Skip to Main Content
To take advantage of the globalization, many current applications use the web browser and java technology as a natural interface to the world. This fact makes necessary to design applications with some form of transactional properties. In particular, storage systems and access methods must cope with concurrency and recovery aspects in order to achieve reliability and to provide the expected services to these applications. However, very few proposals on index structures and access methods focus on this important point. In this paper we present concurrency and recovery mechanisms developed for the Q-tree, a multidimensional index structure which is a relevant part of the search engine used in a knowledge management-oriented application called Xerka.net. We hope that the original ideas on which our proposal is based on may contribute to enhance the current state of art of the access methods -in general- and of the multidimensional indexing structures -in particular.