Selective Delivery of Event Messages in Peer-to-Peer Topic-Based Publish/Subscribe Systems

A distributed system is considered in an event-driven model where states of processes transit on occurrence of events. In this paper, we discuss a peer-to-peer (P2P) model of a topic-based publish/subscribe (P2PPS) system where each peer process (peer) p_i can both subscribe a subscription p_i.S and publish an event message e with a publication e.P. The subscriptions and publications are specified in terms of topics, i.e. keywords. An event message e is notified to a target peer p_i whose subscription p_i.S includes a common keyword with the publication e.P. Each event message e carries a vector e:KV = 〈kv₁, . . . , kv_m〉 of keywords k₁, . . . , k_m. An event message e₁ causally precedes an event message e₂ with respect to a subscription S_i iff not only e₁ causally precedes e₂ but also e₁·kv_h <; e₂·kv_h for every keyword k_h in the publications e₁.P and e₂.P and the subscription p_i.S. A pair of event messages e₁ and e₂ are unnecessarily ordered if e₁.KV <; e₂.KV but e₁ does not causally precede e₂. In this paper, we newly propose a keyword vector (KV) protocol where the linear clock or physical clock is used with keyword vector to reduce pairs of messages unnecessarily ordered. We evaluate the KV protocol and show the number of pairs of unnecessarily ordered messages is reduced.