Exploring Statistical Information for Applications-Specific Design and Evaluation of Hybrid XML storage.
Lena Strömbäck, Valentina Ivanova, David Hall · Databases, Knowledge, and Data Applications · 2011
Modern relational database management systems provide hybrid XML storage, combining relational and native technologies. Hybrid storage offers many design alternatives for XML data and in this paper we explore how to aid the user in effective design of hybrid storage. In particular we investigate how the XML schema and statistical information about the data can support the storage design process. We present an extended version of our tool HShreX that uses statistical information about a data to enable fast evaluation of alternative hybrid design solutions. In addition we show the benefit of the approach by a first evaluation where we discuss how the tool aids in the storage design and evaluation process. on are the XML schema structure and statistical properties of the data set. We first give some further motivation and discuss the goals of our work. This is followed by a discussion of properties and measurements relevant for storage design. We then present a tool that enables fast evaluation and exploration of storage solutions and present a first evaluation to show the feasibility of the tool. The paper is summarized by presenting our future vision. Our long term goal with the work is to present a method that can suggest a set of plausible hybrid storage models for an application.