2. The external environment
The user demands
The user demands The internal environment
Figure 1. The environment range that communication network sensed
Stage one: the network senses current network (SLA), making the reasonable and correct response to the
environments, including external and internal environments. new user demands.
Each node in the communication network can only
monitor the local network state information, how to make III. THE RESEARCH STATES OF COGNITIVE
each node know the whole network state information is a COMMUNICATION NETWORK
problem to be solved firstly. Driven by the process of information globalization, a
Stage two: the communication network is clearly aware large lots of research works on the communication network
of the current demands and the occupied network technology were carried out. The U.S. National Science
resources of various users. Foundation provided 3 billion for the supporting of 100 ×
How to dynamically obtain the global information of the 100, FIND and GENI projects. In Europe, the financial
occupied network resources, and make the communication supports more than 9.1 billion Euro was afforded to the
network rapidly estimating the usages of network resources, researches of ICT (information and communication
is another great challenge. technology). In China, the National High-Tech. R&D
Stage three: dynamically allocates the resources and Program and 973 Program afforded to the researches on this
reconfigures the network adapting to the demands of new kind of communication network technology, and takes it
users with the correct and reasonable response.. as an important direction in information areas, considering
the technologies of cognitive communication network and
reconfiguration as the hot spots problems. Overall, the
U ser dem ands present researches on the cognitive communication network
are mostly about the exploratory on cognitive capabilities,
and are still in its infancy. The cognitive communication
environments
environments
network technology is still an open issue, and the main
External
Internal
Learning and
intelligent decision
research projects are as follows:
(1) The Cognitive Networking (CogNet) Project funded
by the U.S. National Science Foundation (NSF)
The Future Internet Design (FIND) is a long-term
M anagem ent and control of research project afforded by U.S. NSF in 2004. The goal of
netw ork resources
FIND is to achieve the next generation Internet, and its core
features include: security, robustness, manageability, the new
Reconfigurable computing model, integration of new network technology,
netw ork elem ents high standards of service architecture and the new network
architecture theories. The Global Environment for Network
Figure 2 the cognitive technology of communication Innovation (GENI) is officially announced in August 2005
network by U.S. NSF as an action plan. The goal of GENI is to build
a test bed to accommodate a variety of new architecture, and
Communication network will be based on the sensed explore the framework more suitable for the future
network environments, user demands, and the occupied computing environment. The researches of GENI include:
network resources, the communication network does some rethinking the functions of network; designing new network
necessary reconfiguration based on the Services Agreement architecture and services; designing the key capacity of
3. network security and new supporting technologies. As part IV. THE ARCHITECTURE AND TECHNOLOGIES FOR
of the FIND and GENI plan, the main goal of cognitive COGNITIVE COMMUNICATION NETWORKS
Networks technology is to design and achieve the cognitive
network architecture and protocols, develop the prototype 4.1 The architecture of cognitive communication
system of cognitive network protocol stack, and its main networks
work include: As shown in Figure-3, the cognitive communication
networks is composed three parts, including the demands
Adaptive physical layer and programmable media
mapping and policies generating unit, the network cognitive
access control (MAC) layer; and reconstruction unit and the programmable
Adaptive frequency readjusting protocol for communication network.
transmission link; The demands mapping and policies generating unit: On
Unified control layer for monitoring, configuring, and the one hand mapping the demands from different users into
adjusting the data layer module; different end-to-end network goals (specific network
behavior), on the other hand merging the local network
Intelligent network layer protocol for the functions of environmental information sensed from the underlying
topology discovery, service discovery, addressing, communication network to form the global network resource
naming and dynamic routing. information, and then generating the necessary network
(2) The software programmable intelligent network reconfiguration policies based on the relevant service level
(SPIN) project in Stanford University agreements (SLA). The number of the policies generating
Stanford University proposed the software programmable unit in the network is in small quantities, and can be
intelligent network (SPIN) concept, based on the 9 kinds of deployed in central or distributed pattern.
basic problems existing in the current IP networks, and The network cognitive and reconstruction unit: On the
extracted the advantages in the architecture of IP networks, one hand completing the sensing of network environment,
PSTN, cellular networks and ad hoc networks to form a including network topology, the transmission link status,
novel network architecture. The SPIN is essentially a network node state information, network security status, and
network operating system, with the similar status as the geographical and meteorological information, which are
operating systems in computer, and possesses a 3-layer translated into the local sensing information by the cognitive
structure including the transport layer, the control and specification language as the input of the demands mapping
management layer and the cognitive layer. and policies generating unit; on the other hand, translating
(3) The MotoMeshTm test-bed network by Motorola the network reconfiguring policies from the demands
Corporation mapping and policies generating unit into the network
The main research area for Motorola Corporation in configuration parameters through the cognitive specification
cognitive networks is the MotoMeshTM test-bed network language, completing the work of network resource
with the ability of end-to-end reconfiguration. The project allocating and connection managing. The number of the
includes 27 top U.S. universities, equipment providers, network cognitive and reconstruction unit in network is large,
network operators, and standards and policy makers. Its goal corresponding to the cognitive and reconfiguring node.
is to design, develop and test a new network architecture, to The programmable communication network is consists of
guide the design of reconfigurable transmission equipment, the programmable configuration elements (links, nodes, etc.)
network equipment and supporting systems. and the network status sensors, and the programmable
(4) The Wireless Network After Next (WNaN) project by configuration elements are the control points of cognitive
DARPA network, as well the network status sensors completing the
2009, the Defense Advanced Research Projects Agency sensing work of the network status.
