Real-time multitask load balance algorithm for heterogeneous cloud computing platforms

A load-balancing algorithm applying in heterogeneous cloud Computing Platform handling real-time multitasks was proposed. Average hardware resource consumption of jobs running on nodes was measured. Balancing server receive load status of each node in the cluster periodically. A load status vector that reflects the quantity of resources required to finish allocated jobs of each node can be estimated according to the latest load status report and other parameters. As a request is submitted to the cluster, Balancing Server calculates the load status estimation vector of each node, and then dispatches it to the node that possesses the minimal load status estimation value. Experiment results proved that this dynamic load balancing algorithm is reasonable and effective.