3.5. Discussion

The process of adapting to changes in dynamic factors that affect networks involves first detecting and then reacting to the changes. Detecting a change refers to collecting information about a dynamic factor and reacting to a change refers to the action taken by a network in order to improve network performance or maintain service continuity after detecting the change.

Detecting changes is critical to the effectiveness of a network's adaptation scheme because a network's ability to adapt greatly depends on the information collected while detecting a change. The accuracy of the collected information determines the correctness of the network's reaction, and the latency and cost of the information collection impacts various characteristics such as the convergence time and overhead of the whole adaptive scheme. Generally, a node may collect the information either actively or passively. A node may actively collect information about dynamic factors by explicitly sending messages to other nodes. For example, a node in AntNet uses agents to probe the quality of a specific path and updates the probability of forwarding along that path based on its quality. A node may passively collect information about dynamic factors by listening to the network to infer or estimate network state without explicitly requesting such information. For example, intelligent search learns about the availability of application data (i.e., data distribution) and the user behaviors (i.e., query ...

Get Cognitive Networks: Towards Self-Aware Networks now with the O’Reilly learning platform.

O’Reilly members experience books, live events, courses curated by job role, and more from O’Reilly and nearly 200 top publishers.