Data distribution strategy for image computing parallel in cloud

To reduce the total response time on the system and balance the workload of all computer nodes in the cloud computing system.The data distribution techniques is very important.A proportional data distribution strategy for image computing parallel is proposed.The cloud computing system is the master/slave model.The capacity of each node is expressed by performance function.The total data is partitioned according to the performance ratio of all slave nodes and the order of data sended is decided by the bandwidth between slave node and the master node.The experimentations are conducted on our cloud tested.The results show that the strategy can improve the efficiency of data processing.