Packing objects by predicted lifespan in cloud storage

摘要

The method 600 includes receiving data objects 200, determining a predicted lifespan 220 of each data object, and instantiating a plurality of shard files 250. Each shard file has an associated prediction life span 260. The method also includes writing each data object to a corresponding shard file having an associated predicted life span that includes the predicted life span of each data object and storing the shard files in distributed system 140. . The method also includes determining whether any stored shard files meet the compression criteria 280 based on the number of data objects deleted in each corresponding stored shard file. For each stored shard file that meets the compression criteria, the method also includes compressing the stored shard file by rewriting the remaining data objects of the stored shard file into a new shard file.