Edge node workload acquisition for distribution of edge workloads

Patent metadata

Field	Value
Publication number	US-12058047-B2
Application number	US-202217682256-A
Country	US
Kind code	B2
Filing date	Feb 28, 2022
Priority date	Feb 28, 2022
Publication date	Aug 6, 2024
Grant date	Aug 6, 2024

Abstract

A method comprises executing a workload acquiring algorithm at a first edge node of a plurality of edge nodes to pull at least one workload of a plurality of workloads from a workload queue corresponding to a second edge node of the plurality of edge nodes. In one example, the first edge node may be a component of a first edge location, and the method may further comprise providing the first edge node with access to one or more workloads from one or more workload queues corresponding to one or more edge nodes from a second edge location different from the first edge location.

First claim

What is claimed is:

1. A method, comprising:
executing a workload acquiring algorithm at a first edge node of a plurality of edge nodes to pull at least one workload of a plurality of workloads from a workload queue corresponding to a second edge node of the plurality of edge nodes;
wherein the plurality of workloads are assigned to respective ones of a plurality of workload queues corresponding to respective ones of the plurality of edge nodes, wherein a primary queue list comprises references to the plurality of workload queues, wherein respective ones of the references are added to the primary queue list responsive to addition of at least one workload of the plurality of workloads to respective ones of the plurality of workload queues, and wherein the respective ones of the references identify respective ones of the plurality of edge nodes corresponding to the respective ones of the plurality of workload queues;
wherein executing the workload acquiring algorithm comprises:
querying the primary queue list for one or more of the references to the plurality of workload queues;
finding a reference to the workload queue corresponding to the second edge node of the plurality of edge nodes in the primary queue list, the reference identifying the second edge node;
pulling the at least one workload from the workload queue corresponding to the second edge node responsive to the finding of the reference to the workload queue corresponding to the second edge node, wherein the at least one workload is awaiting processing by the second edge node; and
processing, by the first edge node, the at least one workload from the workload queue corresponding to the second edge node;
wherein, after the pulling and the processing, the reference to the workload queue corresponding to the second edge node is maintained in the primary queue list when one or more workloads remain in the workload queue corresponding to the second edge node; and
wherein, prior to the processing, the reference to the workload queue corresponding to the second edge node is removed from the primary queue list when the pulling empties the workload queue corresponding to the second edge node; and
wherein the method is performed by at least one processing device comprising a processor coupled to a memory.

2. The method of claim 1 , further comprising:
processing, by the first edge node, one or more workloads of the plurality of workloads in a workload queue corresponding to the first edge node, wherein the processing is performed until the workload queue corresponding to the first edge node is empty; and
wherein the pulling of the at least one workload from the workload queue corresponding to the second edge node is performed following emptying of the workload queue corresponding to the first edge node.

3. The method of claim 2 , further comprising:
spawning a new workload during the processing of the one or more workloads in the workload queue corresponding to the first edge node; and
adding the new workload to the workload queue corresponding to the first edge node.

4. The method of claim 3 , wherein the new workload is available for acquisition from the workload queue corresponding to the first edge node by an edge node of the plurality of edge nodes other than the first edge node.

5. The method of claim 1 , wherein the plurality of edge nodes are components of a first edge location, and one or more edge nodes from a second edge location different from the first edge location are provided with access to the plurality of workloads from the respective ones of the plurality of workload queues.

6. The method of claim 5 , wherein a given one of the plurality of workloads is pulled from its corresponding workload queue and processed by one or more edge nodes from the second edge location.

7. The method of claim 5 , wherein:
a first primary queue list in the first edge location comprises references to the plurality of workload queues corresponding to the respective ones of the plurality of edge nodes from the first edge location; and
a second primary queue list in the first edge location comprises references to one or more workload queues corresponding to the one or more edge nodes from the second edge location.

8. The method of claim 7 , wherein:
a first primary queue list in the second edge location comprises references to the one or more workload queues corresponding to the one or more edge nodes from the second edge location; and
a second primary queue list in the second edge location comprises references to the plurality of workload queues corresponding to the respective ones of the plurality of edge nodes from the first edge location.

9. The method of claim 1 , wherein the first edge node is a component of a first edge location, and the method further comprises providing the first edge node with access to one or more second edge location workloads from one or more workload queues corresponding to one or more edge nodes from a second edge location different from the first edge location.

10. The method of claim 9 , wherein executing the workload acquiring algorithm further comprises:
pulling, by the first edge node, at least one of the one or more second edge location workloads from the one or more workload queues corresponding to the one or more edge nodes from the second edge location; and
processing, by the first edge node, the at least one of the one or more second edge location workloads from the one or more workload queues corresponding to the one or more edge nodes from the second edge location.

11. The method of claim 9 , wherein the access by the first edge node to the one or more second edge location workloads from the one or more workload queues corresponding to the one or more edge nodes from the second edge location is managed by a workload acquiring communication protocol executed between respective first and second workload schedulers of the first and second edge locations.

12. The method of claim 1 , wherein executing the workload acquiring algorithm further comprises implementing one or more constraints governing whether the first edge node is permitted to pull workloads from the workload queue corresponding to the second edge node, the one or more constraints comprising at least one of an edge node location constraint, a workload processing performance constraint, a workload processing cost constraint, an edge node capability constraint, an energy efficiency constraint, and a security constraint.

13. The method of claim 1 , wherein executing the workload acquiring algorithm further comprises implementing at least one constraint governing whether the first edge node is permitted to pull workloads from the workload queue corresponding to the second edge node, the at least one constraint comprising an edge node location constraint.

14. An apparatus, comprising:
at least one processor and at least one memory storing computer program instructions wherein, when the at least one processor executes the computer program instructions, the apparatus is configured:
to execute a workload acquiring algorithm at a first edge node of a plurality of edge nodes to pull at least one workload of a plurality of workloads from a workload queue corresponding to a second edge node of the plurality of edge nodes;
wherein the plurality of workloads are assigned to respective ones of a plurality of workload queues corresponding to respective ones of the plurality of edge nodes, wherein a primary queue list comprises references to the plurality of workload queues, wherein respective ones of the references are added to the primary queue list responsive to addition of at least one

Assignees

• Dell Products Lp

Inventors

• Bruno Eric
• Seibel Amy

Classifications

• H04L47/12 Primary

  Avoiding congestion; Recovering from congestion · CPC title

• H04L47/62 Primary

  characterised by scheduling criteria · CPC title