Adaptive real-time publish-subscribe messaging for distributed monitoring systems

Many complex distributed real-time applications, monitoring and controlling the external environment, require sophisticated processing and sharing of an extensive amount of data under critical timing constraints. In this paper, we present an adaptive Realtime publish-subscribe (RTPS) messaging service for distributed real-time applications, and four primitives are defined for implementing the adaptive RTPS. Furthermore, adaptive RTPS is incorporated into an Active real-time database (ARTDB) named Agilor by translating RTPS primitives into subscription objects and ECA rules. We also apply real-time scheduling algorithm to protocol processing for improving predictability and minimizing priority inversions during message transmission. The experimental results indicate the idea is feasible, and the current implementation shows better performance.