Technologies for dynamic accelerator selection
Technologies for dynamic accelerator selection include a compute sled. The compute sled includes a network interface controller to communicate with a remote accelerator of an accelerator sled over a network, where the network interface controller includes a local accelerator and a compute engine. The compute engine is to obtain network telemetry data indicative of a level of bandwidth saturation of the network. The compute engine is also to determine whether to accelerate a function managed by the compute sled. The compute engine is further to determine, in response to a determination to accelerate the function, whether to offload the function to the remote accelerator of the accelerator sled based on the telemetry data. Also the compute engine is to assign, in response a determination not to offload the function to the remote accelerator, the function to the local accelerator of the network interface controller.
1. A cloud service provider system for use in providing at least one service in association with at least one node via at least one network, the cloud service provider system comprising:
accelerator circuitry comprised in multiple accelerators in the at least one network, the multiple accelerators comprising one or more certain accelerators that are remote from the at least one node; and
server circuitry configurable to dynamically assign and/or reassign for acceleration, at least in part, based upon physical location information of the multiple accelerators and telemetry data, at least one workload to and/or from at least one of the multiple accelerators;
wherein:
the at least one workload is associated with the providing of the at least one service;
execution of the at least one workload is to be associated with at least one container and/or virtual machine;
the cloud service provider system comprises resources;
the resources comprise the accelerator circuitry;
the server circuitry is configurable to dynamically allocate and/or dynamically deallocate, based upon current resource usage data, resource utilization trend data, and predicted future resource utilization data, at least one portion of the resources of the cloud service provider system for the execution, at least in part, of the at least one workload; and
the server circuitry is configurable to generate the current resource usage data, the resource utilization trend data, and the predicted future resource utilization data based, at least in part, upon the telemetry data.
2. The cloud service provider system of claim 1 , wherein:
the server circuitry is to determine accelerator failure based upon the telemetry data.
3. The cloud service provider system of claim 2 , wherein:
the resources comprise physical resources.
4. The cloud service provider system of claim 3 , wherein:
the server circuitry is configurable to dynamically assign and/or reassign for the acceleration, at least in part, the at least one workload to and/or from the at least one of the multiple accelerators, in association with achievement of quality of service and/or latency considerations.
5. The cloud service provider system of claim 3 , wherein:
the server circuitry is configurable to dynamically allocate and/or dynamically deallocate, the at least one portion of the resources of the cloud service provider for the execution, at least in part, of the at least one workload, in association with achievement of quality of service and/or latency considerations.
6. The cloud service provider system of claim 3 , wherein:
the at least one workload is associated with machine learning; and
the accelerator circuitry comprises graphics processing unit hardware.
7. A method implemented using a cloud service provider system, the cloud service provider system to be used in providing at least one service in association with at least one node via at least one network, the cloud service provider system comprising accelerator circuitry and server circuitry, the accelerator circuitry being comprised in multiple accelerators in the at least one network, the multiple accelerators comprising one or more certain accelerators that are remote from the at least one node, the method comprising:
dynamically assigning and/or reassigning, by the server circuitry, for acceleration, at least in part, based upon physical location information of the multiple accelerators and telemetry data, at least one workload to and/or from at least one of the multiple accelerators;
wherein:
the at least one workload is associated with the providing of the at least one service;
execution of the at least one workload is to be associated with at least one container and/or virtual machine;
the cloud service provider system comprises resources;
the resources comprise the accelerator circuitry;
the server circuitry is configurable to dynamically allocate and/or dynamically deallocate, based upon current resource usage data, resource utilization trend data, and predicted future resource utilization data, at least one portion of the resources of the cloud service provider system for the execution, at least in part, of the at least one workload; and
the server circuitry is configurable to generate the current resource usage data, the resource utilization trend data, and the predicted future resource utilization data based, at least in part, upon the telemetry data.
8. The method of claim 7 , wherein:
the server circuitry is to determine accelerator failure based upon the telemetry data.
9. The method of claim 8 , wherein:
the resources comprise physical resources.
10. The method of claim 9 , wherein:
the server circuitry is configurable to dynamically assign and/or reassign for the acceleration, at least in part, the at least one workload to and/or from the at least one of the multiple accelerators, in association with achievement of quality of service and/or latency considerations.
11. The method of claim 9 , wherein:
the server circuitry is configurable to dynamically allocate and/or dynamically deallocate, the at least one portion of the resources of the cloud service provider for the execution, at least in part, of the at least one workload, in association with achievement of quality of service and/or latency considerations.
12. The method of claim 9 , wherein:
the at least one workload is associated with machine learning; and
the accelerator circuitry comprises graphics processing unit hardware.
