Research on Indexing Mechanism for Massive Distributed Data Fragment Storage

Authors

  • Huan Wang, Yong Peng, Huajian Huang, Junyi Deng, Huajun Huang, Jingxian Liu

Abstract

Currently, massive distributed storage systems store a large number of interrelated heterogeneous data, but lack of a perfect indexing mechanism to solve the query problem of massive heterogeneous data. Therefore, this paper proposes an index model based on the combination of associative index and auxiliary index. The model uses the description of the common entity between structured data and unstructured data to establish a relationship, and creates an index as a keyword to establish an associated index layer. At the same time, a secondary index layer is established at each node to create an R tree index for the structured data, and an entity-based inverted index is created for the unstructured free document, thereby improving the query efficiency and accuracy. Experimental results show that the indexing system can not only effectively support mixed queries of heterogeneous data, but also improve the accuracy of query results.

Published

2020-02-28

Issue

Section

Articles