Secure Self Re-Organizing of Nodes Using Closeness Technique in Cluster MANET

Text Sample:

Chapter 3.2 SECURITY ISSUES IN MANET:

Wireless networks provide connection flexibility between users in different places. In addition, the network can be extended to any place or building without the need for a wired connection. Wireless networks are differentiated into two classification namely infrastructure networks and ad hoc networks. An Access Point (AP) acts as an essential coordinator for all nodes in infrastructure networks. The nodes are linked to the network with the support of access point. In addition, AP systematizes the connection between the Basic Set Services (BSSs), facilitating easy access of route at the time of requirement. But the limitation of handling infrastructure network is the huge volume of the routing table. Additionally, the AP is connected to the router through wire in infrastructure environment leading to unsecure communication. [.].
Figure 3.1 describes the types of wireless network communication. The dotted line represents wireless connection and the solid lines represent wired connection. The two different types of wireless communications are wired AP connection with wireless nodes and complete wireless node connection. The purpose of AP is to coordinate the nodes for better communication and to support transmission as AP is connected to router/hub. On the other hand, the complete wireless node connections are connected to mobile devices. Moreover, the possibility of attack is high on communication between mobile device and nodes. Additional concentration is required in complete wireless node connections.
Unlike in infrastructure network a central coordination point is needless in ad hoc network. Moreover, sending and receiving packets are denser than infrastructure networks. Currently, with the enormous development in wireless network applications like PDAs and cell phones, handheld computers, researchers are optimistic to progress the network services and routine. One of the demanding proposal issues in wireless ad hoc network is supporting secure mobility in MANET. The independent behavior of nodes in MANET enhances the difficulty of the routing protocols and the amount of connection flexibility. But, the possibility of permitting nodes to connect, depart, and transfer data to the network cause security challenges.
In a wireless network, a transmission range of node usually uncovers the entire network, so end-to-end communication requires routing information through some nodes. As a result, ad hoc networks are referred to as multi-hop networks. The hop is a direct link between two nodes. In an ad hoc network, nodes are termed as routers or terminals. As ad hoc network is an environment without infrastructure, the co-operation among the routers is worst. Independent nature of nodes creates the problem of routing framework leading to nodes misbehavior in the network. The misbehaved node types are referred to as selfish nodes. The lifespan of the network automatically decrease when the selfish nodes in the network increases. The main solution to address these problems is providing secure routing.
Fadlullah et al (2010) proposed techniques for securing routing in ad hoc network such as Certificate-Based Cryptography (CBC) and ID-Based Cryptography (IBC). For a fixed network, a secure communication is effectively done through Public Key Infrastructure (PKI) or a centralized certification authority. But these methods are not feasible for ad hoc network. Distribution of a signing key and Centralized Authority (CA) functionality over multiple nodes is a promising solution to this problem. In ad hoc network, the reputation mechanism is presented to dynamically evaluate the authentication of neighboring nodes. The reputation system selects the trusty nodes to support trustworthy behavior in transferring packets.
The three main goals of reputation system are providing useful information about the authenticated and unauthenticated nodes in the network, encouraging the authenticated node to involve in the communication facilitating a secure channel and discouraging the unauthenticated node. After eliminating unauthenticated node, the authenticated nodes are involved in the network communication. The ranking model is also used in ad hoc network for a secure communication. The process involved in ranking model is to rank the node based on their activities and forwarding time. Based on rank, the ranking model selects the nodes to transfer the packet data.
3.2.1 Reputation Based Self Re-Organized Node Clustering:

