4.3. Theories for Designing Self-Managing Networks

To date, many autonomic computing and/or networking architectures and frameworks have been proposed in the literature. Rather than present their architectural details, we focus on the promising theories that aid in the design of self-managing networks.

4.3.1. Design Patterns

Design patterns offer an effective method of capturing human knowledge in how to cope with specific problems from an architectural standpoint. Applying appropriate design patterns in self-managing networks not only results in intelligent organization of autonomic components at runtime, but also provides a concrete guideline to component interactions. More importantly, autonomic systems developed using design patterns are guaranteed to yield desirable system output and correct system states. In a broad sense, design patterns could be general design practices, such as the patterns proposed in []:

  • Goal-driven self-assembly: this pattern is useful for self-configuration. System configuration decisions are made a priori and are assigned to the system components as goals. When the component joins an autonomic system, it knows how to contact a service registry to obtain resources and services based on its goal description.

  • Self-regenerating cluster: two or more instances of a particular type of autonomic component are tied together in a cluster. They share the same input element and process external requests assigned by some scheduling algorithm. Instances in a cluster ...

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.