Specialist Dissertation Services
Dissertation Examples

Chapter Two

2.1 Introduction to CRNs 

   In the first place, fixed spectrum assignment policies of the wireless applications significantly lead to wasting spectrum resources which are valuable as a result, there is will be spectrum scarcity. Cognitive radio technology, introduced by Mitola, proposes an efficient way to exploit the unused available channels through wireless spectrum band with automatically detection way based on DSA. (Joseph, 2000).

Cognitive radio systems are classified depending on the criterion used to allow secondary users to use the licensed spectrum band  into three main paradigms: Underlay Paradigm: The secondary  user can transmit simultaneously with the primary user as long as the interference caused by the secondary user is below a specific threshold. Actually, it’s common using in the licensed spectrum (e.g. ultrawideband (UWB) communications) Furthermore, it’s also be used in unlicensed spectrum bands to provide different users services.

Overlay Paradigm: The secondary transmitter knows the channels moreover, the messages (and their codebooks) of the primary user. It can transmit simultaneously with the primary user as long and the interference is mitigated by some collaboration, for instance via relaying.  In licensed bands, secondary users would be allowed to use the band by sharing with the licensed users since the cognitive user would not interfere with, and might even improve, their communications by exploiting this knowledge in different ways to either cancel or mitigate the interference seen at the SU and PU. In a similar way, cognitive users in unlicensed bands improve using the channel with a higher spectral efficiency.

Interweave Paradigm: The secondary user work in an opportunistic transmission mode to access to spectrum holes or white spaces in licensed spectrum band, e.g., TV white spaces, to transmit its messages. (Andrea et al., 2009).

Hybrid schemes using a combination of the previous paradigms to improve the efficiency of spectrum sharing. In other words, allows secondary users to maximize their transmission rate. (Zhiqiang et al., 2007)

In figure (2.1) show three main paradigms in cognitive radio. (Lorenza ,2009)

Figure (2.1) cognitive radio interweaved, underlay, and cooperative overlay .

The cognitive radio node is classified based on cognitive functionality through the cognitive cycle. Where it is sense to  a set of environment parameters , it has the ability to the self-organized setting as well as adapts to reconfigurable itself in very flexible and transparent way.

Cognitive capability :it can be summarized as follows

• Spectrum sensing : This characteristic is described by the ability of the cognitive radio node to detect the spectrum holes in the frequency band, taking into account not to interfere with the primary user or limited interference as described in the previous description and then shared this acquisition case with neighboring nodes.
• Location identification: Its describe as the ability of the cognitive node to know its location as well as the location of the other transmitter nodes within the network. There are different ways and models of how to obtain the information of the location, including it is a manual way or it depends on another external way with another extended geographical location system.
• Network/system discovery: Naturally, so that the cognitive node can achieve the best way to communicate and share network characteristics it must first discover the networks available within the coverage area. Consequently, that network may have access to them via directed one hop communication or multi-hop nodes
• Service discovery : After discover and access to the network, it is possible to access the network services, so this feature is linked to network discover capability and then the possibility of finding the appropriate service suited to the node requirements.

Reconfigurable Capability :The ability of the cognitive node to reconfigure itself Consequently, the characteristic is described in the following points:

• Frequency agility : capability of the cognitive node to change the parameters of operating frequency in a dynamic way to select the best frequency depending on the signal sensitivity of the other transmitters. on the other hand, its dynamic frequency selection way.
• Adaptive modulation/coding (AMC) : This property has developed as an approach channel capacity in fading channels. Where the ability of the CR node to modify the characteristics of the transmitter to improve the access to the spectrum and its optimal use by choosing the best modulation type  for use to support interoperability between systems.
• Transmit power control (TPC) :This feature enables the device to dynamically change between many several transmission power levels in the data transmission process through  a more selective mechanism to lower effective power level.
• Dynamic system/network access : This feature is very important to work in Heterogeneous wireless networking environment thus CR reconfigure itself to  run different protocols that required to access the multiple Communication Systems/Networks.

Self-organized capability : Depending on the previous characteristics, the cognitive node has become very intelligent to communicate and compatible with the devices. The CR node is characterized by a set of characteristics at this stage, for example, Spectrum / radio resource management, Mobility and connection management and Trust / security management to provide maximum network performance. (Kwang et al., 2009)

Advantage of cognitive radio network

Compared with traditional radio, CR is more flexible, it’s a key enabling technology for increasing spectrum availability in wireless network.  In other words, CR has the ability to sense the surroundings in an RF sense and then adapt, in both a digital and RF sense, to that set of parameters. Depending on the application requirements such as voice or data, low or high priority, low or high security. the radio can select the most appropriate means of information transfer.

