Abstract-

The scope of this paper is to introduce an innovative paradigm for cognitive self-managed elements of the Future Internet. The present Internet model is based on clear separation of concerns between protocol layers, with intelligence moved to the edges, and with the existent protocol pool targeting user and control plane operations with less emphasis on management tasks. Future Internet shall be engineered based on cognitive behavior with a high degree of autonomy, by proposing the operation of selfmanaged Future Internet elements around a novel hierarchical feedback-control cycle. The concepts are based on a hierarchical Distributed Cognitive cycle for System & Network Management (DC-SNM) which aims at facilitating the promotion of distributed management. The management approach encompasses a hierarchical distribution of cognitive cycles, breaking down the execution and decision making levels to (autonomic) network elements, network domain types and up to the service provider realm in order to address management, dynamic organization and (re)configuration of future internet elements.

No abstract

Thanks to exponential growth rates, the Internet has doubled in size many times since its inception, nearly five decades ago. However, this tremendous growth has brought on an accumulating complexity in management tasks, thus undercutting the Internet's manageability and proving detrimental to further development. To address this predicament, the concept of a capacity for autonomic operation was proposed and later extended with artificial intelligence features to thus converge on the vision of a self-managed Future Internet. The capacity for self-management builds on specific technological enablers, including a network monitoring infrastructure, mechanisms for decision making and decision appraisal as well as advanced artificial intelligence features (e.g., machine learning). This paper establishes the role of these capacities in realizing a self-managed mode of operation for wireless Internet systems. We present a case study where self-management features resulting from the combined application of device and network monitoring and decision making are employed in a solution of a well-known planning and management problem in wireless networking.