Clustering-Based Materialized View Selection in Data Warehouses

TL;DR

This paper introduces a clustering-based framework for selecting materialized views in data warehouses, aiming to optimize query performance and storage costs through a systematic, cost-driven approach validated by experiments.

Contribution

It presents a novel view selection method using clustering and view merging algorithms combined with cost models, improving efficiency over traditional techniques.

Findings

The strategy reduces query response time significantly.

It performs well even with limited storage space.

Experimental validation confirms its effectiveness.

Abstract

Materialized view selection is a non-trivial task. Hence, its complexity must be reduced. A judicious choice of views must be cost-driven and influenced by the workload experienced by the system. In this paper, we propose a framework for materialized view selection that exploits a data mining technique (clustering), in order to determine clusters of similar queries. We also propose a view merging algorithm that builds a set of candidate views, as well as a greedy process for selecting a set of views to materialize. This selection is based on cost models that evaluate the cost of accessing data using views and the cost of storing these views. To validate our strategy, we executed a workload of decision-support queries on a test data warehouse, with and without using our strategy. Our experimental results demonstrate its efficiency, even when storage space is limited.