| In cloud computing platform application technology, resource scheduling algorithm is critical, because it is responsible for the tasks assigned to the worker nodes, the cloud computing platform performance is good or bad is mainly reflected in the reasonable degree of tasks allocation. Many existing resource scheduling algorithm, some classical algorithms such as Min‐min algorithm, genetic algorithm, WQR algorithm and LATE algorithm,these algorithms in the design of resource scheduling algorithms proposed some very important ideas. But they all are not good enough.Because of various problems of existing algorithms, this paper presents a work queue priority resource scheduling algorithms. The algorithm is based on the WQ algorithm that combines the advantages of Hadoop cloud platform, innovative way to join the working node fuzzy rating strategy. node selection algorithm and node migration algorithm are independent. The algorithm used the average completion time for the work node rating, avoiding the evaluation of information according to the hardware difficulties. Work node selection algorithm to find the order‐class queue, the level found in the work of randomly selected node, and then assign a task to worker nodes to run. Node migration algorithm comparison and different level through peer comparison process, the gradual completion of the class queue. Mathematical theory papers also compare performance of the algorithm, by calculating the probability of working nodes selected to the average performance of the algorithm were compared.Paper is also conducted simulation experiments through cloudSim to compare the average project completion time for the task, the task completion time and total resource consumption. Faster convergence of the algorithm can be applied effectively in the cloud computing platform. Paper finally also designed scheduling module, detailing the composition of modular components, and the actual development work on future have a good guide. |
PDF Full Text Request