Security Improvisation through Node Trust Prediction Approach in Mobile Ad Hoc Network

—Identifying trusted nodes for safe communication is a key challenge in mobile ad-hoc networks. Node compromises a service and leads to uncertainty in node behaviour. Computing the node trust and node management will enhance the security aspect in MANETs. This paper proposes a security improvisation based on a Node Trust Prediction Approach (NTPA). NTPA aims to prevent the interference of an anomalous node in a MANET. There by improving the security and data delivery output . The NTPA calculates the node trust prediction by evaluating the four most frequent actions that are performed by a node in the communication process. Node authorization is a key aspect in the evaluation of an ad hoc network’s security. In the proposed method, we monitor the valid and Invalid Authorization of a node. Data delivery reliability is measured with S uccess of Packet delivery and Loss or Drop of packets. In this paper, NTPA is compared with SAR (Security-Aware Routing) and AODV (Ad hoc on-Demand Distance Vector), to evaluate the efficiency in an ad hoc network. The empirical results show that there is an increase of 25% packet delivery and a 40% reduction in routing overhead.


Introduction
A mobile ad-hoc network (MANET) is a network having a number of wireless devices that are proficient in interacting with each other.Due to dynamic environment, sharing of channels and computation constraints, MANETs are more vulnerable to security attacks compared to traditional wireless networks.For a possible multi-hop communication among non-adjacent nodes, former nodes must function as routers.This constraint poses a great challenge to find a reliable node that acts as a router for secure communication in MANETs.
To accomplish benchmark performance in reliable communication, the routing technique must be adaptable to dynamic environmental format.The mobility of the nodes can result in loss of existing links.Hence there is a need to discover new paths to overcome the communication disruption.The presence of faulty nodes hinders ad hoc networks through entering "incorrect routing updates", "responding to out dated routing information", "changing routing updates or advertising false routing information" and "dynamic characteristics of ad hoc networks" [4], [9], [11], [12].Many frameworks [1], [2] and methods [5], [7], [3] are available in relational and reliable computational-based security models that are more effective for limited resource communication.
The mechanism of routing in MANET relies entirely on the coordination and participation of neighbourhood nodes [11].A misbehaving node causes data loss and imbalances the network.A powerful, steady and secure routing protocol is required to achieve quality, secure performance standards, effective maintenance of the node-link and mobility of a MANET.
The trust system will keep the available network services in safe state.For instance, "quality assessment", "access control", "authentication", "malicious node detection", and "secure resource sharing" of information received [3], [6], [8], [9], [13].As a result, it is significant to regularly estimate the trust values of nodes based on certain metrics and calculations.In this paper, we propose the Node Trust Predicting Approach (NTPA) to recognize the Node Trust.It is used to establish more secure communication on Mobile Ad hoc Networks.Many strategies related to trust computation [15], [16], [17] have been proposed in these networks.In the proposed methodology, a comprehensive node trust prediction based on "Valid and Invalid Authorization", and "data routing actions" have been carried out to improve the security of MANET by increasing node-level trust characters.
The remaining paper is structured as follows.Section 2 discusses correlated research on trust-based routing and security enhancement.The proposed Node Trust Prediction Approach (NTPA) mechanism is discussed in Section-3.Section-4 explains the experiment and outcome assessment, followed by conclusions in Section-5.