There are very large number of cognitive radio benefits in different application environments naturally, this is because of it being a multidisciplinary technology. We are include some of these benefits in the following below :

• Dynamic spectrum access
• Self-organizing networks
• Real-time spectrum efficiency
• Improving spectrum utilization
• Improving interoperability between emerging systems
• Improving link reliability, for instance, multipath poor environments to increase channel capacity
• Low-cost radios
• Mitigate Interference with Cognitive jamming systems
• Enhancing SDR techniques through using a set of intelligent algorithms called a cognitive engine (CE) to create a cognitive radio (CR) that provides a user with a required quality of service (QoS), low-cost, multi-band and multimode operation for interoperability with SDR technology.
• Extended battery life when it’s used in different application implemented with wireless sensor network and Internet of Things (IoT) application on the other hand its enhance energy efficiency or throughput in WSN.
• Extended coverage area particularly that one in rural network
• Efficiency Radio Resource Management
• Secure Radio Resource Management in Computing environments
• etc.

Challenges and Disadvantages of Cognitive Radio Network

• The hidden terminal problem in Cooperative Spectrum Sensing
• The complexity becomes a significant challenge to the Network lifetime
• Challenges in CRN routing
• Developing efficient information sharing algorithm and increased the complexity
• Challenges for machine learning research
• Detection sensitivity in a wideband spectrum
• Challenges CR in Wireless Sensor Network (CR-WSN) such as Hardware requirements (storage and energy) and Topology Changes that effected on affects the network lifetime in WSNs.in addition,  Fault Tolerance, Quality of Service (QoS), Security and Scalability in WSNs
• etc.

Cognitive Radio Networks Architecture

generally, the importance of the cognitive radio node to improving network performance especially channel capacity was also explained in the previous section .The nature of the cognitive network is heterogeneous networks compound of many communication systems. The heterogeneity exists in different location within communication system level for instance Wireless access technologies, networks, user terminals, applications, service providers and etc. the main goal of the design of the cognitive radio network architecture to improve network utilization. From the point of view of the user, the network utilization means that they can always to fulfill its requirements at anytime and anywhere through associate CRNs. while from the point of view of the services provider, they can not only provide better services to CR node but also accommodate efficient way to allocate network resources . The basic units of the cognitive radio network are classified into Mobile Station (MS), base station / access point (BSs / APs) and backbone/core networks. These are the basic parts of the three architectures :  Infrastructure, Ad-hoc and Mesh Architectures. (Kwang et al., 2009) .

Infrastructure Architecture :This architecture consists of the three parts of the network components For example, in fig()illustrates how  Mobile Station (MS) can only access a base station / access point (BSs / APs)  in the one-hop way as it show in the scenario(A). MSs under same coverage transmission range can communicate with each other through the BS/AP as it show in the scenario(B). while communications between different cells are routed through backbone/core networks

Infrastructure architecture of a CRN

Ad-hoc Architecture :This kind of architecture does not depend on infrastructure support but rather than its set up on the fly. in fig()illustrates when one MS node is targeted by another MS node within each other’s Radio Frequency (RF) range want to communicate and sending their data through communication standards/protocols for example (E.g., WiFi, Bluetooth) or dynamically using spectrum holes. it can directly establish communication links on demand.

Ad-hoc architecture of a CRN

Mesh Architecture: This architecture integrates both privious types infrastructure and infrastructure-less (ad-hoc) architectures. Where the node within this architecture is directly connected to BSs / Apsor using another node MSs as multi-hop relay This architecture allows more flexibility to create different types of structuresSome BSs/APs may connect to the wired backbone/core networks and function as gateways to another network.In addition it has ability to be aware of the spectrum hole to access to the channel and communicate with each other.Thus allowing the highest performance So the capacity of wireless Communication links between cognitive radio BSs / APs is large .  on the other hand, it allow the wirelessbackbone with optimal channel exploitation to serve more traffic

Mesh architecture of a CRN

 Common Application areas of cognitive radio network

The cognitive radio network has many applications in various fields, for instance, it can be used in places where the wireless  sensor network is used where CR developed as general technology within the applications of this network CR-WSNs, for example Facility management, precision agriculture,  machine surveillance, defensive maintenance, Medicine, telemetries, logistics,  object tracking, intelligent roadside, safety, actuation and maintenance of complex systems, monitoring of the indoor and outdoor environment. . (Gyanendra et al, 2015)

Leased Networks : Most CRN applications have the secondary users exploiting the resources during non-existence of the primary networks in general case without being beneficial to the primary networks in any way. However, a primary network can benefit from leasing a portion of its licensed spectrum to the secondary user to operate adaptively to opportunistically access the spectrum. Thus improving the general performance of the cognitive network by the very important property to increase co-operation between nodes through spectrum sharing capability will be significantly mitigated by connection failure and channel capacity. (Fanzi et al, 2016).in figure 4 we illustrate concept of leased network .

