Distributed Strategic Learning for Building Network Embedded Coding to Achieve Reliable Event Notification

Event-driven communications characterize several key industrial projects for realizing future federated architectures, and publish/subscribe services are considered as key building blocks in such projects. Most of these projects are also affected by strong requirements for the reliability and timely data sharing, since involved in providing decision support for critical activities. In the current literature, reliable multicast is always achieved at the expenses of violations of temporal constraints, since retransmissions are used to recover lost messages. In this paper, we present a solution to apply a proper coding scheme so as to jointly achieve reliability and timeliness when multicasting over the Internet. Such a solution employs game theory so as to select the best locations within the multicast tree where to perform coding operations. Our driving idea is to take a distributed strategic learning in order to determine such locations without a formal characterization of the payoff functions within the non-cooperative game. We prove the quality of this solution by using a series of simulations run on OMNET++.