| Summary: | In many networked applications, ranging from information dissemination services to distributed simulations and virtual worlds to file-sharing applications, a vast amount of information can potentially be accessed by a large number of widely distributed network users. In most cases, each user is interested in only a small subset of all information available. This stimulates tremendous demand for interest-based content retrieval and content dissemination. In this thesis, we first investigate interest-based content retrieval where information consumers actively seek content of interest and then initiate content transmission. We examine such a pull mechanism by analyzing various types of peer-peer file-sharing systems, in which peers first locate their files of interest and then issue requests to download matched files. We next investigate interest-based content dissemination mechanisms in which information providers publish (push) their content through a so-called publish/subscribe (pub/sub) system, and information consumers receive content of interest to them from this system. In particular, we study a method of utilizing multiple multicast groups to scope content dissemination. We also consider an alternative approach that provides content filtering along the content-delivery paths based on the requests of receivers to perform interest-based content dissemination.
|
|---|