13. At least one non-transitory machine-readable storage medium storing instructions to be executed by at least one machine that is to be associated with a cloud service provider system, the cloud service provider system to be used in providing at least one service in association with at least one node via at least one network, the cloud service provider system comprising accelerator circuitry and server circuitry, the accelerator circuitry being comprised in multiple accelerators in the at least one network, the multiple accelerators comprising one or more certain accelerators that are remote from the at least one node, the instructions, when executed by the at least one machine, resulting in the cloud service provider system being configured to perform operations comprising:
dynamically assigning and/or reassigning, by the server circuitry, for acceleration, at least in part, based upon physical location information of the multiple accelerators and telemetry data, at least one workload to and/or from at least one of the multiple accelerators;
wherein:
the at least one workload is associated with the providing of the at least one service;
execution of the at least one workload is to be associated with at least one container and/or virtual machine;
the cloud service provider system comprises resources;
the resources comprise the accelerator circuitry;
the server circuitry is configurable to dynamically allocate and/or dynamically deallocate, based upon current resource usage data, resource utilization trend data, and predicted future resource utilization data, at least one portion of the resources of the cloud service provider system for the execution, at least in part, of the at least one workload; and
the server circuitry is configurable to generate the current resource usage data, the resource utilization trend data, and the predicted future resource utilization data based, at least in part, upon the telemetry data.
14. The at least one non-transitory machine-readable storage medium of claim 13 , wherein:
the server circuitry is to determine accelerator failure based upon the telemetry data.
15. The at least one non-transitory machine-readable storage medium of claim 14 , wherein:
the resources comprise physical resources.
16. The at least one non-transitory machine-readable storage medium of claim 15 , wherein:
the server circuitry is configurable to dynamically assign and/or reassign for the acceleration, at least in part, the at least one workload to and/or from the at least one of the multiple accelerators, in association with achievement of quality of service and/or latency considerations.
17. The at least one non-transitory machine-readable storage medium of claim 15 , wherein:
the server circuitry is configurable to dynamically allocate and/or dynamically deallocate, the at least one portion of the resources of the cloud service provider for the execution, at least in part, of the at least one workload, in association with achievement of quality of service and/or latency considerations.
18. The at least one non-transitory machine-readable storage medium of claim 15 , wherein:
the at least one workload is associated with machine learning; and
the accelerator circuitry comprises graphics processing unit hardware.
19. At least one data center for use in association with at least one node, the at least one data center to be used in a cloud service provider system, the cloud service provider system to be used in providing at least one service, the at least one data center comprising:
at least one network;
multiple accelerator nodes and multiple server nodes communicatively coupled together via the at least one network, the multiple accelerator nodes comprising respective accelerator circuitry, one or more of the multiple accelerator nodes being remote from the at least one node, the multiple server nodes comprising server circuitry, the server circuitry being configurable to dynamically assign and/or reassign for acceleration, at least in part, based upon physical location information of the multiple accelerator nodes and telemetry data, at least one workload to and/or from at least one of the multiple accelerator nodes;
wherein:
the at least one workload is associated with the providing of the at least one service;
execution of the at least one workload is to be associated with at least one container and/or virtual machine;
the cloud service provider system comprises resources;
the resources comprise the respective accelerator circuitry;
the server circuitry is configurable to dynamically allocate and/or dynamically deallocate, based upon current resource usage data, resource utilization trend data, and predicted future resource utilization data, at least one portion of the resources of the cloud service provider system for the execution, at least in part, of the at least one workload; and
the server circuitry is configurable to generate the current resource usage data, the resource utilization trend data, and the predicted future resource utilization data based, at least in part, upon the telemetry data.
20. The at least one data center of claim 19 , wherein:
the server circuitry is to determine accelerator failure based upon the telemetry data.
21. The at least one data center of claim 20 , wherein:
the resources comprise physical resources.
22. The at least one data center of claim 21 , wherein:
the server circuitry is configurable to dynamically assign and/or reassign for the acceleration, at least in part, the at least one workload to and/or from the at least one of the multiple accelerator nodes, in association with achievement of quality of service and/or latency considerations.
23. The at least one data center of claim 22 , wherein:
the server circuitry is configurable to dynamically allocate and/or dynamically deallocate, the at least one portion of the resources of the cloud service provider for the execution, at least in part, of the at least one workload, in association with achievement of quality of service and/or latency considerations.
24. The at least one data center of claim 23 , wherein:
the at least one workload is associated with machine learning; and
the accelerator circuitry comprises graphics processing unit hardware.
25. The at least one data center of claim 19 , wherein:
the at least one data center comprises multiple data centers.