Application characteristic-based isomeric group operation self-adapting dispatching method and system

Abstract

The invention provides an application characteristic-based isomeric group operation self-adapting dispatching method and an application characteristic-based isomeric group operation self-adapting dispatching system, wherein the dispatching method comprises the following steps that: an operation dispatching server collects hardware information of all nodes of the isomeric group and divides the nodes into node groups with different characteristics according to the hardware information of each node; an operation dispatching middleware inquires whether the operating characteristics of the operation currently submitted is stored in the data base; if so, the operation dispatching server submits the operation currently submitted to the node group with corresponding characteristics for operation according to the operating characteristics; and otherwise, the operation dispatching server submits the operation currently submitted to any node group for operation, and the operation dispatching middleware records the operating data of the operation currently submitted, analyzes the operating characteristics according to the operating data and stores the application type and the operating characteristics of the operation currently submitted in the data base. The method and the system of the invention can improve the efficiency of the high-performance software and also improve the use efficiency of the isomeric group.

Description

technical field [0001] The invention relates to the high-performance field of computer technology, in particular to an adaptive scheduling method and system for heterogeneous cluster jobs based on application characteristics. Background technique [0002] A cluster is the abbreviation of a computer cluster, which is connected by a group of loosely integrated computer software and / or hardware to complete computing work in a highly closely coordinated manner. Individual computers in a cluster are often called nodes. Clusters are often used to improve the computing speed and / or reliability of individual computers. [0003] The operating characteristics of business software in the high-performance field vary widely, and the resource requirements for clusters are different. Some are computing-intensive, requiring high CPU processing speed, and some are I / O-intensive, requiring high hard disk I / O rates and paging. Rate. However, the traditional cluster job scheduling method doe...

AI Technical Summary

Problems solved by technology

[0004] Aiming at the problem that the cluster job scheduling method in the prior art does not consider the operating characteristics of high-performance software and the characteristics of heterogeneous clusters, which leads to the reduction of the efficiency of ...

Images