource platform basis.
5. The apparatus of claim 1 , wherein the processing circuitry is further to configure at least one of a computing resource platform manager and a computing resource controller to implement the control of power allocation to the particular resources.
6. The apparatus of claim 1 , wherein controlling allocation of the determined amount of power comprises controlling an amount of power provided to the pooled computing resources during execution of the portion of the workflow.
7. The apparatus of claim 1 , wherein controlling allocation of the determined amount of power comprises controlling a bandwidth provided to the pooled computing resources during execution of the portion of the workflow.
8. The apparatus of claim 1 , wherein the processing circuitry is further to detect potential service degradation caused by execution of the workflow, wherein the potential service degradation is detected using a telemetry model that analyzes a level of power consumption for particular workflow operations over time.
9. The apparatus of claim 8 , wherein the processing circuitry is further to reject the request based on detection of a potential service degradation.
10. The apparatus of claim 1 , wherein the processing circuitry is further to detect that the portion of the workflow will not execute on the particular resources based on the determined amount of power to be allocated to the particular resources.
11. A non-transitory computer-readable medium comprising instructions that are, when executed by processing circuitry, to:
process a request to execute at least a portion of a workflow on pooled computing resources, the workflow being associated with a particular tenant;
determine an amount of power to be allocated to particular resources of the pooled computing resources for execution of the portion of the workflow based on a power budget associated with the tenant and a current monetary cost of power, wherein determining the amount of power to be allocated comprises determining a power consumption associated with execution of the portion of the workflow based on a power consumption model for the pooled computing resources, the power budget specifying a monetary cost per unit time and the power consumption model specifying an estimated power consumption per unit data per unit time; and
control allocation of the determined amount of power to the particular resources of the pooled computing resources during execution of the portion of the workflow.
12. The non-transitory computer-readable medium of claim 11 , wherein the portion of the workflow comprises one or more of a network function and a service, and the power budget is specified based on at least one of a per-tenant, per-service, per-network function, per-multiple service, or per-multiple network function basis.
13. The non-transitory computer-readable medium of claim 11 , wherein controlling allocation of the determined amount of power comprises controlling allocation of power on at least one of a per-DIMM and a per-disk basis.
14. The non-transitory computer-readable medium of claim 11 , wherein controlling allocation of the determined amount of power comprises controlling allocation of power on a per-computing resource platform basis.
15. The non-transitory computer-readable medium of claim 11 , wherein the instructions are further to configure at least one of a computing resource platform manager and a computing resource controller to implement the control of power allocation to the particular resources.
16. The non-transitory computer-readable medium of claim 11 , wherein controlling allocation of the determined amount of power comprises controlling an amount of power provided to pooled memory during execution of the portion of the workflow.
17. The non-transitory computer-readable medium of claim 11 , wherein controlling allocation of the determined amount of power comprises controlling an amount of power provided to pooled storage during execution of the portion of the workflow.
18. The non-transitory computer-readable medium of claim 11 , wherein controlling allocation of the determined amount of power comprises controlling a bandwidth provided to the pooled computing resources during execution of the portion of the workflow.
19. The non-transitory computer-readable medium of claim 11 , wherein the instructions are further to detect potential service degradation caused by execution of the workflow, wherein the potential service degradation is detected using a telemetry model that analyzes a level of power consumption for particular workflow operations over time.
20. The non-transitory computer-readable medium of claim 19 , wherein the instructions are further to reject the request based on detection of a potential service degradation.
21. A method for controlling power allocation to pooled computing resources of an edge computing system, comprising:
processing a request at an edge computing device to execute at least a portion of a workflow on pooled computing resources coupled to the edge computing device, the workflow being associated with a particular tenant;
determining an amount of power to be allocated to particular resources of the pooled computing resources for execution of the portion of the workflow based on a power budget associated with the tenant and a current monetary cost of power, wherein determining the amount of power to be allocated comprises determining a power consumption associated with execution of the portion of the workflow based on a power consumption model for the pooled computing resources, the power budget specifying a monetary cost per unit time and the power consumption model specifying an estimated power consumption per unit data per unit time; and
controlling allocation of the determined amount of power to the particular resources of the pooled computing resources during execution of the portion of the workflow.
22. The method of claim 21 , wherein the portion of the workflow comprises one or more of a network function and a service, and the power budget is specified based on at least one of a per-tenant, per-service, per-network function, per-multiple service, or per-multiple network function basis.
23. The method of claim 21 , wherein controlling allocation of the determined amount of power comprises controlling allocation of power on at least one of a per-DIMM and a per-disk basis.
24. The method of claim 21 , wherein controlling allocation of the determined amount of power comprises controlling allocation of power on a per-computing resource platform basis.
25. The method of claim 21 , wherein controlling allocation of the determined amount of power comprises one or more of: controlling an amount of power provided to the pooled computing resources during execution of the portion of the workflow and controlling a bandwidth provided to the pooled computing resources during execution of the portion of the workflow.