Load balancing method and system

A load balancing and task technology, applied in the field of data processing, can solve problems such as reducing system computing efficiency and different computing loads.

Active Publication Date: 2019-04-23
CHINA MOBILE SUZHOU SOFTWARE TECH CO LTD +1
View PDF2 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0006] The Pregel-like processing system adopts iterative operations and tends to gradually converge, resulting in different computing loads for each task; in the roadblock synchronization stage, it will cause a serious "bucket effect" and reduce the overall computing efficiency of the system

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Load balancing method and system
  • Load balancing method and system
  • Load balancing method and system

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0088] The preferred embodiments of the present invention will be described in detail below in conjunction with the accompanying drawings. It should be understood that the preferred embodiments described below are only used to illustrate and explain the present invention, and are not intended to limit the present invention.

[0089] In related technologies, the load balancing technology that can be adopted by Pregel-like processing systems may include: dynamic load balancing technology centered on tasks, dynamic load balancing technology centered on task data blocks, and dynamic load balancing technology centered on task vertices technology; these load balancing technologies do not take into account the topological relationship between graph data (that is, the edge dependence between vertices), resulting in low prediction accuracy; in addition, although from task-centered to vertex-centered, A certain degree of refinement has been carried out, but the refinement granularity is ...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention discloses a load balancing method and system. The load balancing method steps of obtaining related information of each task in an Lth super step; predicting the activation vertex number of each task in a Kth super step according to the related information of each task; predicting the operation time of each task in the Kth super step according to the activation vertex number of each task in the Kth super step; L being a positive integer which is more than or equal to 1, and K being a positive integer which is larger than L; determining a short plank task set and a long plank task set are determined according to the operation time of each task in the Kth super step; predicting migration income and migration cost generated when data on one task in the short plank task set is migrated to one task in the long plank task set; and migrating the data on the task in the short plank task set to the task in the long plank task set when the migration income is not lower than the migration cost. The bucket effect can be avoided, the calculated load of each task is balanced, and the calculating efficiency is improved.

Description

technical field [0001] The invention relates to data processing technology, in particular to a load balancing method and system. Background technique [0002] For the Pregel-like processing system based on the Bulk Synchronous Parallel Computing Model (BSP, Bulk Synchronous Parallel Computing Model), a graph processing job consists of a series of supersteps, and the work of each superstep is completed in parallel by multiple distributed tasks, including Three sequential stages: [0003] (1) Concurrent computing stage: each task uses the data stored in the local memory to complete the computing work; [0004] (2) Communication phase: exchange calculation results between tasks; [0005] (3) Roadblock synchronization stage: When a task completes the preceding stages (1) and (2), it enters the roadblock synchronization stage and waits for other tasks until all tasks complete the work of stages (1) and (2). [0006] The Pregel-like processing system adopts iterative operation ...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
Patent Type & Authority Patents(China)
IPC IPC(8): G06F9/50
Inventor 徐萌鲍玉斌钱岭
Owner CHINA MOBILE SUZHOU SOFTWARE TECH CO LTD
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap