{Market-Oriented} {Meta-Scheduling} for Utility Grids

Abstract

As users increasingly require better quality of service from Grids,resource management scheduling mechanisms have to evolve in order tosatisfy competing demands on limited resources. Traditional algorithms arebased on system-centric approaches which do not consider user requirementsand interests. These system-centric approaches for scheduling userapplications aims to maximize system performance and thus do not consideruser utility. Most importantly, these traditional techniques do not performeffectively when the demand for resources surpasses the supply. Therefore,we propose a complementary approach which uses basic economic and marketprinciples to help allocate limited resources more efficiently and fairlyby balancing supply and demand. This requires good valuation schemes forboth grid resources and user applications. In this work, we designvaluation and allocation mechanisms for mapping jobs with Quality ofService (QoS) requirements to heterogeneous grid resources. Finally, wepropose a Meta-Broker framework for Grid Market to realize these schedulingstrategies with different objectives.