[Paper Review] Materialized View Selection by Query Clustering in XML Data Warehouses

Hadj Mahboubi, Kamel Aouiche+1 more|ArXiv.org|Sep 11, 2008

Advanced Database Systems and Queries23 references20 citations

TL;DR

This paper proposes an automated materialized view selection strategy for XML data warehouses using query clustering to improve query performance. By grouping similar XQuery decision-support queries and materializing common subexpressions, the approach reduces execution time significantly, even for complex queries, as validated through experiments on an XCube-based XML warehouse.

ABSTRACT

XML data warehouses form an interesting basis for decision-support applications that exploit complex data. However, native XML database management systems currently bear limited performances and it is necessary to design strategies to optimize them. In this paper, we propose an automatic strategy for the selection of XML materialized views that exploits a data mining technique, more precisely the clustering of the query workload. To validate our strategy, we implemented an XML warehouse modeled along the XCube specifications. We executed a workload of XQuery decision-support queries on this warehouse, with and without using our strategy. Our experimental results demonstrate its efficiency, even when queries are complex.

Motivation & Objective

• To address performance bottlenecks in native XML database management systems used for decision-support applications.
• To reduce query execution time in XML data warehouses by selecting optimal materialized views.
• To automate the view selection process using data mining techniques instead of manual tuning.
• To evaluate the effectiveness of query clustering in identifying reusable query patterns for view materialization.
• To demonstrate the scalability and efficiency of the proposed strategy on complex, real-world XQuery workloads.

Proposed method

• Clustering the workload of XQuery decision-support queries using a data mining technique to group semantically and structurally similar queries.
• Analyzing query patterns to identify common subexpressions and structural motifs suitable for materialization.
• Selecting materialized views based on cluster characteristics, such as frequency and structural overlap.
• Implementing the XML data warehouse according to XCube specifications to ensure standardization and extensibility.
• Executing the query workload both with and without materialized views to measure performance improvement.
• Using a cost-based evaluation to prioritize views that maximize performance gain per storage cost.

Experimental results

Research questions

• RQ1 How can query clustering be effectively applied to identify optimal materialized views in XML data warehouses?
• RQ2 To what extent does query clustering reduce execution time for complex XQuery queries in XML decision-support systems?
• RQ3 What is the impact of materialized view selection based on clustered query workloads compared to random or heuristic selection?
• RQ4 Can the proposed strategy scale to large and complex XML data warehouse workloads?
• RQ5 How does the performance gain vary with different clustering granularity and view selection policies?

Key findings

• The proposed query clustering-based materialized view selection strategy significantly reduces query execution time in XML data warehouses.
• Even for complex XQuery queries, the performance improvement was measurable and consistent across the workload.
• The method demonstrated scalability and adaptability to varying query patterns and data structures.
• Materialized views derived from clustered queries provided higher reuse rates and better performance gains than non-clustered approaches.
• The experimental results on the XCube-based XML warehouse confirmed the strategy’s effectiveness in real-world decision-support scenarios.
• The approach achieved notable performance gains without requiring manual tuning or prior knowledge of query patterns.