………

Primary Users

Secondary Users

Leased Network

Emergency Network: CRNs can be used for such emergency networks for instance ambulance, fire, police, and rescue. provide a significant amount of bandwidth especially natural disasters that can cause collapse and crash to the communications infrastructure so it became necessary to work in CR environment to handle the expected huge amount of sensitive real-time multimedia.(Qiwei et al, 2006). In figure 4 we illustrate how Cognitive radio Rapid Deployment and Interoperability between network units to make connection to all communication units after destroyed communication in Disaster Relief environments. (Saim et al., 2014)

Emergency Network scenario

 

Military Applications : Naturally, achieving reliable and secure communications in modern battlefields has become a more challenging task nowadays and the focus of researchers by means of the security considered the main factor in the first place and then bandwidth. ideal military CRN network would have a distributed, that enable CR military nodes to transmit on any frequency band while simultaneously managing the network without a central unit. (Qusay, 2007). In figure 3 we illustrate CR with spectrum awareness, collaborative sensing, fast response time, awareness /collaboration with other network resources. In addition, the cognitive radio network is to help gather a variety of different wireless links.(Elektrobit, 2013)

Intelligent actical Military network.

Wireless Network Powered by Distributed Green Generators :One of the most important applications is to reduce wireless network constraints like power and radio spectrum scarcity parameters.  Throughutilizing the cognitive radio   network and its characteristics to improve radio spectrum for data communication.Furthermore, take advantage of energy and electromagnetic spectrum efficiently based on traditional energy resources like wind and solar .etc.Therefore, adopting the Energy Harvesting (EH) through green energy can be exploited to strengthen wireless power networks. In figure 2 we illstrate how power farm (Wind Turbine, Solar Pannel) enhance power flow through power sharing with neighboring cells between network components(Xueqing et al, 2015).

Green energy powered cognitive radio network.

2.2 Security of CRNs benefits

The CRNs have provided a very large motivation for the development of various applications, especially those that need to ensure security in a very reliable way for their effective impact by making very effective decisions that need accuracy, safety and real-time access as much as possible for instance military applications from a point of view network performance , the cognitive radio network has greatly improved the performance of the network through the following points highlighted:

• Speed of adaptation
• Interoperability
• Usage of network resources
• Security

CR self-adaptive networks are able to Quickly respond to any changes that occur at the level of services and performance requirements and any other changes that occur during the development stage to take effect very quickly on the entire network, for example, may include updates or changes to the topology of the network , resources allocation even at the level of security modification. it reduces the delay of changes which is generated through manual network planning and configuration In addition, its speed of cooperation and compatibility of the tactical networks . (Anssi, 2015)

One of the most important requirements for the creation of an efficient network is to work in an environment that allows the highest level of interoperability between network components, interfaces, and protocols. CR, software-programmable network devices enable full adaptation of network protocols and parameters. allowing different types of systems to communicate with each other, improving the process of compatibility between Nodes, obtaining the highest level of network coverage area, furthermore, it allows the flexibility of communication and transfer of the information between the elements of the network and thus obtaining the highest quality of information flow. (Fette, 2015).

Within the concept of wireless network resources management means the optimal use of electromagnetic radio spectrum as a key factor of this network is spectrum usage. As previously described the purpose for developing CR to provide a maximum spectrum bandwidth capacity, security control, and other resources enhancement. In the future, dynamic spectrum usage will be a mandatory capability as the number of wireless devices keeps growing to meet the requirements of technology.(Fette, 2015).

Security is the most interesting factor in this thesis CR system has the ability to adapt to security parameters and techniques according to goals set by the system administrator through from the network security policies during initial stage then leaving the system to operate in an automatically manner to the required security level, which minimizes security vulnerabilities caused by human errors and omissions. Within this automatically way CR managing security parameters, building situational awareness and protecting mission-critical networking to provide the best acceptable protection against current threats. (Anssi, 2015)

2.3 The general concepts for the Security of CRNs

CRNs has attracted the researchers interest in a various fields related to the development of wireless networks application especially, those related to spectrum usage, but it remained a hot subject for researchers, especially related to security concept that need more attention. A CRNs security issue has the same scenario as other networks. It is divided into two lines of protection. The first line focuses on preventing attacks using cryptographic primitives. While the second line to detect and identify the attacker on the other word, detect the malicious behavior, eliminate or mitigate it and then attempt to protect the network from future attacks.(Bace et al, 2000). There are many trends more biased to the second line of protection, of course there are different CRNs architectures either infrastructure-based or infrastructure-less (decentralized) certainly in the case of the distributed architectures for instance, ad hoc networks, attack detection process is very difficult due to the difficult for the control on the node movement its add another complexity which is responsible for the control over these nodes so it made this structure more vulnerable to attacks .

