Efficient metascheduling in a cloud extended grid environment

This paper explores the application of an efficient technique for scheduling in grids and further extends it with clouds. Our approach, takes into account the available bandwidth of the link before selecting resources for job scheduling. Thus, this strategy selects the resource in such manner that along with computational capability, the ability of the resource to quickly respond to a task is also taken into account by means of using available bandwidth. Also, we are proposing here to extend this with cloud resources on demand if need arises in case of lack of grid resources. Thus, if peak demand arises in a grid environment, this approach instantiates an on demand cloud resource customized to meet the grid user requirements. The transfer time is lowered as a result of including network awareness in scheduling, which is evident from the experimental results. Providing on demand cloud resources is expected to lower the waiting times of jobs in the queue, thus improving performance. This also ensures reliability by means of providing a cloud resource thus assuring job completion.