Technologies for determining and storing workload characteristics

Abstract

Technologies for determining and storing workload characteristics include an orchestrator server to identify a workload to be executed by a managed node, obtain a profile associated with the workload, wherein the profile includes a model that relates an input parameter set indicative of one of more characteristics of the workload with an output parameter set indicative of one or more aspects of resources to be allocated for execution of the workload, determine, as a function of the input parameter set and the model, resources to allocate to the managed node to execute the workload, and allocate the determined resources to the managed node to execute the workload. Other embodiments are also described and claimed.

1 . An orchestrator server to manage workload profiles, the orchestrator server comprising:
 one or more processors;   one or more memory devices having stored therein a plurality of instructions that, when executed by the one or more processors, cause the orchestrator server to:
 identify a workload to be executed by a managed node; 
 obtain a profile associated with the workload, wherein the profile includes a model that relates an input parameter set indicative of one or more characteristics of the workload with an output parameter set indicative of one or more aspects of resources to be allocated for execution of the workload; 
 determine, as a function of the input parameter set and the model, resources to allocate to the managed node to execute the workload; and 
 allocate the determined resources to the managed node to execute the workload. 
   
     
     
         2 . The orchestrator server of  claim 1 , wherein the plurality of instructions, when executed, further cause the orchestrator server to:
 receive telemetry data indicative of resource utilization and workload performance as the workload is executed;   determine, as a function of the telemetry data, whether one or more threshold objectives are satisfied by the execution of the workload with the allocation of resources;   adjust, in response to a determination that the one or more threshold objectives are not satisfied, the allocation of resources to the managed node to satisfy the one or more threshold objectives; and   adjust the model to include the adjustment to the allocation of resources.   
     
     
         3 . The orchestrator server of  claim 2 , wherein to adjust the model comprises to adjust the model to produce an output parameter set of resources that represents the adjusted allocation of resources, in response to the input parameter set. 
     
     
         4 . The orchestrator server of  claim 2 , wherein to adjust the model comprises to adjust the output parameter set to change one or more architecture features of the resources to be allocated, wherein the architecture features include one or more of support for an extended instruction set, support for preloading of processor cache with one or more predefined values, support for accelerated cryptographic operations, and support for accelerated data compression operations. 
     
     
         5 . The orchestrator server of  claim 2 , wherein the plurality of instructions, when executed, further cause the orchestrator server to:
 generate, as a function of the telemetry data, landscape data indicative of conditions across the set of managed nodes in a data center; and   wherein to determine whether the threshold objectives are satisfied comprises to determine whether the threshold objectives are satisfied based additionally on the landscape data.   
     
     
         6 . The orchestrator server of  claim 2 , wherein the plurality of instructions, when executed, further cause the orchestrator server to:
 obtain resource allocation objective data indicative of one or more thresholds to be satisfied during the execution of the workload; and   wherein to determine whether the execution of the workload satisfies one or more threshold objectives comprises to determine whether the execution of the workload satisfies the resource allocation objective data.   
     
     
         7 . The orchestrator server of  claim 6 , wherein to obtain the resource allocation objective data comprises to obtain one or more thresholds indicative of a target power consumption, a target life expectancy, a target heat production, and a target performance of one or more resources allocated to the managed node. 
     
     
         8 . The orchestrator server of  claim 2 , wherein to adjust the resource allocation to satisfy the one or more threshold objectives comprises to adjust one or more settings of one or more architecture features of the allocated resources. 
     
     
         9 . The orchestrator server of  claim 2 , wherein to obtain a profile comprises to obtain a profile that includes a model that relates an input parameter set that is further indicative of a type of the workload, a category of the workload, resource utilization behavior, or one or more of the threshold objectives to be satisfied during the execution of the workload to the output parameter set. 
     
     
         10 . The orchestrator server of  claim 1 , wherein to obtain a profile associated with the workload comprises to obtain a profile that includes a model that relates the input parameter set with an output parameter set that is further indicative of a target location of the resources. 
     
     
         11 . The orchestrator server of  claim 1 , wherein to determine the resources to allocate comprises to select, from the profile, a pre-stored output parameter set mapped to the input parameter set. 
     
     
         12 . The orchestrator server of  claim 1 , wherein to obtain the profile comprises to:
 determine whether a pre-stored profile is associated with the workload; and   generate, in response to a determination that a pre-stored profile is not associated with the workload, the profile from a reference profile.   
     
     
         13 . The orchestrator server of  claim 1 , wherein to identify the workload to be executed by the managed node comprises to receive a request from a client device to execute the workload. 
     
     
         14 . The orchestrator server of  claim 13 , wherein to obtain the profile comprises to receive the profile with the request from the client device. 
     
     
         15 . One or more machine-readable storage media comprising a plurality of instructions stored thereon that, in response to being executed, cause an orchestrator server to:
 identify a workload to be executed by a managed node;   obtain a profile associated with the workload, wherein the profile includes a model that relates an input parameter set indicative of one or more characteristics of the workload with an output parameter set indicative of one or more aspects of resources to be allocated for execution of the workload;   determine, as a function of the input parameter set and the model, resources to allocate to the managed node to execute the workload; and   allocate the determined resources to the managed node to execute the workload.   
     