Related Work
Network to establish trust management mechanism, effectively improves network security [1], [2].Trust has been noticed in several areas of security systems and has become more and more important in wireless networks [4], [10].Each security document method has its own subject qualification and filtering issues.Trust-based security techniques are important in MANET-based approaches and have been studied in numerous recent literatures [18], [19], [20].The rich literature surrounding trust and its management in the network makes us strongly suggest that this is an important and exciting area of research.The trust has an extensive diversity of alterations and functions as a concept that leads to disagreement over trust management terminology.And while prevention-based approaches prevent inappropriate behavior, malicious nodes still have the opportunity to contribute to the routing process and disrupt appropriate routing concern.Commencing the familiarity of wireless network security design and the multi-level security mechanism it is highly required for future secure communication.K. Govindan et al. [3] conducted a detailed investigation of various trusted computing methods for MANET.It provides MANET designers multiple viewpoints on the impression of trust, a considerate of the attributes that have to be well thought-out in extending trust metrics, and within reach on how to calculate trust.It suggests an allinclusive assessment of a variety of trust calculation methods, as well as comparisons of diverse "attack models" and "computational prerequisites".It also analyzes diverse documents of "trust dynamics", such as "trust propagation", "aggregation" and "forecasting".
Z. Wei et al. [5] projected a" trust management scheme".The trust model has two components: "direct surveillance of trust" and "indirect surveillance of trust".Direct surveillance from the observer node, the trust value is derived using "Bayesian presumption", which is an uncertainty assumption, whilst the complete "probability model" be able to be characterized.On the other side, indirectly observing second-hand information about neighbor nodes, also called observer nodes, using DST (Dempster-Shafer theory) to derive trust values, DST is another type of uncertainty inference that can be derived indirectly.By uniting these two constituents in the trust model, a further perfect trust value for observing nodes in MANETs can be obtained.
A. Pirzada et al. [21] proposed routing-based direct trust computing.It describes the trust as a fractional value in [x0, 1] and assesses the performance of the AODV and DSR protocols and analyzes it with the proposed trust scheme.In this scenario, nodes always monitor neighbors to build and update trust relationships.Sun et al.Consider the uncertainty of trust as being properly implemented by observed nodes, using entropy to develop a trust model, and assessing trust values through straight surveillance.The "Indirect or indirect information" possibly imperative in assessing the trust of observing nodes as compared to direct observations in confidence assessments.For illustration, a set of proofs as of neighboring nodes be able to identify that a malicious node achieves a good situation for one observer instead of another.
A routing protocol based on "Security-Aware Routing (SAR)" [22] mechanism transforms the "AODV routing protocol" [14] to comprise the trust hierarchy of integration nodes for path assessment and assortment.The protocol executes the trust level in terms of to the organization level and uses the "shared key" of every one layer in order to facilitate nodes be able to specify security prerequisites whilst apply for routes.Merely nodes that assemble this necessity can contribute to the routing.But how to classify "node trust", "key distribution" and other key awareness of nodes is a significant scope of the current research works.
Predicting node trust dynamics means that node trust should change according to its behavior.Non-transitive means that "Node-A does not unavoidably trust Node-C if Node-A trusts Node-B and Node-B trust Node-C".In asymmetry, it means that "Node-B does not essentially trust Node-A if Node-A trusts Node-B".Perspective dependent resources that trust estimation is usually derived from the behavior of nodes.The dissimilar phase of action can be accessed through dissimilar trusts.For instance, if a node consumes a smaller amount of power than it could not be capable of self-assured the message to its neighbors.In such case, the energy trust of this node will decrease, however, the "security trust" of this node determination not change because of the status.To calculate the level of trust on a node, it is important to un-derstand the various trust features used for trust definitions, metrics, and trust calculations.The trust of a meticulous node is a prejudiced consideration of the reliability and accuracy of the information the node's agent receives or passes through in that given context.The MANET routing protocol is used to evaluate the proposed protocol for SAR [22].The following sections define the process of distributing trust keys and trust calculation and routing mechanisms.

Proposed Node Trust Prediction Approach
The proposed Node Trust Prediction Approach (NTPA) is shown in Fig. 1.It is a three-step process.First step involves, a node to secure itself in a given network (A in Fig1).To secure itself, it acquires a Trust Certificate from Certified Authority (CA), which consists of a "Public Key as CApub_key" and "Private Key as CApvt_key".Based on these authorized keys it authenticates the node during data routing.In the second step (B in Fig1), it performs the Node Trust prediction and finally(C in Fig1) based on individual trust prediction of a node, it performs a secure trust-based routing.

