Effects of routing computations in content-based routing networks with mobile data sources

Abstract

This paper presents the first quantitative evaluation of the role of routing computations on performance when mobility is introduced to a content-based routing network. Additionally, the paper identifies the factors that affect the performance of a distributed publish/subscribe architecture supporting mobile publishers, formalizes publisher mobility protocols for distributed publish/subscribe systems, and develops and evaluates protocols that reduce the costs associated with supporting mobile publishers in publish/subscribe systems. Our results show that ignoring route computation time paints a false picture of the scalability of content-based routing networks, but that with appropriate protocols the adverse effects can be mitigated.

Load Balancing Content-Based Publish/Subscribe Systems

“…For example, newspaper agencies rarely change, but the number of news readers subscribing and unsubscribing from a news service is comparatively higher. For scenarios where the above assumption about message type frequencies does not hold, protocols exist that can handle these situations [Muthusamy et al 2005]. Unadvertisement messages are used by the publisher whenever it disconnects from the system.…”

Publisher Placement Algorithms in Content-Based Publish/Subscribe

“…Muthusamy et al [31], [41] proposed several publisher migration protocols with different optimization techniques to study the effects of publisher and subscriber migration on system performance. The publisher migration protocol that we propose in this paper is different in three ways.…”

“…First, instead of rebuilding the advertisement tree rooted at the new broker, we simply revise the last hop of the existing advertisement only on brokers along the migration path as in [31]. In terms of overhead message count, our approach generates O(log N) messages, whereas the approach in [41] generates O(N) messages, where N is the total number of brokers in a treeoverlay network with typical fan-out greater than one. Second, the advertisement/subscription tree rebuilding period is known in our approach.…”

“…This allows our publishers to know precisely the earliest time to resume publishing at the new broker with assurance that those messages will be delivered to all matching subscribers in the network. Third, the objective in [41] is to analyze the impact of supporting mobile publishers on system performance, whereas here, our objective is to minimize average end-to-end delivery delay and system load …”

On interest locality in content-based routing for large-scale MANETs

“…Content-based routing is intrinsically data-centric [17] and human-centric [25], i.e., contents are routed based on consumers' specified interests while consumers' behaviors follow some kind of social features such as mobility, community, membership, locality, etc. In [15], McPherson et al identified the homophily theory through hundreds of case studies to prove the community and homophily feature of human society.…”