Attacker Model : System design steps should be considered concept of the attack model that determines the attacker’s capabilities and misbehavior process that effected for CRNs services precisely, there are a different types of the security issues which are included within attack model

Detection methodology : Detection models collect the information from different sources. For instance, this information is generated within the same system depending on the characteristics and techniques of the work of that system or generated from external sources, for example, from other adjacent models, in the final analysis for the process of building an knowledge system of previous information and new one can help the system to distinguish between normal and abnormal behavior of nodes within the network and so on to identify malicious behavior  (Stefan, 2000).

There are three main techniques of intrusion detection: misuse detection anomaly detection, and specification-based detection.

• Misuse detection : In this technique, gathered information are compare  with the predefined “signatures” of well-known attacks. This type is very interactive but only for the specific types of attacks and certainly, predefined one but does not have the ability to detect new attacks.
• Anomaly detection: In this type a particular pattern is stored, which is considered as a normal behavior state. as well as, this pattern is constructed by the concept of Machine learning techniques and training. as a matter of fact, the system response against any change happen to this pattern.
• Specification-based detection: This type also depends on the deviation from a normal behavior but in this type the normal behavior is based on manually defined specifications.

The detection system of attacks for the cognitive radio network should integrate these three techniques to be more compatible with different CR characteristic, e.g. such signature-based scheme to detect Lion Attacks or an anomaly-based technique to detect OF attacks.

General Requirements : When designing the second line of defense (identifying and detecting attacks) it is very necessary that certain requirements Which are fit the characteristics of the cognitive radio network elements, as follows(Mishra et al., 2004):

• Of course the system design should not produce a new weakness for the system, for instance, the cooperative detection module it must take into account the presence of malignant and defective nodes as well as the presence of DoS attack.
• The system must be able to reconfigurable itself in the other word, the fault-tolerant system to avoid errors or recover from a problematic case, based on prior knowledge stored in the current and previous states to solve the system fault problem.
• The system must provide the users or the network with appropriate mechanism that allow them to know the current attack on the network and take reaction against it and this includes the attack against the system itself In other words, the ability to detect vulnerabilities of the system.
• System design should have the ability not only adding a new detection modules, but also a seamless interaction with existing detection mechanism.

 Security Models and Requirements for CRNs

Security model is a protection scheme has a different security services to protect network information against malicious nodes threats. Its aim to specify and enforce security policies, access rights or any encryption algorithm in the network. There are several cryptography methods and modulation techniques on different locations within the security of communications system use to protect entirely network transmission. In this paper we have make a survey on security models used in CRNs. In (James et al, 2011) the author talked about use of the different security ways and their effectiveness in cognitive security for example (Symmetric-key Algorithm: RC5 [block], Asymmetric Key Algorithms: Elliptic Curve Cryptography [ECC] ), aimed to solve existing and new security threats in a heterogeneous communication network. In (Deepraj et al, 2012)(Katharine et al, 2012) security module are (Symmetric-key Algorithm: AES (block cipher), RC4 (stream cipher), Asymmetric-key Algorithm, Hash Functions ), aimed to identify a new potential threats. Most of the attacks that make use of one of the inherent properties of cognitive radio and evaluate their impact on CRNs performance within their applications in mobile and sensor networks. In (Anssi et al, 2013)focus on security of the main critical CRNs applications like battlefield applications by generate a cyber-security architecture, through taking into consideration the finding weaknesses and decrease security vulnerabilities by enhance security of tactical military networks. In (Kresimir et al, 2015) discuss the security and protection for related topics with cognitive radio such as wireless systems, Software Defined Radio. They are different security models in this topic(Symmetric-key algorithm, A5 (stream cipher), KASUMI (block cipher), Wired Equivalent Privacy (WEP) : Rivest Cipher 4 (stream cipher), Wi-Fi Protected Access (WPA) based Temporal Key Integrity Protocol (TKIP), WPA2 based Advanced Encryption Standard (AES), Spread spectrum techniques like direct-sequence spread spectrum (DSSS) ), A symmetric-key algorithm as (digital sing) (Blesa et al, 2015) to detect PUE attack, besides he studies the impacts of cognitive radio technology on tactical battlefield solutions ,also  study the principles and practical solution related to Radio Frequency (RF) jamming and anti-jamming problems, proposes a game-theoretical framework (Kresimir et al, 2015) . In (Bhagavathy et al, 2014) discussed the possibility of secure CR by using the combinations of encryption algorithms (Symmetric-key Algorithms, Asymmetric Key Algorithms) such as (Rivest-Shamir-Adleman, Elliptic, Secure Hash Algorithm , Digital Signature, etc.) and spread spectrum modulation for mitigate different type of layered attack, also Kerberos algorithm for reliable session establishment by limited shared keys through prevent attacker from creating the session key to access for highly secure communication.in (Wendong et al, 2007) dynamic frequency hopping (DFH) is proposed for WRAN data transmission, meanwhile reliable spectrum sensing and efficient channel usage in parallel. In(Mahmoud et al, 2014)spread spectrum strategies are procedures: DSSS (Direct Succession Spread Range) and FHSS (Frequency Hopping Spread Spectrum) in CRNs under the parameters of Data Drop Rate, Detection Time and Throughput , also as a supporting methods for security model deals with interference in cognitive radio network. In (Mohandass et al, 2014) different types of multi carrier modulation schemes and spread spectrum techniques proposed like Direct Sequence Code Division Multiple Access (DS- CDMA), Multicarrier Code Division Multiple Access (MC-CDMA). They are use various types of enhancement models to mitigate from Inter Symbol Interference (ISI), multiuser interference (MUI) and robustness against narrowband interference (NBI).

