Configuration and Optimization in Self-Managing Networks: a Tutorial

One of the key goals of network administration is configuring a network that is always operating at an optimal state. However, this goal is increasingly time-consuming and expensive as networks increase in size and complexity. One path to offsetting the increasing cost is in the implementation of a network that manages itself – known as a cognitive network. This paper will introduce the concepts of self-configuration and self-optimization as they relate specifically to cognitive networks and more generally to the discipline of artificial intelligence from a nonimplementation specific view. A possible model for the structure of a cognitive network will be presented as a result of these concepts. A minimal knowledge of network architecture is required to understand the following material. This paper is intended in part to be an introduction to the newly formed concept of cognitive networks.