Two-stage dynamic optimization scheduling mechanism based on cloud storage

Two-stage dynamic optimization scheduling mechanism based on cloud storage

Ninchown Yang Dongju

In the research of distributed storage, how to use storage space efficiently is a hot issue. Storage cluster, the storage capacity of each data node can not be exactly the same, because the main node to select the Random data node, the selected data node disk may be close to full, at this time the main node will automatically do the storage load balance, occupy the data transmission bandwidth, not only affect the performance of data transmission, but also This paper presents a two-stage dynamic optimal scheduling mechanism based on cloud storage: The first stage, by calculating the optimal ratio of replica storage, using the local optimal storage scheme based on greedy algorithm, selecting storage nodes and balancing the space of copy placement; the second stage adopts real-time monitoring storage cluster, dynamically adjusts the copy placement node, Achieve efficient utilization of storage resources. Finally, the experiment verifies that the scheduling mechanism can effectively place the copy, reduce the data transmission between nodes and improve the efficiency of file access.

Two-stage dynamic optimization scheduling mechanism based on cloud storage