Content-based publish/subscribe is a powerful data dissemination paradigm that offers both scalability and flexibility. However, its nature of high expressiveness makes it difficult to analyze or predict the behavior of the system such as event delivery probability and end-to-end delivery delay, especially when deployed over unreliable, best-effort public networks. This paper proposes the analytical model that abstracts expressiveness nature of content-based publish/subscribe, along with uncertainty of underlying networks, in order to predict quality of service in terms of delivery probability and timeliness based on partial, imprecise statistical attributes of each component in the system. Furthermore, the paper leverages the proposed prediction algorithm to implements heuristic-based subscriber admission control algorithms to maximize system utility when the system cannot support all subscribers. The evaluation results yields good prediction accuracy and admission rates.
展开▼