See in figure one we show security scheme block diagram for a group of thesis and papers deals with security of cognitive radio network.

Fig 1: CRNs Security Scheme Block Diagram

 

Security Requirements for CRNs: Cognitive radio networks such as any type of wireless network have security issues. Moreover, the open-air medium (wireless) is a wide vulnerable to attack. Cognitive radio networks have some special characteristics for example (high sensitivity to weak PU signals, scarcity of common control channel, missing PU receiver location, etc.). So attacker are trying to attack the weaknesses of these characteristics and others on various layers, protocols and associated technologies .Security measurement and various evaluation policies should be applied to reduce the probability of attack by malicious nodes in a wireless network(Mahmoud et al, 2014),   however security requirements in CR wireless network nodes are authentication  , availability, confidentiality, integrity, authorization and non-repudiation(Jacobs et al, 2011).

 

Layered and Cross-Layers attacks Against CRNS

The risks in CRNs disaggregated by the target TCP/IP five layers begin with: physical layer attacks and ends with application layer attacks in addition to, Cross-Layers attack in case effected on any layer can arrive to another layer .

A. Physical Layer attacks : The basic platform layer of TCP/IP model it’s the physical medium of the channel that’s establish connection between two or more devices with each other for example the network cards, cables, or the atmosphere as wireless networks. Cognitive radio network is differ from traditional wireless networks because of the cognitive radio uses dynamically Opportunistic Spectrum Access while traditional wireless networks use fixed frequency band. Spectrum sensing to access unallocated spectrum bands, and open air medium as physical layer channel in CRNs are vulnerable to many security issues that’s make the attacker exploit them during spectrum sensing process (Mahmoud et al, 2014).

These are a group of the most common attacks on CRNs channel and prevent access to physical layer of the cognitive radio network (Primary User Emulation (PUE) Attack, Objective Function Attack (OFA), Jamming Attack, Eavesdropping, Primary Users’ Location Attack and Learning Attack (LA) )(Yi-cheng et al, 2014).

Primary User Emulation (PUE) attack : Malicious node hide themselves and change behavior to be similar the PU behavior by transmitting special signals same as using by legal primary user in the licensed band, and thus lead to wrong sensing process by the secondary user and make he believes in the presence of primary user (Wassim et al, 2011).

objective Function Attack (OFA) : The task of cognitive radio is the adaptive according to certain parameters in the environment, Radio parameters include center frequency, bandwidth, power, modulation type, coding rate, channel access protocol, encryption type, and frame size (Deanna et al, 2011). cognitive engine work depends on these parameters, the best choice according to the requirements needed by the application and used to solving one or more objective functions such as reduce power and maximum throughput, etc. Attacker tries to exploit the weaknesses of the techniques for evaluating these parameters, manipulate them and thus lead to undesirable results to calculate the value of objective function according to the requirements of user applications to the network(Olga et al, 2010).

Jamming attack: An attacker sends a packet during a connection to exploit and reduce the signal , it’s considered a familiar for denial of service ,However, There are other cases caused a bottleneck or congestion as a result of messages exchange between the nodes will be affected on the signal quality .