(DARPA) started the Wireless Network After Next (WNaN) 4.2 The Technologies of cognitive communication
project, based on the XG project and adaptive network networks
technology, aiming to develop the quickly cognitive-based Part of the network nodes and links in cognitive
network construction technology for the battlefield communication network possess the learning and reasoning
environment, lower cost of network equipment and handheld ability, and can dynamically make decision and reconfigure
wireless terminals, covering a wide range of connections for the network elements, by measuring or predicting the
each soldier. June 2009, the BBN Corporation in United network environmental parameters, to meet the targets of
States won the finance of 1100 million for the development self-adapting network environment, optimal use of network
of this network project. resources (Frequency, the link bandwidth, node processing
power, etc.) and optimizing the performance of end-to-end
information transferring. Its technologies include:
4. User demands
End-to end
goals
Demands mapping Service priority policy Delay-control policy
Transmission Bandwidth-control demands mapping and
Global network
protecting policy policy policies generating
information
Network environment Policies
Other policies
generating
Cognitive Specification Language-CPL
Topology info
Network resources configuration
and management
Network Link states info
sensing control network cognitive
and reconstruction
Node configuring Dynamic resource Security Connection
and states info allocation management management
Security states
Geographical and
meteorological info
programmable communication network
Figure-3 the architecture of cognitive network
(1) The rapid sensing technologies of communication mapping technologies
network environment The network resources dynamically allocating
The rapid measuring technologies of network technologies
environment parameters The cognition-based network security technologies
The rapid merging and processing technologies of the (4) The cognition-based link efficiently transferring
global environmental information technologies
The global network environment situation The link transferring technologies with the learning
(including network security situation) generating and reasoning capabilities
technologies The link reconfiguring technologies
(2) The User demands mapping and access control The cognitive radio based cross-layer routing
technologies optimizing technologies
The user demands mapping and forecast analyzing (5) The cognition-based network organizing and
technologies information exchanging technologies
The cognition-based user access controlling The network topology on-demand configuring
technologies technologies
(3) The policy-based network resource managing and The end-to-end connecting and QoS providing
controlling technologies technologies
The service level agreement (SLA) and network policy The network nodes and switching mode reconfiguring
5. technologies VI. CONCLUSION
V. THE IDEAS AND SUGGESTIONS ON THE DEVELOPMENT For the application environment and user demands
dynamically changed multi-domain complex networks, it is
OF COGNITIVE COMMUNICATION NETWORKS
needed to multi-angle sensing the network environment, user
To propose a new communication network architecture demands, network resources usage, service supporting,
and build a cognitive communication network with the security level, etc. Implementing a communication network
ability of cognition and reconfiguration is no easy task, with cognitive ability can improve the learning and self-
requiring long-term researching and investigating. The organizing capacity of the communication network,
current implementations and relevant technology on dynamically configure and restructure the network, and
cognitive network is in its progress, and has not been a broad better adapt to the complex and diverse changes of
consensus. However, the direction and focus of the application environments and user demands.
researches are already clear, and the key technology
researches have begun. These technologies (such as REFERENCE
communication network environment rapid detection, policy- [1] Teamwork and Collaboration in Cognitive Wireless Networks.Nolan,
based communication network resource management and K.E.; Doyle, L.E.;Wireless Communications, IEEE Volume 14, Issue
control, the cognition-based and efficient link transmission, 4, August 2007 Page(s):22–27.
the cognition-based networking and information exchanging [2] Utilization of Location Information in Cognitive Wireless
technology) reflect the important characteristics of cognitive Networks.Celebi, H.; Arslan, H.;Wireless Communications, IEEE
Volume 14, Issue 4, August 2007 Page(s):6–13.
communication networks, and are the important ideas for
[3] Cognitive Wireless Networks.Mahonen, P.; Zorzi, M.;Wireless
improvement and technological changes on the existing Communications, IEEE Volume 14, Issue 4, August 2007 Page(s):4-
communication network architecture. First, break through the 5.
key technologies in cognitive communication networks is [4] Guest Editorial -Adaptive, spectrum agile and cognitive wireless
needed. Second, realize the heterogeneous converged networks.Cordeiro, C.; Daneshrad, B; Evans, J; Mandayam,
networks with the local cognitive capability based on the N.;Marshall, P.;N, S.S.; Cimini, L.;Selected Areas in
existing network conditions. Finally, implement the wide- Communications, IEEE Journal on Volume 25, Issue 3, April 2007
Page(s):513–516.
area cognitive communications network.
[5] Towards Cognitive Telecommunication Networks. Safak, Mehmet;
Signal Processing and Communications Applications, 2007. SIU 2007.
IEEE 15th 11-13 June 2007 Page(s):1-4.