We propose a novel access control system for graph-based models, which supports schema constraints and constraint rules to protect the data, as well as user context rules. We consider systems with huge volumes of data, where the efficient processing of aggregation operations is of paramount importance. To comply with this goal, we introduce an architecture with modules for rewriting, planning and executing queries in parallel, respecting the access constraints. Performance tests show the efficiency of our distributed query processing mechanism. Compared to a centralized approach, it reduces execution time from 25% to 68% for conjunctive queries and from 12% to 59% for queries involving aggregation.
展开▼
