Efficient data retrieving in distributed data-streaming environments

Abstract

In a potential distributed application, automobile tracking system (ATS), automobile location data is continuously generated, kept in a distributed manner. As large amount of traffic will be incurred during search process, it is critical to construct an efficient overlay multicast structure for the ATS so as to distribute traffic to all the physical links evenly, as well as balance the load among group members. In this paper, we propose a distributed protocol, MMT scheme, in which end hosts self-organize into multiple multicast trees. We evaluate the performance of MMT with comprehensive simulations. Experimental results show that MMT outperforms existing approaches in load balance, and its performance penalties are low from the network and the application perspectives.