Cognitive Management Systems

OneFIT applies, through the CMONs and the CSCIs, cognitive techniques (e.g., [1], [2], [3], [4], [5]) for the management of the opportunistic networks and for coordinating the infrastructure. This innovation will lead to robustness and dependability. The approach capitalizes on the self-management features and also on the learning capabilities that must be intrinsic to cognitive systems.

Figure 1 shows the capabilities (in terms of structure and functionality) of a cognitive management system.

Figure 1. Structure of a cognitive management system (click image to enlarge)


A OneFIT cognitive management system hosts and implements capabilities for: (i) context acquisition and reasoning, profile management, and policy-awareness; (ii) the cooperation with other cognitive management systems, through the exchange of profiles, policies and context information; (iii) building and sharing knowledge, which, in principle, refers to the situations (contexts) typically encountered, the policies applied, the optimization decisions taken, and the resulting efficiency achieved; (iv) decision-making through cross-layer optimization functionality that takes into account the context of an operation, the profiles, the policies (potentially, of various business level stakeholders), and the acquired knowledge and experience.

The operational context describes aspects like the: (i) geo-area and time period in question; (ii) applications requested; (iii) mobility levels; (iv) radio quality; (v) element or device status. The profile component provides information on the capabilities of devices and infrastructure-elements, the characteristics of applications, and the requirements and preferences of users. The policies provide rules for context handling, in terms of objectives to be achieved, strategies to be used for the optimization, and constraints to be respected.

CMONs and CSCIs will have functionality and will collaborate for performing the following tasks:

  • Determination of the suitability of the opportunistic network approach. This includes node/infrastructure discovery, identification of candidate nodes, identification and generation of spectrum opportunities from the infrastructure side, interference coordination through the exploitation of results from off-line studies.

  • Opportunistic network creation. This includes the selection of the optimal, feasible configuration of the opportunistic networks. A configuration includes the selection of participant nodes, spectrum and routing pattern.

  • Opportunistic network maintenance. This involves QoS control (monitoring and corrective actions) of the data and control flows served by the opportunistic network, and the realization of reconfiguration actions in the case of alterations in the node status, and the spectrum and routing conditions.

  • Handling of forced terminations of the opportunistic network. This means to try to preserve the provision of applications as much as possible, even when the opportunistic network has to be terminated before the data session ends.

[1] End-to-End-Efficiency (E3) project Website,
[2] R. Thomas, D. Friend, L. DaSilva, A. McKenzie, "Cognitive networks: adaptation and learning to achieve end-to-end performance objectives", IEEE Commun. Mag., Vol. 44, No. 12, pp. 51-57, Dec. 2006
[3] J. Kephart, D. Chess, "The vision of autonomic computing", IEEE Computer, Vol. 36, No.1, pp. 41-50, January 2003
[4] P.Demestichas, G.Dimitrakopoloulos, J.Strassner, D. Bourse, "Introducing reconfigurability and cognitive network concepts in the wireless world", IEEE Vehicular Technology Mag., Vol. 1, No. 1, pp. 33-39, June 2006
[5] P.Demestichas, D.Boscovic, V.Stavroulaki, A.Lee, J.Strassner, "m@ANGEL: autonomic management platform for seamless wireless cognitive connectivity to the mobile Internet", IEEE Commun. Mag., Vol. 44, No.6, pp. 118-127,June 2006