Research Of Technology About Data Broadcast In Mobile Computing

The emergence and popularization of wireless communications, mobile device and Internet accelerate the development of mobile computing. Different from traditional distributed computing environment, the mobile computing environment has some distinct features: mobility, disconnection, weak connectivity, resource scarcity, diversity and asymmetry of network communication etc. These features restrict the use of existing classical database technologies in mobile environments, and produce new research fields. Due to the good expansibility, data broadcast is used in mobile computing widely. In this paper, several key technologies about data broadcast are investigated.Data scheduling is the most important issue in data broadcast. In this paper, the push-based data scheduling policies is investigated in detail and a multidisk scheduling algorithm is analyzed. This algorithm is easy to apply contrasting to traditional scheduling policies. On-demand channel is always used in many broadcast systems for individuation of users, which usual has a faster response time, so it should be used in full. On the other hand it could not be overloaded. In this paper, an adaptive balance algorithm between on-demand channel and broadcast channel is investigated, which can adapt to the change of system's load.In order to reduce query latency, mobile device should cache "hot" data. To maintain the consistency of mobile cache, system always broadcast invalidate report. If the invalidate report is long then the delay will be great. In order to reduce the length of invalidate report, a consistency policy based on data absolute validity interval is proposed. A prefetch policy based on d-value is future investigated to improve the hit ratio of cache. This policy can adapt to the energy of mobile device.In order to save energy, the mobile should reduce access to wireless channels. So index broadcasting is always adopted. Due to the data access skew, index technology based on data access probability is proposed. At last ORD algorithm is proposed to create optimal broadcast sequence according to given index tree.