Method and system for collaborative workload placement and optimization

Patent metadata

Field	Value
Publication number	US-11748171-B2
Application number	US-202016821924-A
Country	US
Kind code	B2
Filing date	Mar 17, 2020
Priority date	Mar 17, 2020
Publication date	Sep 5, 2023
Grant date	Sep 5, 2023

Abstract

A method and system for collaborative workload placement and optimization. Specifically, the disclosed method and system entail implementing metric exchange between hosts in a clustered or non-clustered environment to perform infrastructure workload placement and optimization decisions. The exchange of metrics may be performed through a lightweight service, which would not impact any existing environment resource utilization. Further, resource utilization maps may be used to offer near real-time tracking of resource utilization across the aforementioned hosts, thereby facilitating resource utilization forecasting for the efficient placement and optimization of workloads throughout the environment.

First claim

What is claimed is:

1. A method for collaborative workload placement, comprising:
obtaining, by an origin infrastructure node (IN), a first workload request comprising a first workload configuration for a first workload, wherein the first workload configuration specifies at least an identifier of the first workload and minimum computing resource requirements for implementing the first workload, wherein the minimum computing resource requirements specify at least a virtualization resource and a networking resource;
identifying, by the origin IN and by examining a local resource utilization map (RUM) based on the first workload configuration, a plurality of participants that comprises a first target IN, wherein the first target IN is not the origin IN;
multicasting, by the origin IN, a workload allocation request for the first workload to the plurality of participants operatively connected through a metric exchange network, wherein the workload allocation request specifies the first workload configuration sought to be implemented, wherein the workload allocation request comprises a RUM snapshot of the local RUM maintained on the origin IN;
unicasting, by the origin IN, a workload allocation confirmation for the first workload to the first target IN, upon receiving the workload allocation confirmation from the plurality of participants;
receiving, from the first target IN, a workload allocation confirmation response verifying that the first target IN comprises available computing resources satisfying the minimum computing resource requirements specified in the first workload configuration;
deploying, by the origin IN and based on the workload allocation confirmation response from the first target IN, the first workload to the first target IN, wherein the confirmation response indicates that the first target IN is able to implement the first workload; and
executing the first workload by the first target IN.

2. The method of claim 1 , wherein identifying the first target IN, comprises:
comparing the minimum computing resource requirements against the local RUM comprising a plurality of metric sets corresponding to the plurality of participants; and
identifying the first target IN from the plurality of participants based on a metric set, corresponding to the first target IN, at least meeting the minimum computing resource requirements.

3. The method of claim 1 , wherein the workload allocation request further comprises target IN metadata describing the first target IN.

4. The method of claim 3 , wherein the target IN metadata comprises an IN name and a network address associated with the first target IN.

5. The method of claim 3 , wherein each of the responses to the workload allocation request from the plurality of participants reflects that a remote RUM, maintained on a respective participant, is consistent with the RUM snapshot.

6. The method of claim 1 , wherein the minimum computing resource requirements further specify a processing resource and a storage resource.

7. The method of claim 1 , further comprising:
receiving, from the first target IN, a multi-casted metric exchange message comprising a new remote metric set reflecting computing resource availability on the first target IN following an implementation of the first workload thereon; and
updating the local RUM by using the new remote metric set to replace an existing remote metric set that reflected computing resource availability on the first target IN prior to the implementation of the first workload thereon.

8. The method of claim 1 , further comprising:
obtaining, by the origin IN, a second workload request comprising a second workload configuration for a second workload;
identifying, based on the second workload configuration, a second target IN, wherein the second target IN is the origin IN;
making a determination that a local metric set satisfies the minimum computing resource requirements specified in the second workload configuration;
allocating, based on the determination, local computing resources at least equivalent to the minimum computing resource requirements; and
implementing the second workload locally using the local computing resources.

9. The method of claim 1 , further comprising:
obtaining, by the origin IN, a second workload request comprising a second workload configuration for a second workload;
identifying, based on the second workload configuration, a second target IN, wherein the second target IN is the origin IN;
making a determination that a local metric set fails to satisfy minimum computing resource requirements specified in the second workload configuration;
based on the determination:
updating the local RUM using the local metric set; and
identifying, based on the second workload configuration, a third target IN, wherein the third target IN is not the origin IN.

10. A non-transitory computer readable medium (CRM) comprising computer readable program code, which when executed by a computer processor, enables the computer processor to:
obtain, by an origin infrastructure node (IN), a first workload request comprising a first workload configuration for a first workload, wherein the first workload configuration specifies at least an identifier of the first workload and minimum computing resource requirements for implementing the first workload, wherein the minimum computing resource requirements specify at least a virtualization resource and a networking resource;
identify, by the origin IN and by examining a local resource utilization map (RUM) based on the first workload configuration, a plurality of participants that comprises a first target IN, wherein the first target IN is not the origin IN;
multicast, by the origin IN, a workload allocation request for the first workload to the plurality of participants operatively connected through a metric exchange network, wherein the workload allocation request specifies the first workload configuration sought to be implemented, wherein the workload allocation request comprises a RUM snapshot of the local RUM maintained on the origin IN;
unicast, by the origin IN, a workload allocation confirmation for the first workload to the first target IN, upon receiving the workload allocation confirmation from the plurality of participants;
receive, from the first target IN, a workload allocation confirmation response verifying that the first target IN comprises available computing resources satisfying the minimum computing resource requirements specified in the first workload configuration;
deploy, by the origin IN and based on the workload allocation confirmation response from the first target IN, the first workload to the first target IN, wherein the confirmation response indicates that the first target IN is able to implement the first workload; and
execute the first workload by the first target IN.

11. The non-transitory CRM of claim 10 , comprising computer readable program code, which when executed by the computer processor, further enables the computer processor to identify the first target IN, by:
comparing the minimum computing resource requirements against the local RUM comprising a plurality of metric sets corresponding to the plurality of participants; and
identifying the first target IN from the plurality of participants based on a metric set, corresponding to the first target IN, at least meeting the minimum computing resource requirements.

12. The non-transitory CRM of claim 10 , wherein the workload allocation request further comprises target IN metadata describing the first target IN.

13. The non-transitory CRM of claim 12 , wherein the target IN metadata comprises an IN name and a

Assignees

• Dell Products Lp

Inventors

• Chaganti Ravikanth
• Patel Dharmesh M
• Ali Rizwan

Classifications

• G06F9/5083 Primary

  Techniques for rebalancing the load in a distributed system · 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