TBSMR: A Trust-Based Secure Multipath Routing Protocol for Enhancing the QoS of the Mobile Ad Hoc Network

Mobile ad hoc network (MANET) is a miscellany of versatile nodes that communicate without any fixed physical framework. MANETs gained popularity due to various notable features like dynamic topology, rapid setup, multihop data transmission, and so on. )ese prominent features make MANETs suitable for many real-time applications like environmental monitoring, disaster management, and covert and combat operations. Moreover, MANETs can also be integrated with emerging technologies like cloud computing, IoT, and machine learning algorithms to achieve the vision of Industry 4.0. All MANET-based sensitive realtime applications require secure and reliable data transmission that must meet the required QoS. InMANET, achieving secure and energy-efficient data transmission is a challenging task. To accomplish such challenging objectives, it is necessary to design a secure routing protocol that enhances the MANET’s QoS. In this paper, we proposed a trust-based multipath routing protocol called TBSMR to enhance the MANET’s overall performance. )e main strength of the proposed protocol is that it considers multiple factors like congestion control, packet loss reduction, malicious node detection, and secure data transmission to intensify the MANET’s QoS. )e performance of the proposed protocol is analyzed through the simulation in NS2. Our simulation results justify that the proposed routing protocol exhibits superior performance than the existing approaches.


Introduction
Mobile ad hoc network (MANET) is an assortment of mobile nodes that communicate without any fixed physical framework. MANETs have many striking features like varying topology, rapid setup, and multihop wireless communication. All these features make MANET suitable for various time-sensitive applications [1,2]. Ad hoc network renders a promising communication facility where physical infrastructure is difficult to establish. Furthermore, MANETs allow mobile nodes to exchange information without any physical framework and administrative activities. Hence, these networks are dynamic, self-organized, and autoconfigured, allowing nodes to move arbitrarily during communication. MANETs also play an essential role in Industry 4.0. ese networks can be extended and integrated with emerging technologies like cloud technologies, IoT, and machine learning techniques to develop smart applications for automating industrial needs [3,4]. e structure of the MANET is depicted in Figure 1. Implementation of a secure routing protocol by ensuring QoS is a challenging task in the MANET because of its dynamic topology [5][6][7]. MANET facilitates open communication infrastructure through which any versatile node can easily join the network and participate in data transmission [8]. is open communication infrastructure provokes security breaches in the MANET [9].
Consequently, the secure and reliable routing in such a network is difficult to accomplish. Generally, routing protocols in MANETs are classified into three categories based on their design and routing process: (i) proactive routing protocols, (ii) reactive routing protocols, and (iii) hybrid routing protocols. e proactive routing protocol establishes and maintains all the routes in a routing table prior to the communication. In this category of routing protocols, route setup and route maintenance tasks are accomplished by periodically exchanging the control packets. Transmission of these control packets for route establishment and maintenance leads to routing overhead in the MANET. ese proactive routing protocols are suitable for miniature networks.
Unlike proactive protocols, reactive routing protocols establish routes whenever the nodes require them. is route establishment process reduces the network overhead that occurs due to the periodic exchange of control packets in proactive routing protocols. Reactive protocols can determine an optimal path with low packet delay and network overhead compared to the proactive routing protocols. Furthermore, reactive routing protocols apply to more extensive networks. Another category of routing protocols includes hybrid routing protocols. ese protocols commingle the benefits of both proactive and reactive routing protocols.
Ad hoc on-demand distance vector routing protocol (AODV) is a predominantly used reactive routing protocol in the MANET. Many researchers introspected AODV's performance by considering multiple factors and also identified various reasons that cause security breaches. is protocol has the following flaws: (i) ere is no mechanism to handle congestion. (ii) e existing protocol does not support multipath routing. (iii) It is susceptible to various security attacks [10]. (iv) It does not have any predefined mechanism to handle packet losses. (v) It does not have any mechanisms to ensure QoS. (vi) ere is absence of power optimization concept.
AODV is highly susceptible to various attacks like black hole attack, wormhole attack, DoS attack, and so on. To resolve these security breaches, it is necessary to upgrade the AODV protocol.
Many researchers have proposed different flavors of AODV protocol to handle the issues mentioned above. But no versions of AODV protocols handle all the issues discussed above together as a single protocol. erefore, the main objective of this proposed protocol is to provide a secure and efficient routing by reducing the packet losses and thereby enhancing the QoS in the MANET. In this paper, we proposed a TBSMR protocol to perform routing by considering the following factors for intensifying the network's efficiency: In MANET, the fundamental reasons for packet loss are the presence of noxious nodes and lack of sufficient battery power of the nodes [11]. e proposed TBSMR protocol integrates all the properties as mentioned above for enhancing the QoS in the MANET.  bushfire in Australia. MANET-based real-time applications demand precise time-sensitive data transmission. Also in emergency situations, safe sheltering points and escape routes must be accurately delivered to the people in a timely manner. Many existing routing protocols do not emphasize these constraints. Furthermore, the existing routing techniques are based on low-level parameters like; delay, routing overhead, bandwidth, and so on.

