Cognitive Networks: Towards Self-Aware Networks Review

Average Reviews:

(More customer reviews)There are several questions that immediately come to mind when contemplating the notion of a cognitive network. Some of these include:
1. In what sense are cognitive networks "intelligent", thus justifying the adjective "cognitive"? If these types of networks are intelligent, is there a methodology for quantifying the intelligence of these networks that would allow for example a partial or even well-ordering of cognitive networks according to their intelligence? Quantifying artificial intelligence is of both academic and practical interest, since in the former it will permit an ongoing assessment as to how far researchers have progressed in obtaining machine intelligence while in the latter it will allow network managers to assess the costs of deploying machine intelligence.
2. Do cognitive networks have an "artificial immune system", i.e. a strategy or collection of strategies for recognizing malicious attempts to disrupt or suppress the data flows in these networks or gain access to sensitive information that exists on them? And if so, is this artificial immune system aggressive in the sense that it will take action against intrusion or other forms of malicious activity? Or is it passive, i.e. merely alerting the system administrator to its presence? There are two articles in this book that address these questions, at least to some extent, although the expression "artificial immune system" is not used. 3. Do cognitive networks need the ability to "think in many domains", i.e. do they need to be capable of processing (and differentiating) information originating in my different contexts? This question is partially addressed in the article entitled "The Semantic Side of Cognitive Radio" wherein the authors discuss the need for a common language that radios will have to share in order to communicate effectively.
4. To what extent are cognitive networks able to recognize novel information that is presented to them? As an example, cognitive networks will need to think about network traffic, which of course is always changing. 5. What are the computational requirements of cognitive networks? What advantages do cognitive networks have over (non)-cognitive ones? For example, what performance hit will one take by deploying a cognitive structure over a wireless network?
6. Are cognitive networks able to troubleshoot themselves, i.e. find the cause(s) of problems or issues within the network itself? Are there any operational or commercial examples of cognitive networks that have this ability and how does it compare with networks whose problems are troubleshot with human engineers? The article entitled "Self-Managing Networks" addresses this question in the context of what its authors refer to as "self-healing". They define this as the ability to automatically detect, diagnose, and recover from faults. One would expect that a network that could troubleshoot itself would use some sort of inductive or abductive reasoning patterns in order to find out the cause(s) of problems on the network. The authors do not discuss induction or abduction in this article, but instead focus on what they call "design patterns". These are methods that emulate how humans deal with specific problems on a network. The authors also discuss inductive reasoning patterns in this article, which they define to be the ability to generate training instances using background knowledge or simply by providing explanations using regularities in the data. The rise of inductive logic programming in the last two decades has brought some of this discussion to fruition, and there are commercial systems that use hybrids of inductive and abductive reasoning patterns, although these systems are not discussed in this article or anywhere else in the book.
7. Are there any commercial examples of cognitive networks that are working out in the field right now? What kind of technology is required to implement and maintain a cognitive network in practice? This book does not include any discussion of commercially available cognitive networks, but it does include some discussion of the theoretical apparatus that has been developed for the design of such networks. 8. To what degree are existing networks "cognitive" in the sense that they have capabilities or properties that were not really intended to be "intelligent" by their designers but nevertheless deserve to be characterized as such? This question is not addressed in the book, but there are sound arguments that can be made that some characteristics of modern networks are intelligent without this being the intent of the designers.
9. Does the modeling and simulation of cognitive networks present any special challenges over and above that of ordinary networks? The answer to this question is relevant since, as is the case for ordinary networks, the decision to deploy cognitive networks will typically involve the use of modeling and simulation to a large degree. None of the articles in this book discuss the simulation and modeling of cognitive networks.10. Are cognitive networks more secure than ordinary ones, or does their complexity raise even more issues for security? There are two articles in this book that address the security issues in cognitive networks, both of them dealing with security in the abstract, with no real case studies given (although certain cognitive network architectures are discussed). One of the articles, entitled "Security Issues in Cognitive Networks" points to "spectrum handoff" as being a particular concern for security in these types of networks. Ordinary ad hoc mobile networks do not have this capability. Even though it does not discuss real cases studies of attacks on cognitive networks, this article is worth studying because it discusses those attacks that are unique to these kinds of networks. The other article that addresses security entitled "Intrusion Detection in Cognitive Networks" does not really point to any security weaknesses in cognitive networks that one could find in ordinary networks. If one is to deploy cognitive networks, one expects that their intelligence would make them very resilient to attacks, over and above what one would find in ordinary networks. If they are even more vulnerable to attacks, this would vitiate their utility in real-world applications.
11. How resistant will the network engineering and business community be to the implementation of cognitive networks? The incorporation of intelligent technology has met resistance in the last fifty years, but this resistance has decreased considerably as the technology is proved viable and profitable.

Click Here to see more reviews about: Cognitive Networks: Towards Self-Aware Networks

Cognitive networks can dynamically adapt their operational parameters in response to user needs or changing environmental conditions. They can learn from these adaptations and exploit knowledge to make future decisions.
Cognitive networks are the future, and they are needed simply because they enable users to focus on things other than configuring and managing networks. Without cognitive networks, the pervasive computing vision calls for every consumer to be a network technician. The applications of cognitive networks enable the vision of pervasive computing, seamless mobility, ad-hoc networks, and dynamic spectrum allocation, among others.

In detail, the authors describe the main features of cognitive networks clearly indicating that cognitive network design can be applied to any type of network, being fixed or wireless. They explain why cognitive networks promise better protection against security attacks and network intruders and how such networks will benefit the service operator as well as the consumer.
Cognitive Networks
Explores the state-of-the-art in cognitive networks, compiling a roadmap to future research.
Covers the topic of cognitive radio including semantic aspects.
Presents hot topics such as biologically-inspired networking, autonomic networking, and adaptive networking.
Introduces the applications of machine learning and distributed reasoning to cognitive networks.
Addresses cross-layer design and optimization.
Discusses security and intrusion detection in cognitive networks.

Cognitive Networks is essential reading for advanced students, researchers, as well as practitioners interested in cognitive & wireless networks, pervasive computing, distributed learning, seamless mobility, and self-governed networks.

With forewords by Joseph Mitola III as well as Sudhir Dixit.

Click here for more information about Cognitive Networks: Towards Self-Aware Networks