Dynamic partition and replication algorithm for storage capacity limited distributed social network

Large scale online social networks (OSN) are usually based on distributed storage systems. The data of users are distributed on multiple storage and computing nodes to provide concurrency and redundancy. How to store and access user data efficiently in OSN system with limited storage has practical significance. This paper proposes a dynamic partitioning and replication algorithm, which partitions and replicates user data periodically in line with the frequency of user interaction, in order to improve local access ratio and reduce system communication load. Through experiment on a real-world dataset, this paper testifies that compared with random partitioning and replication, the proposed algorithm could improve local access ratio greatly, and finally reduce access latency.