Efficient Implementation of Best-Fit Scheduling for Advance Reservations and QoS in Grids

In this work we consider the problem of providing QoS guarantees to Grid users by means of advance reservation of resources. Advance reservation mechanisms provide with the ability to allocate resources to users based on agreed-upon QoS requirements and increase the predictability of a Grid system. However, incorporating such mechanisms into current Grid environments has proven to be a challenging task due to the dynamism and heterogeneity of Grid environments. In our previous work we introduced a suite of scheduling algorithms for Grids capable of allocating resource efficiently and providing with QoS guarantee. In this paper we extend our previous work by introducing an efficient implementation of the best-fit scheduling algorithm. We evaluate the performance of the best-fit algorithm by means of extensive simulation and analysis against the scheduling algorithms introduced previously. Our results show that best-fit algorithm performs well across several metrics that reflect both user and system specific goals.