×

Transferring a partial task in a distributed computing system

  • US 10,140,177 B2
  • Filed: 12/06/2012
  • Issued: 11/27/2018
  • Est. Priority Date: 01/31/2012
  • Status: Active Grant
First Claim
Patent Images

1. A method comprises:

  • determining, by a first dispersed storage and task (DST) execution unit, whether partial task processing resources of the first DST execution unit are projected to be available, wherein the determining is based on;

    comparing a number of partial tasks in a current snapshot of a first local task queue to a current queue threshold, anddetermining a number of partial tasks in a projected snapshot of the first local task queue, and a projected queue threshold, wherein the determining is based at least partially on a first expected partial task performance factor and a first expected partial task allocation factor,wherein the first expected partial task performance factor is based on either when the first DST execution unit will perform one or more of the partial tasks in the projected snapshot of the first local task queue or how long the first DST execution unit will take to perform one or more of the partial tasks in the projected snapshot of the first local task queue, and wherein the first expected partial task allocation factor is based on the number of partial tasks in the first local task queue and the types of partial tasks in the first local task queue;

    when partial task processing resources of the first DST execution unit are projected to be available, determining, whether partial task processing resources of a second DST execution unit are projected to be overburdened, wherein the determining is based on;

    comparing a number of partial tasks in a current snapshot of a second local task queue to a current queue threshold associated with the second local task queue, anddetermining a number of partial tasks in a projected snapshot of the second local task queue, and a projected queue threshold for the second local task queue, wherein the determining is based at least partially on a second expected partial task performance factor and a second expected partial task allocation factor,wherein the second expected partial task performance factor is based on when the second DST execution unit will perform one or more of the partial tasks in the projected snapshot of the second local task queue or how long the second DST execution unit will take to perform one or more of the partial tasks in the projected snapshot of the second local task queue, and wherein the second expected partial task allocation factor is based on the number of partial tasks in the second local task queue and the types of partial tasks in the second local task queue;

    when partial task processing resources of the second DST execution unit are projected to be overburdened, receiving, by the first DST execution unit, a partial task of a set of partial tasks from the second DST execution unit, wherein the partial task is associated with a data portion, wherein the data portion includes a plurality of data partitions, wherein each of the plurality of data partitions comprises a group of encoded data slices and further wherein the first DST execution unit and the second DST execution unit are each assigned at least one partial task of the set of partial tasks to perform on the group of encoded data slices; and

    executing, by the first DST execution unit, the partial task to produce a partial resultant.

View all claims
  • 5 Assignments
Timeline View
Assignment View
    ×
    ×