The reputation of the node defined based on the quality of the node behaviors. The reputation systems helps the clustered nodes to collect, distribute and aggregate the information about all the nodes in the cluster based on its past behavior. Each node maintains a reputation table which contains information regarding the node's current transmission. The information maintained in reputation table is collected either directly or indirectly from each node in a MANET. The direct way to collect the reputation information of node in the network is through observation of node characteristics and the indirect way to collect the reputation information of node in the network is through observation of related other nodes characteristics with respect to specific node.
In addition to utilizing past observations in gathering reputation information, SKM presents a new reputation system. The working of reputation system in SKM facilitates the process of identifying the unauthenticated nodes. SKM uses reputation detection to ensure that old reputations fade away providing more possibility for nodes to retrieve their reputation. The authenticated node is decided based on constant behavior of the nodes in a cooperative manner.
A secondary response supports to react against any neighbor nodes with originally bad unauthenticated reputation. The wild unauthenticated nodes show early signs of node misbehavior afterwards to avoid selfish nodes from network. The response of misbehaved nodes is recorded in the reputation table in form of reputation value of nodes. The reputation value of misbehaved nodes in the network is indicated as negative. The reputation noise detection in avoiding misbehaved node and cancellation, variation test and secondary response implemented in SKM in order to increase the accuracy and reliability of the reputation system.
The reputation system in the SKM is performed with the identification of malicious nodes based on high reputation and high symmetric. The level symmetric is considered in the task of performing high symmetric rate. In SKM, Level symmetric is used in order to choose the good number of significant authenticated nodes to support in the role of selecting other nodes in the cluster. The selected other nodes based on level symmetric builds trust into other less popular nodes in the network and act as public leaders. Nodes with higher symmetry contain higher chance of getting in contact with many authenticated nodes than nodes with low symmetric.
The source of indirect reputation information is the recognition of both high symmetric and high reputation. The identification of reputation in the nodes grows even more significant in high mobility MANET network communication. As nodes are independent to each other, with few links, there is much possibility of frequent connection change causing more asymmetric nature. Argument of secure key model is nodes with higher symmetric and higher reputation are major nodes to give highly trusted authentication opinions about other nodes in MANET in a self reorganized manner.
A symmetric of personality networks is used for each node to attain localized view of its adjacent nodes to permit fast reputation union and consequently higher throughput. Node clustering is performed with the reputation system and level symmetric. Node clustering based on reputation symmetric is explained below in Figure 3.2. [.].
Figure 3.2 is an example for clustered node, where N1, N2, N3, N4 are the neighbor nodes to A, B and C. In the cluster, the nodes A, B, C cannot communicate with each other without the help of its neighbor nodes. Node A possibly sends message from nodes N1 to N4 either directly or indirectly through N1-B-N2-A-N4. In indirect communication, initially node B is responsible for transferring message from N1 to N4 with the help of reputation observation of node A. As node B starts the communication, it is more important for successful message delivery. Therefore, level reputation-based symmetric approach is performed on node B. On performing level reputation-based symmetric approach, node B collects symmetric measure based on Node A's adjacent nodes reputation estimation of that node A. Indicating that node B is susceptible against attacks, as it knows the characteristics of other nodes around it through clustering. The information around node B protects it from the attacks, restricting the vulnerable ones. Hence, the implementation of level reputation-based symmetric approach makes Node B more robust against attacks.
Nodes reputation information is gathered with the aid of symmetric node characteristics. Node reputation information is classified into high, medium, uncertain, low and negative. Similarly, symmetric characteristic of nodes is categorized into high, medium and low. [.].
Table 3.1 shows classification of experimental nodes into sectors based on their reputation and symmetric. On monitoring the state of the node regarding the characteristics like path travelled, status of node behavior and trust, the nodes are sectored to high, medium and low. As per in Table 3.1, high reputation and high symmetric in sector 1 denotes high trust. Node N4 falling into Sector 1 i.e., high reputation and high symmetric which denotes that this node holds the high trust and clear path travelled characteristics. Hence, node N4 is highly trusted with wider view of the network. Moreover, the node N4 under the high trusted sector is highly advantaged. The nodes classified as belonging to high trusted sector is exceptional from the variation test and decision-making termination time. The variation test finds the unauthenticated node and authenticated node based on low and high reputation respectively. The high reputation or authenticate node reduces termination time records.
On the other hand, nodes falling into Sector 6 are classified as misbehaving nodes due to low reputation and low symmetric which indicate the trustless characteristics. Therefore, a node N2 falls into low reputation and low symmetric which is termed as unauthenticated node as well as misbehaving node. Therefore, the presence of node with negative reputation value like node N2 is eliminated from network communication to avoid attacks. Similarly, the identified misbehavior nodes are rejected from involving in network communication.
Based on the node behavior characteristics, the nodes are categorized in reputation table as depicted in Table 3.1. Node N4 is recognized as authenticated node in Sector 1 for transferring packets by establishing trusted connection. At the same time nodes N2 is identified as unauthenticated node because the node lies in Sector 6 with untrusting connection. So, the node A rejects the connection to node N2 and searches trusted node like N4 for packet forwarding. The packet is forwarded from Node A to Node B on estimating the shortest path. So, the packet forwarding establishes connection through node N1 as it is already identified as authenticated node.
Nodes falling in sector between 1 and 4 encompass special levels of recognition and the different constraints are used to reveal their present sector. Nodes arrangement varies over time. The alteration of nodes leads to change over from high reputation to low reputation node. The higher reputation node begins to behave maliciously causing less trusted node falling to less favorable sector 6. The secure key model allows the network to develop into a multiple clusters of various trustworthiness levels. These different levels of trustworthiness tolerate higher layer applications to edge their communication only to one selected sector versus any other sectors.
Reputation procedure with a reputation table shows the interaction between the key components of reputation model. Reputation system is designed in order to offer automatic and autonomous routing decisions to the fundamental routing protocol based on the available adjacent reputations. The routing protocol used in secure key model is authenticated group key agreement protocol.
3.2.1.1 Reputation management:

The major task of reputation management is the responsibility of maintaining nodes reputation records. The reputation management works in the process of managing nodes storage and retrieval with respect to nodes adjacent reputation information. Reputation management organizes the process of the other components and act as the attention point for all the events held inside the reputation system. Adjacent Reputation Record is the object that represents reputation observation. Each node holds N adjacent reputation records where N is determined by the node's memory capacity, CPU powers maintenance in updating records and other resource utilization. Nodes with higher reputation and symmetric range sufficient reputation records about other nodes in order to provide enough reporting of the nodes in its own sector. Nodes reprocess the records using termination time to provide stability in different overheads with the requirement to attain adequate reputation.