Constructing XML Views over Binary Data for XML-based Data Integration

摘要

XML has been widely used for exchange, integration and management of data. Therefore, there is a trend that applications should support XML-formatted data. However, their primary data are still encoded in application-specific binary formats because of backward compatibility, I/O performance, and space efficiency. When we can expand the scope of XML technologies to such binary data, they can be shared by many areas beyond the original application area. This paper discusses XML views over binary data to make the existing XML technologies applicable to them. We first introduce an XML schema language to define an XML view over such binary data. Then, W3C DOM API is provided for them. We show an implementation technique of W3C DOM API which enables partial instantiation of DOM nodes to limit amount of consumed memories. The paper also shows techniques to realize efficient accesses to values based on properties of binary data. In addition, techniques to improve XPath query processing will be mentioned.