IPFS-DKRM: An Efficient Keyword Retrieval Model of IPFS Based on ART

The content-based addressing method of Interplanetary File System (IPFS) leads to the lack of the function of retrieving relevant data through key information. To solve this problem, this paper proposes an efficient IPFS keyword retrieval model – IPFS-DKRM (IPFS-Distributed keyword retrieval model). This model combines the global index with Adaptive Radix Tree, and optimizes the storage mode of IPFS network and node-local data index: The model adopts the global index method and uses ART to store the key information of global index locally and stores the complete index information in IPFS to reduce the time of data retrieval and update; The nodes of network through the Publish-Subscribe Pattern synchronization index, and the use of Conflict-free Replication Data Type (CRDT) to maintain the final consistency of the global index of each node, to ensure that all nodes in the local to provide efficient retrieval services. In the simulation experiment, the index of open source data set Crosswikis was constructed, and the performance was analyzed based on the results of Siva data. The experimental results showed that compared with the Siva model, the response retrieval time of IPFS-DKRM was reduced by 75%, and the space occupied by node local storage index was reduced by 70%. It proves that the model only needs to occupy a small amount of space to store the global index information in the system to provide efficient retrieval function for IPFS, so that IPFS can meet more application scenarios in the future.