An efficient scalable parallel view maintenance algorithm for shared nothing multi-processor machines

Abstract

The problem of maintenance of materialized views has been the object of increased research activity recently mainly because of applications related to data warehousing. Many sequential view maintenance algorithms are developed in the literature. If the view is defined by a relational expression involving join operators, the cost of re-evaluating the view even incrementally may be unacceptable. Moreover, when views are materialized, parallelism can greatly increase processing power as necessary for view maintenance. In this paper, we present a new parallel join algorithm by partial duplication of data and a new parallel view maintenance algorithm where views can involve multi-joins. The performances of these algorithms are analyzed using the scalable and portable BSP1 cost model which predicts a near-linear speedup. Key words PDBMS2, materialized view maintenance, parallel incremental algorithm, data-warehouse, multi-joins, data skew, dynamic load balancing.