System and method for an optimizing reservation in time of compute resources based on prioritization function and reservation policy parameter

Patent metadata

Field	Value
Publication number	US-11960937-B2
Application number	US-202217697403-A
Country	US
Kind code	B2
Filing date	Mar 17, 2022
Priority date	Mar 13, 2004
Publication date	Apr 16, 2024
Grant date	Apr 16, 2024

Abstract

A system and method of dynamically controlling a reservation of resources within a cluster environment to maximize a response time are disclosed. The method embodiment of the invention includes receiving from a requestor a request for a reservation of resources in the cluster environment, reserving a first group of resources, evaluating resources within the cluster environment to determine if the response time can be improved and if the response time can be improved, then canceling the reservation for the first group of resources and reserving a second group of resources to process the request at the improved response time.

First claim

What is claimed is:

1. A method of operating a compute environment having a plurality of compute nodes so as to provide workload prioritization, the method comprising:
identifying, at a first time, first workload which cannot be processed immediately due to one or more resource requirements associated with the first workload exceeding an amount of resources which are available at the first time;
utilizing priority data associated with the first workload to prioritize processing of the first workload within the compute environment relative to one or more other workloads;
identifying one or more additional resources that have become available after the first time, and which enable processing the first workload;
based at least on the identifying the one or more additional resources, causing reserving of at least the one or more additional resources; and
subsequent to causing the reserving, causing processing of the first workload using at least the reserved one or more additional resources; wherein:
the utilizing of the priority data associated with the first workload comprises utilizing priority data submitted to the compute environment via a submission from a computerized user process;
the method further comprises utilizing at least one software object to control access by one or more identified users or the computerized user process to a prioritization function associated with the priority data;
the at least one software object comprises a class having one or more attributes, and the priority data comprises data identifying the class;
the one or more attributes comprise a host constraint; and
the controlling access by the one or more identified users or the computerized user process to the prioritization function comprises enabling only the one or more identified users or the computerized user process to generate data determining prioritization of the processing of the first workload within the compute environment relative to the one or more other workloads, the generated data comprising a value representative of a reservation policy parameter which indicates that the reserving of at least the one or more additional resources will be maintained until the processing of the first workload regardless of whether the one or more other workloads bypass the first workload in priority.

2. The method of claim 1 , wherein the at least one software object restricts access based at least on the class and utilization limit data.

3. The method of claim 1 , wherein the utilizing of the priority data associated with the first workload to prioritize the processing of the first workload within the compute environment relative to the one or more other workloads comprises utilizing the priority data to enable out-of-order processing of the first workload relative to the one or more other workloads, the one or more other workloads otherwise having a higher priority than the first workload due to being able to be scheduled using then-available resources of the compute environment.

4. The method of claim 1 , wherein the identifying the one or more additional resources that have become available after the first time, and which enable processing the first workload, comprises identifying the one or more additional resources that have been made available via at least cancellation of the one or more other workloads.

5. The method of claim 4 , wherein the cancellation is based on a cost associated with cancelling the one or more other workloads and reserving of at least the one or more additional resources.

6. The method of claim 1 , wherein the identifying of the one or more additional resources that have become available after the first time, and which enable processing the first workload, comprises:
monitoring one or more parameters of at least one of the plurality of compute nodes; determining that processing of at least one workload by the at least one of the plurality of compute nodes has been at least one of (i) canceled or (ii) completed; and
based at least on the determination that the processing of the at least one workload by the at least one of the plurality of compute nodes has been at least one of (i) canceled or (ii) completed, causing the identifying.

7. The method of claim 1 , wherein: the causing the reserving of the at least the one or more additional resources comprises causing of both (i) the one or more additional resources, and (ii) resources which were available at the first time; and the processing of the first workload using at least the reserved one or more additional resources comprises processing using both (i) the one or more additional resources, and (ii) the resources which were available at the first time.

8. The method of claim 1 , wherein the one or more attributes comprise a default job, and/or a special access constraint that indicates an adjusted job priority.

9. The method of claim 1 , wherein the identifying of the one or more additional resources that have become available after the first time, and which enable processing the first workload, comprises identifying one or more additional resources that meet one or more required criteria and one or more preferred criteria.

10. A method of operating a compute environment having a plurality of compute nodes and a computerized scheduler process, the operating so as to provide workload prioritization, the method comprising:
identifying, at a first time, first workload which cannot be processed immediately due to one or more resource requirements associated with the first workload exceeding an amount of resources which are available at the first time;
utilizing priority data associated with the first workload to prioritize processing of the first workload within the compute environment relative to one or more other workloads, wherein the utilizing priority data associated with the first workload comprises utilizing of the priority data submitted to the compute environment via a submission from a computerized user process in data communication with the compute environment via an internetwork, wherein the priority data (i) is created by the computerized user process and (ii) comprises a priority specified by the computerized user process, and wherein the utilizing of the priority data comprises modifying a default prioritization process where only a highest priority workload receives a priority reservation, the modifying of the default prioritization process effected by adjusting a value associated with a reservation depth parameter which enables any workloads which cannot be processed immediately to receive the priority reservation;
identifying resources, at least a portion of which that have become available after the first time, and which enable processing the first workload;
based at least on the identified resources, causing reserving of the identified resources via the computerized scheduler process;
subsequent to causing the reserving, causing the processing of the first workload using at least the reserved resources; and
controlling access by the computerized user process to a prioritization function associated with the priority data via one or more software objects which restricts access to the prioritization function to only one or more identified users or the computerized user process, wherein the controlling of the access by the computerized user process to the prioritization function associated with the priority data comprises utilizing the one or more software objects comprising a class having one or more attributes, at least one of the one or more attributes which restricts access based at least on the class and utilization limit data, wherein the utilization limit data is configurable by an administrator and associated with how much of the resour

Assignees

• Iii Holdings 12 Llc

Inventors

• Jackson David B

Classifications

• G06F9/5027 Primary

  the resource being a machine, e.g. CPUs, Servers, Terminals · CPC title

• G06F9/4881

  Scheduling strategies for dispatcher, e.g. round robin, multi-level priority queues · CPC title

• G06F9/5011

  the resources being hardware resources other than CPUs, Servers and Terminals · CPC title

• G06F9/505 Primary

  considering the load · CPC title

• G06F9/5072

  Grid computing · CPC title