A Task Schedule Approach for Heterogeneous Storm Based on Particle Swarm-Load Balancing Algorithm

Storm is used for the calculation of streaming big data. However, the default task scheduling strategy does not consider the resource situation of the cluster for polling allocation. This paper presents a MDFDRPSO-EDLB algorithm, considers heterogeneous clusters of resource utilization, network factors to the task static scheduling. Using the storm resource benchmark test, compared with the default scheduling strategy, the sum of the node's cpu and memory utilization can be steadily improved.