Research On Power Conservative Methods For Mobile Transactions Processing In Broadcasting

With the rapid development of mobile application, to improve the data access efficiency and ensure the consistency of mobile transactions have been two main research directions of mobile computing. In a mobile computing environment, improving the data access efficiency not only means reducing the data access time, but also needs to decrease the power consumption while accessing data. Data broadcast based asymmetric communication is an appropriate model in mobile computing environments. In this model, server can support simultaneous access by a huge number of mobile clients. On one hand, efficient usage of scarce bandwidth is allowed. On the other hand, mobile clients can consume less battery power. However, when data are updated and broadcasted simultaneously in a mobile application, the mobile clients may access inconsistent data. Traditional schemes cannot be applied to control concurrency in broadcast-based transaction processing. In order to take full advantage of broadcasting and still remain consistency, new concurrency control protocol is needed.This paper discusses how to use indexing technology to reduce power consumption of accessing data in broadcast channel. The basic idea is to allow a mobile client to go to doze mode after it retrieves a pointer to the required data item on air, and go to active mode while the required data item appeared on air. Meanwhile, it analyze several existing methods for broadcast-based concurrency control, and point out the limitations. The conclusion is that two main research directions of mobile computing are relatively independent. The traditional indexing methods do not support concurrency of transactions issued in mobile clients, and the existed methods for broadcast-based concurrency control do not care the problem of energy efficient. Based on the distributed indexing methods and STUBcast protocols, this paper also designs an energy efficient method for broadcast-based mobile transactions processing. Simulation results show that this method has an excellent average tuning time and a satisfactory average response time.