Eavesdropping Attack:Malicious node attempt to listening on communication between the various legitimate reliable network devices even base station ,to get a useful information and launch further attacks on the basis of these information as start point for more harmful attack on the network (Yi-cheng et al, 2014).

Primary Users’ Location attack : Considered one of the most dangerous attack types because it directly attacks on the devices after identify primary user location, in CRN malicious node calculate distance between itself and primary user by convert signal strength to distance . When more than a malicious node estimated location of the primary user based on this method, they can creation of confluence or crossroads then get the original location and doing physically attack on primary user (Yi-cheng et al, 2014).

Learning Attack (LA) : in this type of attack Attacker announces the wrong messages caused by the wrong process sensing and this false information will be remaining during all next step after spectrum sensing process in cognitive radio network (Nathan et al, 2008).

B.  Data Link Layer attacks : The data link layer is hides the details of underlying hardware (physical medium), attacker in this layer exploit vulnerabilities in MAC address Strategies (Mahmoud et al, 2014).

Spectrum Sensing Data Falsification  (Byzantine attack) : Malicious node will send falsified spectrum Sensing information for misleading and manipulate the decision-making process,  prevent secondary users from using the existing spectrum hole, bring them spectrum band to access the channels and cause excessive interference or reduce throughput to legitimate users in CRNs (Deanna et al, 2011).

Control Channel Saturation DoS Attack(CCSD) : Due to the distribution communication, process is necessary to connect nodes with each other and multiple overlapping manner in a multi-hop CRN, within data link layer MAC control frames are exchanged to preserve the channel. such a complex and multi-state connection, When many CRs want to communicate at the same time, channel can only support a particular number of synchronous data channels A result of this common control channel becomes suffering from a bottleneck. attacker exploits this status (distributed not centralized CRN) by create a fake MAC control frames for the purpose of saturating the control channel and thus decreasing the network performance to deal with this huge number of MAC control frames, the Control Channel Saturation DoS Attack leaves the CRN with a near-zero throughput (Kaigui et al, 2006) (Li Zhu et al, 2008).

Selfish Channel Negotiation (SCN) : The most common applications of cognitive radio based on a multi-hop routing, in this type of attack cognitive radio node refuse to send the forward message to the next node in order to preserve its own throughput which resulted from selfish channel concealment (Kaigui et al, 2006).

2.  Network Layer attacks:  Cognitive radio has network topology like  classic wireless communication networks for example infrastructure-less (ad hoc, mesh), infrastructure-based (mesh) and hybrid (Wireless Sensor Network). Security threats in cognitive radio network targeted routing function by sending the wrong paths to building routing table and make collision leads to drop packets in network layer, anyway there are three main attacks types in this layer are Hello, sinkhole, and Sybil attacks (Mahmoud et al, 2014).

Hello attack:Because of the similarity and convergence in routing strategies used in wireless sensor networks, this type of attack also adversely effect on CRNs. Attacker use a high energy level in order to broadcast message reaches all network nodes with a good signal strength and make nodes illusion that the owner of this message is a neighbor.as soon as these node forward packets to the attacker and find themselves without neighbors (Chris et al, 2003).

Sinkhole Attack: Cognitive radio networks often use multi-hop routing. sinkhole attack aimed to exploit this benefit and advertising itself as the best route to a specific destination node, neighbor nodes helps to spread this property in the network .and thus increase the impact strength and conviction to choose this path .The attacker can commits the selective redirection attack such as forwarding, dropping, modifying, or eavesdropping any packets pass through it (Stephen et al, 2014).        

Sybil attack: Each CRN nodes have a legitimate identity to communicate and use the channel. Sybil attack exploits this feature by create a large number of false identities that effects on results of spectrum decision making process, prevent use of the channel by legitimate users and opportunist overall control channel transmission (Muheet  et al, 2013).

Ripple effect attack: A special type of attacks to cognitive radio because of CR ability to change the spectrum bands during the use of the channel. The ripple effect is similar to the primary user emulation or byzantine attack in that the false sensing information provided especially during any updating for new network topology changes.

Malicious node for this attack affected on the related information for choice spectrum bands in the channel then effect on throughput of the channel in the network .wrong information will be pass hop by hop and make the network in disruption state (Jing et al, 2012).

Transport Layer attacks: Attacker in Cognitive Radio Network (CRNs) trying to exploit vulnerabilities during transmission session (establish connection process) between nodes. They are different attacks type in this layer for example (Key Depletion attack, Jellyfish attack) (Mahmoud et al, 2014).

