Research on Trust-Based Secure Routing in Wireless Sensor Networks

With the increasing scale of wireless sensor networks, the limited resource of sensor nodes and the uncertainty of the layout environment, sensor nodes are very vulnerable to malicious attacks, but the traditional security mechanism is not suitable for wireless sensor networks. This paper summarizes the common attack forms, security technologies and the security mechanisms of trust-based routing protocols in wireless sensor networks proposed in recent years, then analyzes and compares these secure routes.


Introduction
Wireless Sensor Networks (WSN) is a self-organizing, multi-hop wireless network consisting of a large number of stationary or moving sensor nodes [1] . Due to its many functions, WSN has received more and more attention in recent years. The sensor nodes of WSN can continuously perform data acquisition, event detection, event identification, position monitoring, and node control. These characteristics and wireless connection methods of sensor nodes broaden the application prospect of WSN, involving various fields, such as military, ecological environment monitoring, intelligent transportation, etc [2] .WSN is considered to be one of the important technologies affecting the future of human beings. This emerging technology provides people with a new way to obtain and process information. Due to the characteristics of WSN itself, there is a big difference between it and the existing traditional network technology, which presents many new challenges. Since WSN is of great significance to the country and society, the research on WSN at home and abroad is being carried out enthusiastically. It is hoped that it will attract the attention of this emerging technology in the field of measurement and control, and promote the research, application, and development of this new technology with national strategic significance.
The sensor nodes in the WSN have the characteristics of low cost, small size, and certain data processing capability. Because of these characteristics and the size of wireless sensor networks is increasing these years, coupled with limited power, computing ability and storage capacity of sensor nodes, and mostly nodes are deployed in the field, these sensor nodes are very vulnerable to attack, so the security of wireless sensor networks has become a major problem.
Research on low-energy routing protocols for WSN has been a hot research topic. Low energy adaptive clustering hierarchy(LEACH) [3] , proposed by Heinzelman WR et al., is one of the most classic clustering routing algorithms. In recent years, Zhang Haiyan et al. proposed a routing algorithm based on K-means(KBECRA) [4] . Li Liuya et al. proposed a high energy-efficient routing protocol  [5] based on CFSFDP clustering. The routing protocol effectively balances the energy consumption of the wireless sensor network to a certain extent. However, most low-energy routing protocols only focus on energy consumption but have no concern about security issues. This paper summarizes the various types of attacks commonly found in wireless sensor networks and the security mechanisms of various wireless sensor networks at the routing level by studying various wireless sensor routing protocols and trust-based security mechanisms proposed in recent years.

Common Attack
Most of the attacks on wireless sensor networks are usually based on the energy-limited characteristics of sensor nodes. The energy of the sensor nodes is depleted by various means of attack, and the nodes in the network are prematurely invalidated, thereby destroying the structure of the entire network, to make wireless sensor networks paralyzed. Due to the limitations of various aspects of wireless sensor networks, the attack methods for such networks are various. The classification of the attack is shown in figure 1.

Active attack
The common form of active attack is to inject error or bad routing information into the wireless sensor network, resulting in a large amount of useless data in the network, so that the sensor network whose energy is very limited further consumes a large number of resources in data processing, causing network paralysis. It includes but is not limited to the following forms. [6] Sybil attack mainly exists in the peer-to-peer network. The malicious node confuses itself, integrates into the sensor network and acquires the trust of neighboring nodes, and declares to other nodes that it has multiple identities so that it can improve the probability to be selected as a relay node in the routing algorithm. After elected as a relay node, the malicious node can also combine other attacks to cause more damage. This kind of attack is extremely harmful to the routing algorithm based on location information.  [7] In this attack mode, the attacker captures the sensor node and does not forward the data packet completely in the same way as the normal ordinary node in the data forwarding. Instead, it selectively forwards some data or directly discards the data packet to destroy the network. [8] In this kind of attack, the attacker broadcasts a message to other nodes, telling other nodes that it has sufficient conditions to act as a relay node, thereby tricking other nodes to select the malicious node as the next hop of the routing path. Then a black hole is formed near the malicious node, and a large number of data packets are intercepted.

