A new Method for Dynamic Balancing of Workload and Scalability in Multicore Systems

This paper presents a new method for dynamic balancing of workload and scalability in multicore systems. This method is called LobeNOI (Load balancing engine for NOI ‑ Notification-Oriented Inference). NOI is composed of a set of small and reactive computational entities, which communicate only when needed by avoiding both temporal and structural redundancies. The innovative aspect of the considered approach would be that computational entities carry out inference by means of collaborations based upon direct notifications. This paper actually presents a way to improve the utilization of hardware processing power available in modern computers, more precisely, in multicore systems, through distribution of NOI entities in multiple cores. The main contributions of the method are the transparency in the concurrent software design process and the better hardware utilization. Moreover, NOI presents itself as a promising alternative to compose complex, distributed and scalable applications.