Key Depletion attack: Operation of generating encryption keys in transport layer occur by private protocols such as The transport layer security (TLS) and secure sockets layer (SSL). There are a large number of sessions be established for the purposes of communication initial between CRNs nodes, increase opportunity in repeating the key with increasing establish the connection process. On the other hand, There are some protocols that have been proven to contain the weaknesses in this area are the temporal key integrity protocol (TKIP) and the wired equivalent privacy (WEP) which are implemented in IEEE 802.11.Key depletion attacks are aimed to exploit key session repeat through these weaknesses or doing to increase the operations of established connection until it reaches to this point or stops the system (Stephen et al, 2014). (Deanna et al, 2011).

Jellyfish attack: Jellyfish attack is effect on behavior of Transmission Control Protocol (TCP) in transport layer , moreover it performed on the network layer before next step transmission in transport layer (Chetan et al, 2007).

Lion attack: Lion attack has the ability to harm the transport layer for example reduce throughput. Specifically, targets the TCP connection by exploiting vulnerability (Wenkai  et al, 2010).

3.  Application Layer Attacks : As a result, any attack on physical, data link, network or transport layers may have an adverse effect on the application layer in Cognitive Radio Networks (CRNs).
4.    CROSS-LAYERS ATTACKS : Cognitive radio network requires application-aware communication protocols, protocol stacks that have large memory footprints are not desirable. Researchers directions to merge convergent layers as cross-layer design includes the functionalities of two or more layers in a single coherent framework to reduce handoff (handover) latency, fitting the requirements of many live streaming, interactive and reduce unnecessary complexity. There is a need to be given individual attention for such attacks (Mahmoud et al, 2014).

See in Figure 2 we present interconnection between layers as cross-layer framework in the context of  Cognitive Radio Sensor Networks (CRSNs) (Suleiman et al, 2013).

Fig 2: Cross-Layer Framework Communications.

Lion attack: A lion attack is outcome of a Primary User Emulation (PUE) or Spectrum Sensing Data Falsification (SSDF) attacks, they have affect at lower layer (PHY or MAC), lion attack has the ability to harm the transport layer for example reduce throughput. Specifically, targets the TCP connection by exploiting vulnerability. This type of attack causes most effective and damage to the network when moving to another new channel causing the same damage and thus prevents secondary user from transmission new data (Wenkai  et al, 2010).

Lure attack:  Here the harmful node modify packets sent from source to destination during routing process by add false information leading to lure other nodes into the routing lap and deletion of legitimate forwarded packets ,the final result for effect of this attack on the network leads to less efficient of network performance (Juan et al, 2011).

Note from TABLE 2 we summarize Layered attacks and cross-layers attacks against CRNs and their common countermeasure using to mitigate from the negative impact on the cognitive radio networks.

Table (2.1)Layers and Cross-Layers attack with their common Countermeasure.

Attack	Targeted layer	Common Countermeasure
PUE Attack	Physical Layer	Cryptographic Authentication
OFA Attack		Prior identification the threshold value(Olga et al, 2010),  Intrusion Detection System (IDS)
Jamming Attack		Determine and keep track the location identification  of the primary user’s, Frequency hopping spread spectrum Technique
Eavesdropping Attack		Encryption Techniques
Primary Users’ Location Attack		changing the density of signals irregularly
LA Attack		Control Environment especially during learning phase, Constant Reevaluation (Nathan et al, 2008).
Byzantine Attack	Data Link Layer	Misbehavior Detection System (MDS) (Kefeng et al, 2013), Cooperative neighboring cognitive radio nodes (COOPON) (Minho et al, 2013), trust and reputation metrics
CCSD Attack		Trust as Detection Mechanism
SCN Attack		Trust as Detection Mechanism
Hello Attack	Network Layer	Symmetric Key based algorithm
Sinkhole Attack		Geographic routing protocols (Chris et al, 2003)
Sybil Attack		identity validation
Ripple effect Attack		checked and validated necessary information
Key Depletion Attack	Transport Layer	Security Protocols
Jelly fish Attack		not complicated algorithm like direct trust-based detection (DTD) (Vijay et al, 2014)
Lion attack	Cross-Layers	Single Layer Monitoring & Trust Calculation (SLMTC) ,Trust Fusion, Abnormal Detection (Wenkai  et al, 2010).
Lure Attack

Security challenges of Cognitive Radio Network

As a matter of fact, main CR security challenges through  proposed a novel approach are

• Power consumption
• Less complexity
• Reliable spectrum sensing
• Reliability module CR communications
• Optimal allocation of radio resources

(Yanwei, 2014) (Anssi ,2015) (Kresimir ,2015) (Javier ,2015).