Related Work
Many researchers have proposed QoI-based source selection schemes to transfer time-sensitive critical information. Arsalaan et al. [12] propounded a low overhead source selection approach and QoIT that explicitly considers the user requirements to determine an optimal source, to allow information exchange in emergency situations, by avoiding bottleneck issue.
Convergence of MANET with IoT enlightens another possibility of research in smart ubiquitous computing where ad hoc network plays a vital role in implementing smart city applications. Smart city applications integrate different types of applications of various domains that require different types of message exchanges. Routing in such applications is a challenging task because of the diversity of nodes and disparate message structures. Intelligent routing techniques are required to meet the challenges of Industry 4.0. Intelligent routing protocols can be developed by using emerging technologies like machine learning, bioinspired optimization algorithms, soft computing, and so on.

Existing QoS-Based Routing
AODV is a conspicuously used reactive protocol in MANET. But this protocol exhibits many flaws, which are mentioned in Section 2. To overcome these flaws, many researchers developed many AODV-based routing protocols for intensifying the network's efficiency. is section emphasizes the recent flavors of AODV protocols propounded for enhancing the QoS.
Bhagyalakshmi et al. [13] proposed a Q-AODV protocol to determine a noncongested route based on the queue vacancy parameter. is queue vacancy parameter is used to reduce the number of intermediate nodes participating in the route exposure state, thereby reducing control packets' transmission.
Sarkar et al. [14] proposed an enhanced Ant-AODV protocol for optimal route selection in MANET. is protocol uses the ant colony optimization concept for the selection of optimal routes. In this technique, routing is done by computing the pheromone values of all the available paths. A path having the highest pheromone values will be used for transmitting packets from source to destination.
Jhajj et al. [15] propounded the EMAODV protocol for handling congestion. is protocol makes use of the TTL parameter to avoid the flooding of RREQ packets. is TTL parameter is used for identifying the active nodes for forwarding the packets. Only these active nodes are used for forwarding the packets. Unlike active nodes, the other nodes that do not respond to RREQ packets will be treated as silent nodes and are not involved in routing them.
Subramanian et al. [16] developed trust-based AODV in which packets are sent through trusted nodes. A node whose trust value is greater than the threshold value is treated as a trustworthy node; otherwise, it is considered an untrustworthy node. In this protocol, trust values are determined based on the number of request packets, reply packets, and data packets forwarded by each node.
Zhaoxiao et al. [17] proposed an energy-aware EAODV protocol in which a path with low energy cost and having a larger capacity is selected for data transmission. is protocol uses a priority weight parameter to predict the nodes' lifetime based on the present network traffic. Table 1 outlines the recent existing protocols along with the properties considered to accomplish QoS routing. In our literature study, we considered recent AODV-based protocols developed to overcome the flaws of AODV protocol and we perceived that many researchers considered only a few specific aspects in extending the AODV protocol for enhancing the efficiency of the network. In the implementation of the proposed system, we considered diverse factors like congestion control, malicious node detection, packet loss reduction, and available battery power of nodes during packet transmission. We contemplated all these factors for implementing the proposed TBSMR protocol through which the QoS can be enhanced.

