A Resource Scheduling Algorithm Based on Division Load Theory for Parallel and Distributed Applications in Optical Grid

Abstract

Optical network provides cheaper bandwidth to grid services. When considering using lightpath to connect grid distributed resources, to provide flexible and dynamic lightpath scheduling, optical grid is constructed to solve the communication resource allocating problems for parallel and distributed applications. We present an optical resource scheduling scheme orienting parallel lightpath based on division load theory, to allocate resource on the requirement of large amount data transmission in parallel and distributed applications. Optical network resources are optimally utilized to obtain dynamic multi-granularity bandwidth in optical grid via wavelength-routed network. We also illustrate functional architecture for optical grid whose resources are connected via multi-wavelength routed paths and supports parallel and distributed services.