A Global Dynamic Scheduling with Replica Selection Algorithm Using GridFTP

This paper presents a global dynamic scheduling algorithm which schedules a set of data transfer requests made by a batch of data intensive tasks in a wide-area environment like the Grid, This proposed algorithm is adaptive in that it accounts for network bandwidth fluctuations in the wide-area environment. Adaptive replica selection is used for transferring different chunks of the same file by taking dynamic network information into account. Using GridFTP for data transfers and utilizes information from past GridFTP transfers to perform predictive bandwidth estimations. The experimental results show the effectiveness of this scheme on a wide-area testbed.