HDFS Copy Storage Improvement Strategy
Jing Zhang, Hongbo Sun, Shijing Yuan · 2019
To improve the reliability of data storage, HDFS adopts the multi-copy storage strategy of Rack-Aware. However, it does not consider the data block's access heat during the replication process, resulting in that in the case of high concurrent access, the client's request for the hot file cannot be responded quickly, and the storage cluster's I/O throughput rate is significantly reduced.