Self-Maintainability of Graph Structured Views

Materialized views need to be maintained in response to changes of base data. It is desirable to do this maintenance without accessing base data, to reduce overhead and contention for base data. In this paper we study the self-maintenance problem for materialized views defined over a graph-structured database (GSDB). We formally define the notion of self-maintainability for such views, and show how to check if a base update can be incorporated into a view without referencing base data. We also characterize views that are always self-maintainable, and compute the auxiliary information that needs to be stored in order to make a view self-maintainable. 1 Introduction Materialized views copy selected or summarized data of interest to improve query performance and data availability. They are especially useful when the query site is separate from the base data, e.g., views cached for Web browsing, or stored in a data warehouse. In such an environment, materialized views can be stored close ..