The Design and Implementation of Service Reservations in Real-Time SOA

Service-oriented architecture (SOA) provides the flexibility of dynamically composing business processes in enterprise computing. However, they must be enhanced to support real-time activities in future SOA applications such as media streaming, control applications, cyber-physical, and intelligent vehicle systems. In this paper, we present the RT-Llama SOA middleware to support predictability in real-time processes. Given a user-specified service process and deadline, the RT-Llama resource management system will reserve resources in advance for each service in the process to ensure the process can meet its end-to-end deadline. RT-Llama provides a global resource manager (GRM) to compose and reserve services in a business process. The GRM is facilitated by an efficient data structure that keeps track of the utilization of local resources. The data structure, called dTBTree, is updated by each local resource manager (LRM) periodically and cached in the Host Utilization Repository in GRM. We have implemented RT-Llama's resource management components so that they can efficiently adapt to dynamic real-time environments.