Efficient Encoding Technique for Strings-Based Dynamic XML Labeling Schemes

Abstract

Several dynamic XML labeling schemes have been proposed to efficiently process updating in dynamic XML data. In this paper, we focus on one class of these schemes which are using strings of lexicographical order to support dynamic XML. We point out the problems of existing encodings which include memory inefficiencies when initial labeling and achieving labels of sub-optimal size when dynamic labeling. A Full-Tree-based(FT) encoding technique is proposed to overcome these problems. We bring the concept of self-increase into strings, and it assures our encoding technique labeling for initial XML with efficient memory usage, which make it possible to process large XML with limited memory. Moreover, the concept of subtraction is generalized into lexicographical order for the guarantee of achieving new labels with optimal size when XML frequently updates, which has the advantages of both reducing the storage cost and optimizing query performance. Experimental results confirm that our FT technique provides a new method which is intuitional and efficient for dynamic label schemes.