Acquisition of trusted certificate
Before joining the network, every node in the network has to acquire a secure "trust certificate" from a "trusted certification authority (CA)".Security certificates that have been released cannot be "Withdrawn or expired"untilanode exists in the network.If the node's trust value drops below the 40%threshold, the certificate will be invalidated.This means that the legitimacy of the certificate determination is maintained until the credibility is preserved.In such progression, NTPA is capable of recognizing nodes that have illegitimately own a valid certificate and avoid the intrusion of malicious nodes in the routing process.The certificate issued by CA is denoted as  #$%& .

Paper-Security Improvisation through Node Trust Prediction Approach in Mobile Ad Hoc Network
Where, E -Encryption Each node is preloaded with this CAcert certificate before joining the network and it produces  )$* in case of authentication verification.The acquired authentication acts as an attestation for a node, thus making it a valid node in the network.

Node trust prediction computation
The evidence of a node trust was computed utilizing the three monitoring factors of a node: authentication, data delivery, and data loss.The process of authentication involves a node to produce its CA authentication key  )$* to recognize it as valid Avalidandinvalid authorization as Ainvalid.Similarly, the process of data delivery as Rdata and loss as Ldatais measured based on the acknowledgement being received by the destination or intermediate nodes.Avalid/Ainvalid and Rdata/Ldata are utilized for computing the node trust prediction value as NTPvalue of a node.
Each of these values is recorded on each data packet being transmitted through the participating intermediate nodes.The positive outcome of authorization Avalid and successful delivery of data packets Rdata increases a node credit by 1 and at the same time its debited by 1 in case of invalid authorization Ainvalid and in case of data delivery fail or loss Ldata is increased by 1.Each of these parameters can be illustrated as: On the basis of the values of Avalid , Ainvalid , Rdata, and Ldata , we compute each individual node's rate of authorization as Arateand rate of data delivery as Rrate.The percentage of this rate can be computed using the Eq.1, Eq.2 and Eq.3 as given. (1)

Data Loss
Based on Arate, Rrate, and Lrate values, we compute the NTPvalue using the Eq. 4. The NTPvalue is utilized for the runtime trusted node selection to route data from the source node to the destination.The NTPvalue used as the value limit for a node's consideration for communication and this node is used as a source node.In the next section, we discuss the trust prediction routing mechanism based on NTPvalue.

Trust prediction-based routing
The main objective of routing methods in the Adhoc network is to provide efficient data routing.Each node in the proposed protocol sends data through the discovered path and predicting each node's NTPvalue.The protocol presupposes that the entire nodes on the network are, to begin with, reliable and trustworthy.The "trust value" is computed based on the value of this NTPvalue derived using the Eq. 4.
The source sends data packets to the destination through the cached route by the route manager.Initially, all nodes NTPvalue is considered as 100%.To begin with, source selects the shortest hoping path.During routing each node asks its neighbour nodes to produce a CAcert certificate to get authenticated before transmitting the data packet.In case of success it updates Avalid and Ainvalid. .Similarly it updates it Rdata and Ldata in case of successful data delivery or failure.On continuous observation of these values it computes its Arate , Rrate, and Lrate values and finally it computes its NTPvalue and updates the routing table.An illustration of node trust routing table is given in Table-1 The "intermediate nodes" also go after the identical plan as the "original source node" function.Table-2 illustrates the node's routing table for "Node-4" which has two hopes.According to this table input and its first_hop NTPvalue it selects the route R1 node as its NTPvalue value is higher in comparison to R2 node.The source node performs a list of "sequential numbers of packets" sent by it.On arrival of successful acknowledgment, source updates the Rdata of each node in the route.In case of packet loss, it punishes a node by decreasing the Rdata value.This dynamic routing based on the runtime NTPvalue(derived from equation ( 4))provide a reliable and secure routing.Also assures a confirmed delivery in case of arbitrary selection of a node or predefined route nodes.

