Memory cloud storage optimization strategy based on small file

Memory cloud storage optimization strategy based on small file

Yu Yu Lu Liang Liu Mine

Because the memory cloud Ramcloud the use of log segments to store data, when a large number of small files are stored in the Ramcloud cluster, each small file monopolizes the entire segment, resulting in more fragment fragmentation, resulting in less efficient memory utilization and a large amount of memory space wasted. To solve this problem, a ramcloud small file storage optimization strategy based on file classification is proposed. This strategy first divides the small file into the structure related file, the logical correlation file and the independent file three class according to the file correlation characteristic, then uses the file merging algorithm for the structure related file in the storage time, the logical correlation and the independent small file uses the grouping algorithm. The experimental results show that this strategy can effectively improve the memory utilization of the cluster compared with the Ramcloud storage strategy without optimization.

Memory cloud storage optimization strategy based on small file