Update-oriented graph-relational storage and consistency

Maintain consistency across base records, topology structures, and auxiliary indexes in graph-relational database systems while preserving update and query efficiency.

Background

The GRBench U-series evaluates vertex and edge insertions and deletions across graph-relational database systems with different topology representations, including table-backed graphs and RID- or adjacency-based structures. The results show that update costs and supported operations vary substantially across systems because updates may need to maintain base tuples, topology structures, endpoint references, and auxiliary indexes. The unresolved problem is therefore to coordinate these structures without sacrificing either consistency or operational performance.