Method and apparatus for determining resource node, method of controlling resource node, resource node and computer product

Abstract

By transmitting information on a condition enabling execution of one service (A) to a resource broker (702) that manages an assignment state of resource nodes (103) used among services (A, B), specific resource nodes (103) enabling execution of the one service (A) out of a plurality of resource nodes used among services (A, B) are specified. Further, by receiving information on a use state of each specific resource node and based on the information on the received use state, a resource node (103) to be used for the one service (A) from among specified resource nodes is determined.

Description

technical field [0001] The invention relates to a resource node allocation technology in a distributed computer environment. Background technique [0002] Generally, when access from users is concentrated on one service, or a service that requires a large amount of computing work is performed, for load distribution, a distributed computer environment in which a large number of computers are connected via a network, for example, in which different types of computer grid environment. [0003] In this case, it is necessary to automatically and quickly select an optimal computer (hereinafter "resource node") suitable for the purpose of a service among a large number of computers to be allocated to the service. Especially in a grid environment in which different types of computers are mixed with each other, the performance of the entire service greatly depends on the allocated resource nodes. [0004] Therefore, in order to improve the performance of the entire service, resourc...

AI Technical Summary

Problems solved by technology

[0007] Therefore, there is a problem that, in each resource node, when a host machine downtime occurs just before the application starts, a start instruction cannot be received due to a temporary communication problem, etc., because the resource nodes necessary for executing the service are not ready, the unable to provide service

[0009] However, since in many cases the selection criteria are different for each service, the selection criteria for each service must be understood in advance, and resource agents based on them must be developed

Therefore, there is the following problem: when the service to be executed increases or the distributed computer environment is used for other services, the selection criteria cannot be flexibly added or changed, resulting in poor versatility

Images