A popular trend in large enterprises today is the adoption of green IT strategies that use resources as efficiently as possible to reduce IT operational costs. With the publish/subscribe middleware playing a vital role in seamlessly integrating applications at large enterprises including Google and Yahoo, our goal is to search for resource allocation algorithms that enable publish/subscribe systems to use system resources as efficiently as possible. To meet this goal, we develop methodologies that minimize system-wide message rates, broker load, hop count, and the number of allocated brokers, while maximizing the resource utilization of allocated brokers to achieve maximum efficiency. Our contributions consist of developing a bit vector supported resource allocation framework, designing and comparing four different classes with a total of ten variations of subscription allocation algorithms, and developing a recursive overlay construction algorithm. A compelling feature of our work is that it works under any arbitrary workload distribution and is independent of the publish/subscribe language, which makes it easily applicable to any topic and content-based publish/subscribe system. Experiments on a cluster testbed and a high performance computing platform show that our approach reduces the average broker message rate by up to 92% and the number of allocated brokers by up to 91%.

Many publish/subscribe systems implement a policy for clients to join to their physically closest broker to minimize transmission delays incurred on the clients' messages. However, the amount of delay reduced by this policy is only the tip of the iceberg as messages incur queuing, matching, transmission, and scheduling delays from traveling across potentially long distances in the broker network. Additionally, the clients' impact on system load is totally neglected by such a policy. This paper proposes two new algorithms that intelligently relocate publishers on the broker overlay to minimize both the overall end-to-end delivery delay and system load. Both algorithms exploit live publication distribution patterns but with different optimization metrics and computation methodologies to determine the best relocation point. Evaluations on PlanetLab and a cluster testbed show that our algorithms can reduce the average input load of the system by up to 68%, average broker message rate by up to 85%, and average delivery delay by up to 68%.

Distributed content-based publish/subscribe systems suffer from performance degradation and poor scalability caused by uneven load distributions typical in real-world applications. The reason for this shortcoming is the lack of a load balancing scheme. This article proposes a load balancing solution specifically tailored to the needs of content-based publish/subscribe systems that is distributed, dynamic, adaptive, transparent, and accommodates heterogeneity. The solution consists of three key contributions: a load balancing framework, a novel load estimation algorithm, and three offload strategies. A working prototype of our solution is built on an open-sourced content-based publish/subscribe system and evaluated on PlanetLab, a cluster testbed, and in simulations. Real-life experiment results show that the proposed load balancing solution is efficient with less than 0.2% overhead; effective in distributing and balancing load originating from a single server to all available servers in the network; and capable of preventing overloads to preserve system stability, availability, and quality of service.