Proxy-based mobile data caching in cellular systems

Abstract

Mobile Data Access (MDA) has been a more and more popular service along with the development of mobile technologies. In the predictable future, 3G technologies will be widely used and relevant data services will develop very fast. One particular problem in providing data services to mobile users is the service latency. In particular, when a large number of mobile devices simultaneously send requests to data servers over the Internet, their requests will be queued up at the server side, which results in long service latency. This thesis mainly focuses on how to enhance global performance of cellular systems from the perspective of Quality of Service (QoS), and proposes a proxy-based caching approach to reduce service latency. The whole process is divided into three steps: user pattern analysis (including user movement pattern and user access pattern), cache allocation and cache migration. By transforming user movement sampling data to frequency domain, we get some movement period information, based on which we try to predict user future locations. We suggest a new concept of role, to help analyze user access in the mobile environment in which user access patterns may vary and become dependent on the various locations. Then we predict user requests by clustering all the roles. Cache allocation and cache migration are based on a novel cache management approach, named PCM (proxy cluster-based model), whereby we cluster proxies together and adopt inner-migration within a cluster. Experiment results show that our approach achieve a good performance with an acceptable computational complexity. Comparisons with some existing approaches are done on the user movement pattern analysis, user access pattern analysis and PCM approaches; furthermore, the experimental results of the whole approach combining the three steps also show that our proxy-based caching management approach can reduce the service latency significantly.

Date of Award	3 Oct 2005
Original language	English
Awarding Institution	City University of Hong Kong
Supervisor	Qing LI (Supervisor)