A Data Broadcast Scheme Based on Prediction for The Wireless Environment

• Venue: ICPADS '02 Proceedings of the 9th International Conference on Parallel and Distributed Systems
• Year: 2002

Broadcasting is an essential paradigm to transport dataand support scalability in the wireless environment.Traditional index-based broadcast schemes take advantageof tree structures for saving energy, but they record toomuch index information to make the broadcast cycle short.On the other hand, the hash scheme may suffer from longerbroadcast cycles and data collision problems, resulting inlonger access and tuning time. In this paper, we propose anovel broadcast scheme, named predictive indexing, whichallows mobile clients to access data by predicting datalocations in the broadcast cycle. The new scheme not onlyhas very compact structure, but also is flexible enough toallow the server to adjust the index structure for savingenergy or reducing access time. Experimental studies showthat it outperforms the tree-based schemes 2.5 times inenergy saving, while maintaining roughly the same accesstime. It also outperforms the flexible indexing 70% and thehash scheme 80% in energy consumption.