| Objectives : Similar to the current trend of user generated content, the Future Internet will derive its application, services and infrastructures from user generated contributions. Hence, a major objective for the Future Internet architecture is the support of overlays for participation. The future services (e.g. networked applications or network control tasks) will be provided by large and dynamically varying communities of intelligent and coordinated edge nodes. Therefore, the edge nodes should be orchestrated with mechanisms to define overlays with a) application-specific name spaces, b) application-controllable routing and c) self-organizing procedures for resource management. Furthermore, since edge nodes will offer overlay services typically based on chains of contributions from other nodes, the provided resources (e.g. bandwidth or QoS routing services) should be verifiable in their application-specific quality. Furthermore, the resource providers have to offer their supply on small time scales while being sensitive to quality feed back and control actions. Another challenge for a participation architecture is the efficient locating and exchanging of “user-generated content”, such as phonebooks, pod- or videocasts. The question where to store the data might deeply influence the overlay and the Future Internet architecture. Centralized storage concepts are easy to control and are highly efficient while being vulnerable to overload and system faults. Distributed content stores suffer from synchronization traffic, but are more reliable. Hence, scalable, efficient, and controllable edge-based content networking mechanisms are needed. Ideally, they permit to specify the degree of centralization or decentralization at run time. In addition, Internet’s current transition from a network of networks to a network of applications leads to stringent requirements on its control mechanisms. Network control mechanisms need to be highly scalable while operating at predetermined and high quality levels. P2P-based mechanisms, for example, have demonstrated their capabilities to form highly scalable and efficient task-specific overlays. Considering the advantages of these overlays, one may think about transferring similar mechanisms from application layer to the future network layers. However, most cooperation mechanisms are based on voluntary and best-effort contributions, while control mechanisms need performance guarantees. Hence, an important objective is the development of scalable and carrier-grade cooperation mechanisms. |
Home
Site map
