Brief Announcement: Parameterized Maximum and Average Degree Approximation in Topic-based Publish-Subscribe Overlay Network Design

摘要

Designing an overlay network for publish/subscribe communication in a system where nodes may subscribe to many different topics of interest is of fundamental importance. For scalability and efficiency, it is important to keep the degree of the nodes in the publish/subscribe system low. It is only natural then to formalize the following problem: Given a collection of nodes and their topic subscriptions connect the nodes into a graph which has low average and maximum degree and in such a way that for each topic t, the graph induced by the nodes interested in t is connected. We present the first polynomial time parameterized sublinear approximation algorithm for this problem.