Proposed Methodology
e proposed TBSMR protocol functions in three phases for reliable packet transmissions. e three phases of the TBSMR protocol are as follows: (i) Route exposure phase. (ii) e malicious node detection phase. (iii) Information forwarding phase.
is proposed TBSMR protocol is an amended version of the AODV protocol. TBSMR protocol overcomes the flaws of the AODV protocol. In the TBSMR protocol, malicious nodes are detected at every stage in communication. Moreover, the packet loss reduction mechanism is also used for reliable packet delivery. In this protocol, a spurious RREQ is broadcasted by the source node during the initial route revelation process. is spurious RREQ packet contains a fake destination address and destination sequence number. For this spurious RREQ packet, only a malicious node will respond with the RREP packet by claiming that it is having an optimal route to the destination [18,19]. In this way, the source node identifies the malicious node based on the invalid RREPs received. After identifying the malicious node, the source node propagates this information to all other nodes so that the malicious node will not be considered for forwarding packets and detached from the network. In this way, malicious node detection and elimination are done at the earlier stages of communication. All the nodes other than the malicious nodes are treated as trusted or trustworthy nodes. Also, during communication, malicious nodes are detected and eliminated by computing the nodes' trust values. If the trust value of a node is less than Security and Communication Networks the threshold trust value T thresh , then that particular node is marked as a malicious node. A trust value of a node Y is computed based on the trust opinion of its neighbors. Suppose that node X is a neighbor of node Y. Node X can determine the trust opinion on node Y by using a function T (X, Y), where T (X, Y) is a function of three parameters and it is mathematically expressed as where P (X, Y) represents successful packet transmissions from node X to node Y; N (X, Y) indicates failure packet transmissions from node X to node Y; U (X, Y) represents uncertainty factor that is initially set to 1. e uncertainty factor of value 1 represents that node X is not certain about the trustworthiness of node Y. Depending on the subsequent successful or failure packet transmissions from node X to node Y, U (X, Y) will be updated [18]. T (X, Y) is an average of three parameters and usually ranges from 0 to 1. e obtained value of T (X, Y) represents node X's trust opinion on node Y and will be maintained by node X in its routing table as trust_val of node Y. For a node to be trusted, its trust_val ≥ � 0.6. Every node in a network shares trust values of its neighbor nodes with all other nodes periodically, such that only trusted nodes are involved in information exchange while all the nodes whose trust_val< 0.6 will be considered as malicious and eliminated from the network. e trust value of a node is updated based on the number of packets forwarded or discarded by it on behalf of other nodes. is protocol is highly reliable because it supports the transmission of acknowledgment after successful transmission of packets from source to destination. When the destination node receives all the packets from the source node, it sends DR (Data Received) packet to the source node. After receiving the DR packet from the destination node, the source node marks the entire route as trusted through which it has transmitted data packets and received the DR packet successfully. Trust value of a route is expressed as follows: Trust (route) � successful packet transmissions/total packets transmitted.
For a route to be trusted, Trust (route) value ≥ 0.6.

Routing by Handling Congestion
is protocol supports the concept of multipath routing. is protocol allows source node to maintain multiple routes to the destination in a cache. ese multiple routes are used by the source node on occurrence of congestion or link errors. In this proposed protocol, the destination node is allowed to receive multiple RREQs from the same source node for which the destination node sends multiple RREPs. On receiving multiple RREPs from the destination node, the source node selects the best route based on the number of hops for forwarding the packets.
e alternate route towards destination will be stored in the sender's cache used in the future on the occurrence of congestion or link failure. e storage of an alternate path to the destination in a cache avoids invocation of the route discovery process and avoids overwhelming the selected route with packets. In existing protocols, during the routing process shortest and optimal path is opted for sending all the packets, which may lead to the congestion in the selected optimal route. Hence, unnecessary packet loss will occur, which results in reduced throughput. To address this issue, our proposed protocol is implemented so that whenever a selected optimal route is about to get congested, an immediately alternate path stored in the cache will be used by the sender for forwarding the subsequent packets. In this way, load distribution is done by determining the status of congestion of each route.
In this protocol, every node periodically sends the status of congestion to its neighbor using a QS (Queue Status) field in the HELLO packet. Each node determines the status of its available avg. queue by using the following equation: (1) Here, Wq is the queue weight and is a constant (Wq � 0.002 from RED, Floyd, 1997) and Instant_queue is instantaneous queue size [1].
If Instant_queue > Minthresh, indicates congestion. Based on the above calculations, QS field is set to either 0 or 1. is QS field is set in the HELLO packet and sent periodically by each node to its neighbors. Also, before transmitting a packet, every node checks the status of congestion of neighbor nodes by transmitting special acknowledgment packets.

Minimizing Packet Loss
A node may discard packets intentionally or due to lack of sufficient battery power for forwarding the packets. ese two cases result in unnecessary packet losses, leading to the degradation of network throughput [20][21][22][23][24][25][26][27][28][29]. In all cases as mentioned above, a node is treated as a malicious node.