Still open problem

Still open problems are design protocols that enable detection malicious activity while minimizing the message exchanges and delay which are the result of protocol running in distributed model, reliable spectrum sensing by control on malicious behavior, enhancement security and privacy to eliminate risks that related to the patient health within CWSNs healthcare, design security module take into consideration the power saving model and less as possible power consumption in cognitive wireless sensor networks environment. (Yanwei, 2014) (Anssi, 2015) (Kresimir, 2015) (Javier ,2015)

2.3 Future Work

CR is the next-generation wireless communication system, It has become necessary to focus on building an integrated security system and very strong one at the same time adapted to the quality of information exchanged between the cognitive radio network devices . in future direction we propose intelligent context meaning for any information exchange between CRNs inside military application information, through checking semantic meaning for incoming data and search about keyword to determine where this data need minimize or maximize security scheme, regularly decide which security scheme will be choice as the best for this status depend on modulation techniques such as (FHSS, DSSS, DFH, CDMA, Hybrid, etc. ) and cryptography techniques (vigenere cipher, AES ,RSA, etc. ) so system strength , generally is directly proportional to the sensitivity of the information .

2.4 Simulation tools used to enhance security of Cognitive Radio Network

They are different types of simulation tools and programming language to simulate the concept of security of cognitive radio network anyway, common simulation tools explained briefly in the following bellow :

MATLAB :  is the easiest and most productive software to simulate different type of network security architecture for instance in (Yanwei, 2014) the author proposed game theoretic approach and evaluate the performance for security purpose and energy consumption in mobile ad-hoc networks. While in (Kresimir et al., 2013) Their results were simulated, With an efficient working to estimate security issues in Cognitive Radio Networks (attacks and its counter-measure).

NS-2 : One of the most important open source simulators, which is used for a variety of  protocols and applications framework for security threats in Cognitive Radio. It is a key tool for working within this environment in research community (Elena et al., 2012). Energy management, collaborative, scalability and learning are not mentioned or implemented in cognitive strategies for security in wireless sensor networks (Javier et al., 2015).

NS-3 : it’s also open source simulator as a great step developing in order to enhance the features of NS-2 . There are many simulation framework  that is suitable for large networks. especially that one deals with several CR capabilities, such as primary user detection, spectrum sensing, spectrum acquisition moreover take into consideration execution time and memory usage (Abdulla et al., 2014)

OMNeT++:It’s another open source simulator can execute the code under a powerful graphical user interface it makes the internal events of a simulation model fully visible to the end user: it displays the network graphics, animates the message flow and lets the user peek into objects and variables within the model in the simulation running state, it provides a modular library, in addition, the tracing/debugging capabilities does not require extra code to be written by the simulation programmer. It is used within the cognitive radio network to apply a new cognitive model over an existing WSN by adding Castalia framework simulator to simulate CWSN devices with different radio standard interface. Furthermore it also used to develop CR architecture (Javier, 2015).CR routing strategy also simulated using omnet++. (Yang, et al., 2009)

OPNET: Its  provide several advantages as it provides a GUI to develop, deploy, debug and modify different algorithms and parameters for different network topologies design for instance, it has a performance data collection and the display module. in addition, allow data filtering and intrusion detection strategies to mitigate intrusion and vulnerabilities in Cognitive Radio Networks. (Ohaeri, et al., 2015). However, OPNET has a level of challenge and difficult to accurately implement the PHY layer in Hybrid cognitive validation platform for Wireless Body Area Networks (WBANs). (Sabin, et al., 2015) .

Qualnet: Its A simulation tool for developing and managing network systems such as security and privacy protocols for Cognitive Wireless Sensor Networks(CWSN) through evaluating the impact of jamming attacks in different radio settings . Moreover, examined the effect jamming attack on the control protocols (UDP, TCP) on channel switching delay and the packet size. (Jaydip, 2013)

MONTE CARLO :It’s deals with series of dependent variables with probability distribution approach . in cognitive radio network used in different module such as Cooperative Spectrum Sensing in Cognitive Radio Networks .(Marco et al., 2009). as well as another work proposed Optimized Sensing and enhances the performance of detection probability for given signal to noise ratio (SNR) in Cognitive Radio Network(Abhinav et al., 2017)

Simulation Tool	Official Website	Supported programming language
MATLAB	https://www.mathworks.com/	C, C++, C#, Java, Fortran and Python
NS 2	https://www.isi.edu/nsnam/ns/	OTcl, C++
NS 3	https://www.nsnam.org/	Python, C++
OMNeT++	https://omnetpp.org/	C++, Java, C#, NED
OPNET	www.opnet.com	C, C++
Qualnet	http://www.qualnet.ca/	Parsec, C++
MONTE CARLO	http://www.goldsim.com/Home/	JavaMonte