Index Challenges in Native XML Database Systems

Abstract

Today, more and more enterprises process XML data, many of them already in XML database systems. Once systems start to grow in size, scalability becomes an issue. One of the core operations during insert processing is the index maintenance. Typical relational systems only have few indexes per table, however, XML database users already look at creating hundreds of XML indexes per XML column. A similar situation is present during query processing when many indexes are present which the system can choose from and has to combine them for optimal performance. This creates new challenges for the database system - unknown in the traditional relational world. In this paper, we discuss those index challenges and possible solutions.