Estimation of round trip time in distributed real time system architectures

Real-time distributed systems often utilize centralized nodes to manage and coordinate the exchange of information; in particular, for users to are unable to communicate directly because of their location behind firewalls or proxy servers. Identifying these nodes is crucial for successful operation of the entire distributed system. The geo-locational placement of these relay node can be determined based on network properties such as Round Trip Time (RTT). In this paper a real-time e-learning system - Remote Access Laboratories (RAL) are discussed as an example of a distributed real-time system. RTT is an important factor directly impacting the Quality of Experience (QoE) of users. In a distributed RAL environment, there may be multiple users who can be both learners (service users) and providers in an activity. As such, the equipment and corresponding learning materials are not concentrated at one location, but spread over multiple nodes. A method to determine the suitable node or relay based on population distribution in a geographic region and projected RTT based on distance between them is proposed here. A mathematical model and simulation data are presented to support the effectiveness of the method. The results are compared with peer-to-peer network model, which is optimal, to establish the difference in performance.