Case 2.
Another reason for packet loss at a particular node is the lack of available battery power for forwarding the packets. In this protocol, before data transmission, the source node gets the status of its neighbor's available battery power by sending a DREQ packet. After receiving the DREQ packet, the node will send its available power to the source through the DREP packet. A node's available power or energy can be determined as follows: resh_Pow � 0.10 * Total_Power.where resh_Pow represents threshold power and Total_Power is the battery power of node.
If any node's Avail_Pow ≤ � resh_Pow, then this node will not be selected by the source node for forwarding the packets. In this way, a node with a lower energy level will not be considered for packet forwarding. e source node will select an alternate node with sophisticated energy for data transmission. is process will improve the packet delivery ratio through which the QoS can be enhanced.
For reliable data delivery, this proposed protocol scheme uses the concept of acknowledgment packets. Whenever the source intends to send packets to its neighbor, it sends DREQ packet to its neighbor requesting its neighbor node's status. If the neighbor node is active, it immediately responds with the DREP packet. After receiving the DREP packet, the source node sends its data. After completion of data transmission, the source node expects acknowledgment from the neighbor node. After receiving the data, the neighbor node sends DR packet to the source node. is data transmission process continues till all the packets sent by the source node reach the destination successfully. Finally, one acknowledgment packet is sent to the source from the destination node after receiving all the packets. Once the source gets an acknowledgment from the destination node, it considered the route and the intermediate nodes across the route are trusted, and at the same time, trust values are updated. In this way, the proposed scheme ensures reliable packet transmission by minimizing the packet loss; hence, the throughput of the network can be intensified. e required step to accomplish data transmission in the proposed TBSMR is shown in Figure 2. In Figure 2, node S is the source node, and it is having a route to the destination node D via intermediate node A. Before sending data packets, the source node first checks the trust value of node A. If node A is a trusted node, it checks node A's status by sending the DREQ packet. On receiving DREQ, if node A is having sufficient power for forwarding the packets and is not congested, node A will immediately respond with DREP. In this way, packets are transmitted by considering nodes' trust values, congestion status, and sophisticated energy availability.
During packet transmission, a malicious node may repeatedly send DREQ packets to the source node to capture data packets. Whenever a node sends three DREQ packets consecutively, it is marked as a malicious node by the source, and this information is propagated to all other nodes in a network.
e algorithm for malicious node detection is explained as follows (Algorithms 1 and 2). where nrecv is the number of received packets, nfowd represents the number of packets forwarded, and size_of_pkt means the packet size.

Simulation and Analysis
To analyze the performance of this proposed protocol, we used the NS2 simulation tool. Table 2 represents the parameter considered for simulation in NS2. e formulation of the MANET by using the proposed protocol is shown in Figure 3. e proposed protocol allows communication between trusted nodes only. To evaluate the performance of the proposed protocol, we considered the metrics like PDR, PLR, average end-to-end delay, and throughput.

Packet Delivery Ratio (PDR).
It is defined as the ratio between the total number of packets received by the total number of packets actually sent.

Security and Communication Networks
Step  Figure 4 shows that the proposed TBSMR protocol exhibits better PDR than the existing approaches. Figure 5 illustrates that the proposed TBSMR protocol has less packet loss ratio than the existing routing techniques. Figure 6 justifies that the proposed routing technique exhibits a lower average end-to-end delay in comparison with the other existing routing schemes. e proposed TBSMR protocol has better throughput than the existing routing approaches considered in this study, and the same is depicted in figure 7. Table 3 demonstrates precisely that the proposed routing technique outperforms by considering multifactors to intensify the QoS in the MANET.    All the simulated results justify that the proposed routing protocol exhibits better performance than the existing approaches in enhancing the QoS and making it suitable for real-time applications.

Conclusion
In this work, we proposed a routing protocol called TBSMR to enhance the QoS of the MANET. It is applicable for more extensive networks and considers multifactors like congestion, trust values of the nodes, and the available battery power of nodes during the routing process, which results in better performance with reduced overhead. Moreover, this proposed protocol supports multipath routing that minimizes the floating of unnecessary control packets for route establishment in congestion or node failure. is protocol also ensures secure communication by detecting malicious nodes. Our simulation results justify that the proposed TBSMR protocol gives better performance in PDR, PLR, average end-to-end delay, and throughput compared to the existing routing techniques. Overall, this proposed TBSMR routing approach enhances the QoS of the MANET besides ensuring secure communication.
In the future, we emphasize the implementation of security algorithms by incorporating encryption, decryption, and blockchain approaches for providing high security to the MANET.

Data Availability
Access to all the supporting data about this article will be given based on the requests directed through the data access committee and institutional review board.