I-DG and HI-DG: Scalable and Efficient Techniques for Subscription-Based Data Dissemination via IP Multicasting

摘要

This paper proposes the IP Multicast-enabled Drop Groups (I-DG) and the Hybrid I-DG (HI-DG) techniques for solving the problem of efficiently and securely disseminating information to a large number of subscribers in the Internet. Possible application areas include stock market information distribution, refreshing the caches of edge servers on the Internet, pay-per-view information delivery. The proposed I-DG technique uses an appropriate IP multicasting protocol as the base and addresses the problems, such as key distribution and subscription, introduced by large scale data dissemination. I-DG provides dynamic and secure subscription to multicast data, and proposes a solution to content organization in broadcasts. The HI-DG technique employs a combination of multicasting and unicasting to reduce the packet load on the network. Simulation shows the efficacy of the two methods. I-DG performs the best when most of the clients subscribe to the same set of items, and HI-DG is best when a subset of data items is more popular.