An XML Publish/Subscribe Algorithm Implemented by Relational Operators

Abstract

Abstract. An XML publish/subscribe algorithm needs to store large numbers of XPath or XQuery subscriptions and match subscriptions with published XML documents. Since the number of the subscriptions may be very large, the performance and the scalability of the algorithm may be critical. The scalability of the method of constructing a large finite state automata or decision tree for all subscriptions is limited by amount of available physical memory. In this paper, we will introduce an XML publish/subscribe algorithm which is consisted of the publish algorithm, subscribe algorithm, and matching algorithm. The matching algorithm uses relational operators to match subscriptions with publications inside a relational database, so the scalability of the algorithm is no longer limited by amount of available physical memory. Experimental results show that the matching algorithm has very good performance and scalability.

An Investigation on Flexible Communications in Publish/Subscribe Services

“…The widely-adopted tree-based format is XML, which specifies the structure of data content by a combination of opening and closing tags. In fact, there has been an increasing demand for XML-based pub/sub services, which support flexible document structures and subscription rules expressed by powerful languages, such as XPath and XQuery [5].When using XML, the publisher transforms a data instance in an XML document by placing the content of its fields between tags with the same name of the given field. Such document is further converted into a byte stream and delivered to the subscriber, which uses such XML document to feed a data instance by assigning to each field the value between tags equal to the field name.…”