New generation of transport protocols for autonomous systems

Requirements and constraints of the communication services required by autonomous systems raise new challenges to the traditional transport protocol layer. Indeed, classical transport protocols (i.e. TCP and UDP) were well dimensioned for the standard distributed applications operating over best-effort network services. Today, more sophisticated transport protocols are required in order to self-configure and self-adapt to varying network resources (e.g. wireless network systems) guided by the requirements and preferences of both the autonomous systems and their applications. This new generation of transport protocols should follow a component-based and service-oriented approach intended to smoothly incorporate current and future protocol components and their specialization, based on the semantic associated to their functions and capabilities. Moreover, these transport protocols should be designed following an autonomic computing approach in order to incorporate autonomic managers in charge of continuous network monitoring, anomalies detection, selection and enforcement of correction strategies to adapt transport components. Furthermore, competing autonomic managers operating over shared and resource-limited networks should be coordinated by higher-level managers in order to better share the available resources. The autonomic management of these resources should be done based on the systems and applications requirements and priorities in order to offer a collaborative orchestrated transport service. This paper presents a framework for defining autonomic transport composite services aimed at providing the basis for designing and developing this new generation transport layer. This framework is intended to dynamically compose reusable transport components based on the semantic associated to each component and guided by the final service expected from the composite architecture. An ontology-driven and service-oriented architecture approach has been followed in order to - - integrate autonomous systems requirements as well as the characterization of autonomic transport components in order to guide the composition and the service selection processes.