Efficient transmission method for mobile data synchronization based on data characteristics

Abstract

In this paper, the modified Huffman encoding method is proposed for the Open Mobile Alliance (OMA) Data Synchronization (DS) packages. The proposed method encodes each DS element and property names of personal information management system (PIMS) data based on the characteristics of DS packages. In the proposed method, the occurrence probabilities for all DS elements and the property names of PIMS data are predetermined. And then two Huffman tables are constructed. Clients and a DS server maintain these Huffman table instead of exchanging their Huffman tree. This is possible since DS elements are sequence of alphabetic symbols and the occurrence probability of each DS element is predictable. Some experiments show that, the average compression ratios of WAP Binary XML (WBXML) were 67.94% for session connection and disconnection packages, and 51.25% for data transmission packages, while the ratios of the proposed method were 70.63 and 62.57%. And the gap between two methods is ever widening as the number of synchronization data increases. Since data transmission packages are exchanged much more frequently, the proposed method based on Huffman is far more effective in mobile data synchronization.