DOMI: A Dynamic Ordering Multi-field Index for Data Processing

The information retrieval from Big Data requires more efficient techniques for data indexing. According to the work in this paper, a Dynamic Order Multi-field Index (DOMI) structure has been introduced and implemented. The proposed DOMI indexing structure allows dynamic rather than sequential ordering of fields, in addition, compacting values with common prefixes. Hence, the DOMI allows efficiently indexing of huge data sets and answering queries that involve multi-fields, as well as, queries that involve a single field. A comparative study by building a compositefield index among the proposed DOMI and other popular indexing data structures such as B and B+ trees has been implemented. The comparison results show that DOMI composite indexing soutperform other composite indexing structures in case of answering a single field query that addresses a non-leading field. The performance of DOMI is slightly better than that of B+ in case of answering a composite-field query. Therefore, DOMI covers different cases of data retrieval more efficiently. Generally, DOMI could be the most suitable indexing structure for intensive-data such as Big Data because it is based on radix trees which compacts common prefixes into a single occurrence. Our experiments show that DOMI is effective, scalable, and efficient for supporting structural data queries.