     
         16 . The one or more machine-readable storage media of  claim 15 , wherein the plurality of instructions, when executed, further cause the orchestrator server to:
 receive telemetry data indicative of resource utilization and workload performance as the workload is executed;   determine, as a function of the telemetry data, whether one or more threshold objectives are satisfied by the execution of the workload with the allocation of resources;   adjust, in response to a determination that the one or more threshold objectives are not satisfied, the allocation of resources to the managed node to satisfy the one or more threshold objectives; and   adjust the model to include the adjustment to the allocation of resources.   
     
     
         17 . The one or more machine-readable storage media of  claim 16 , wherein to adjust the model comprises to adjust the model to produce an output parameter set of resources that represents the adjusted allocation of resources, in response to the input parameter set. 
     
     
         18 . The one or more machine-readable storage media of  claim 16 , wherein to adjust the model comprises to adjust the output parameter set to change one or more architecture features of the resources to be allocated, wherein the architecture features include one or more of support for an extended instruction set, support for preloading of processor cache with one or more predefined values, support for accelerated cryptographic operations, and support for accelerated data compression operations. 
     
     
         19 . The one or more machine-readable storage media of  claim 16 , wherein the plurality of instructions, when executed, further cause the orchestrator server to:
 generate, as a function of the telemetry data, landscape data indicative of conditions across the set of managed nodes in a data center; and   wherein to determine whether the threshold objectives are satisfied comprises to determine whether the threshold objectives are satisfied based additionally on the landscape data.   
     
     
         20 . The one or more machine-readable storage media of  claim 16 , wherein the plurality of instructions, when executed, further cause the orchestrator server to:
 obtain resource allocation objective data indicative of one or more thresholds to be satisfied during the execution of the workload; and   wherein to determine whether the execution of the workload satisfies one or more threshold objectives comprises to determine whether the execution of the workload satisfies the resource allocation objective data.   
     
     
         21 . The one or more machine-readable storage media of  claim 20 , wherein to obtain the resource allocation objective data comprises to obtain one or more thresholds indicative of a target power consumption, a target life expectancy, a target heat production, and a target performance of one or more resources allocated to the managed node. 
     
     
         22 . The one or more machine-readable storage media of  claim 16 , wherein to adjust the resource allocation to satisfy the one or more threshold objectives comprises to adjust one or more settings of one or more architecture features of the allocated resources. 
     
     
         23 . The one or more machine-readable storage media of  claim 16 , wherein to obtain a profile comprises to obtain a profile that includes a model that relates an input parameter set that is further indicative of a type of the workload, a category of the workload, resource utilization behavior, or one or more of the threshold objectives to be satisfied during the execution of the workload to the output parameter set. 
     
     
         24 . The one or more machine-readable storage media of  claim 15 , wherein to obtain a profile associated with the workload comprises to obtain a profile that includes a model that relates the input parameter set with an output parameter set that is further indicative of a target location of the resources. 
     
     
         25 . A orchestrator server to manage workload profiles, the orchestrator server comprising:
 means for identifying a workload to be executed by a managed node;   means for obtaining a profile associated with the workload, wherein the profile includes a model that relates an input parameter set indicative of one or more characteristics of the workload with an output parameter set indicative of one or more aspects of resources to be allocated for execution of the workload;   means for determining, as a function of the input parameter set and the model, resources to allocate to the managed node to execute the workload; and   circuitry for allocating the determined resources to the managed node to execute the workload.   
     
     
         26 . A method for managing workload profiles, the method comprising:
 identifying, by an orchestrator server, a workload to be executed by a managed node;   obtaining, by the orchestrator server, a profile associated with the workload, wherein the profile includes a model that relates an input parameter set indicative of one or more characteristics of the workload with an output parameter set indicative of one or more aspects of resources to be allocated for execution of the workload;   determining, by the orchestrator server and as a function of the input parameter set and the model, resources to allocate to the managed node to execute the workload; and   allocating, by the orchestrator server, the determined resources to the managed node to execute the workload.   
     
     
         27 . The method of  claim 26 , further comprising:
 receiving, by the orchestrator server, telemetry data indicative of resource utilization and workload performance as the workload is executed;   determining, by the orchestrator server and as a function of the telemetry data, whether one or more threshold objectives are satisfied by the execution of the workload with the allocation of resources;   adjusting, by the orchestrator server and in response to a determination that the one or more threshold objectives are not satisfied, the allocation of resources to the managed node to satisfy the one or more threshold objectives; and   adjusting, by the orchestrator server, the model to include the adjustment to the allocation of resources.   
     
     
         28 . The method of  claim 27 , wherein adjusting the model comprises adjusting the model to produce an output parameter set of resources that represents the adjusted allocation of resources, in response to the input parameter set.