4
Empirical Assessment

Simulation set-up
Experiment simulation is performed using Glomosim Simulator, we have modified the "AODV protocol" to evaluate the "NTPA protocol", and evaluate the effectiveness of our projected protocol with "SAR" [28] and "AODV" [16].As we added security parameters, the route request and the size of the routing's packet header are increased.We configure the simulation by means of the subsequent setup constraints as illustrated in Table-3.
We execute the experimentation based on the Table-3 simulation factor for a time of "600 seconds" with an RWP movement behavior model with varying speeds between "0 to 100 m/s".We execute the simulation in six dissimilar speed as configured in Table-3.For data routing, we used "15 source-target pairs" of "constant bitrate (CBR)" traffic of "4 packets per second", and each "512 bytes" in size.The assessment was conducted in two different situations.First, there were not any misbehaving nodes in the network, followed by 25% of the misbehaving nodes added.The experimental outcomes demonstrate the "overhead introduced" caused by security enhancements and "throughput" comparisons.
All nodes behave normally during route discovery.However, 25% misbehaving nodes are randomly selected by route simulator.Nodes will behave abnormally discarding all data packets and generating incorrect trust prediction.However, using signature verification in NTPA can detect any type of packet modification attack.Packet drops can isolate abnormal nodes from the network.For performance assessment, we determined the following "throughput" and "control overhead" metrics.

Result analysis
Throughput: Throughput is determined to utilize the "Total Number of Delivered" by "Total Number of Packets Send".To provide performance comparison analysis for a better insight of our simulation results shown in Fig. 3 and 4 show the throughput outcome of the protocol.In the absence of malicious nodes, all protocols show similar results.Compared with the AODV and SAR protocols that have malicious nodes, NTPA shows improvement.Increase in the throughput is due to secure data routing by the trusted nodes.In the nonexistence of malice, it shows an average outcome as a result of security overhead.NTPA achieved a 25% improved packet transfer compared to other approaches, while others showed a 10-20% drop with 25% of malicious nodes presences in the network.Control overhead: Control overhead determination is based on the "total number of control packets originated and forwarded" by the approaches for the period of complete communication progression.To provide performance comparison analysis for a better insight of our simulation results shown in Fig. 5 and 6show the control overhead between NTPA and other methods.In the "absence of malicious nodes", the entire protocols comprise a comparable overhead ratio.
However, in presence of malicious node NTPA shows lower routing overhead.Because NTPA uses trust prediction to identify misbehaving nodes and discards them to minimize control packets.In SAR repeated security checks are perform during communication.In AODV, during speed changes high speed links will fail.Also malicious nodes add high control packet switching.This raises their routing overhead compare to NTPA.We propose an innovative "Trust-based secure routing protocol", NTPA, for mobile ad hoc networks.NTPA authenticates the routing node based on the "trust certificate" and "trust prediction" computed during communication.NTPA supervises multiple routes to reach the target node.Each node in the network accumulates a "local trust value" of each other node and maintains the routing table.The NTPA calculate the trust value of all nodes in the first hop.Intermediate nodes, route, data packets by choosing a path with a higher value of trust nodes.The mechanism for security improvisation based on NTPA helps in improvising the throughput of PDR during communication.The empirical results show a 25% of high PDR with minimal overhead.In mutual cases, NTPA commences a practical network load to ascertain more packet transmission rates.This increase can result in shorter trust values and convergence time.In addition, more work needs to be done in the future to measure the effect of any change in protocol parameter values and to find the best value for different settings.

Table 1 .
. Source Node Routing TableFor example, in Table-1, it shows 5 routes to destinations and each route 1st hop and NTPvalue, the most efficient and shortest route is R1, but as per the NTPvalue de-

Table 2 .
Routing Table for Node-4 as Intermediate Node