Optimal Update Policies for Distributed Materialized Views

In this paper we present an analysis of the problem of determining optimal policies for updating distributed materialized views. We demonstrate the general application of materialized views, and define the concept of materialized view currency and allow a query to specify its currency requirement. We also allow a materialized view to be updated from either a base relation or another materialized view. This flexibility provides an opportunity for further reduction in the cost of maintaining distributed materialized views. We model the problem of optimal update policies to capture currency and policy constraints, replicated data, and various view update policies. The optimization incorporates a minimum-cost objective function as well as user's response time constraints.

• 저널Management Science · 37(7) · 851–870
• 토픽Distributed systems and fault tolerance · Computer Networks and Communications
• DOI10.1287/mnsc.37.7.851
• 저자Arie Segev, Weiping Fang