Efficient XML storage based on DTM for read-oriented workloads

Abstract

We propose an XML storage scheme based on Document Table Model (DTM) which expresses an XML document as a table form. When performing query processing on large scale XML data, XML storage schemes on secondary storage and their access methods greatly affect the entire performance. For this reason, we developed an XQuery processing scheme in which an XML document is internally represented as a set of DTM blocks and can be directly stored on secondary storage. Our scheme is tailored for read-oriented workloads, and an XML document is stored on disks as arrays of nodes. We analyzed the actual data access patterns to DTM appeared in processing XML queries, and employed the combination of informed prefetching and scan-resistant buffer management based on the analysis. Our experimental results showed that our storage scheme outperforms competing schemes with respect to I/O-intensive workloads, and our sophisticated prefetching and caching increase overall throughput without significant drawbacks.