Passive Attack
Passive attacks focus on the stealing of information. Because wireless sensor networks mainly transmit data through wireless radio frequency communication, attackers can obtain information in the network by listening to channels, etc, thereby achieving the purpose of the attack. It mainly includes camouflage rivals, data stream analysis, and monitoring, etc.

Key management
Encryption technology is an important security technology foundation in all computer-related fields, and its application in the field of wireless sensor networks also plays a fundamental role. Key management technology is the core of encryption technology. There are three types of key management techniques commonly used in wireless sensor networks.

Centralized Key Management
This method mainly refers to the distribution mode based on the Key Distribution Center (KDC), such as SPINS [9] .

Distributed Key Management
This method mainly refers to the key management mode using the key pre-allocation scheme. The most typical representative is E-G [10 ] . Later, many E-G improvements have appeared, such as Q-composite [11] , DDHV [12] , etc.

Hierarchical Key Management (Semi-distributed Management Mode)
This method clusters the nodes of the WSN and uses the cluster heads to operate on key assignment, negotiation, and update between sensor nodes in the cluster. Mainly including LEAP [13] , URKP [14] etc.

Refer security as a design goal of routing protocols
In recent years, the design of routing protocols for wireless sensor networks is mostly aimed at energy consumption. If security is the original design goal in the design process of routing protocols, this will make the design very complicated. However, it can better evaluate the energy balance problem after joining the security mechanism.

Extend the security of existing routing protocols
There are many mature energy-efficient wireless sensor routing protocols, such as LEACH, KBECRA, and CFSFDP, which can be improved on this basis and added security mechanisms.

Combine a variety of security mechanisms
Key management, authentication technology, timestamp mechanism, and reputation mechanism are currently relatively complete security mechanisms. For example, using the reputation evaluation value in WSN. The node is used to evaluate the behavior of the neighbor node or use the base station for monitoring and detection. After the routing protocol is improved, it can effectively resist the Sybil attack, the worm attack, and the selective attack.

Intrusion Detection System
In the related analysis of wireless sensor network intrusion detection, the research of intrusion detection architecture is the basis, and its design and application must be carried out for specific scenarios. In the current intrusion detection structure, according to the special relationship between detection nodes, can be roughly divided into the following three types.

Distributed detection system
The intrusion detection system is installed on all nodes of the wireless sensor network, and the node itself handles the intrusion behavior.

Peer relationship detection system
The nodes in the structure work together to detect intrusion behavior in the network.

Hierarchical detection system
It is mainly used in heterogeneous wireless sensor networks. The nodes in the sensor network are divided into common nodes, aggregation nodes or cluster head nodes. Each node handles the intrusion behavior to different degrees.

Early research
The trust-based routing protocols in the early literature were mainly based on existing protocols, with some special methods to deal with trust factors, including T-RGR [15] , EMPIRE [16] , etc. With the deepening of research, more trust-based routing modes have emerged, and multiple attributes have been combined more comprehensively, including ERRM [17] , TCLM [18] , TARF [19] , and SETM [20] , etc. [21]. This literature proposes a framework based on distributed trust and a mechanism for selecting trusted cluster heads. The mechanism proposed in this paper reduces the probability of malicious nodes being elected cluster heads to a certain extent but does not pay much attention to trust evaluation. [22]. This literature adds a trust management module based on the original LEACH protocol and uses a trust information exchange mechanism to construct and maintain trust information between adjacent sensor nodes. The key is to prevent malicious nodes from becoming cluster heads to creating larger harm. Although the algorithm is robust to attacks by malicious nodes, it is vulnerable to collusion attacks.

Recent research
In recent years, wireless sensor networks have become a hot research topic, and there are more and more researches on secure routing for wireless sensor networks. This section summarizes some work in this field in recent years through the research of trust-based routing protocols, and analyze the innovation, advantages, and disadvantages. [23]. This literature proposed a fuzzy-based method to improve the security of routing protocols. The node uses the trust model to calculate the trust degree of its neighboring nodes and uses the fuzzy rules to select reliable routes. However, the energy consumption of the proposed algorithm is very large. There is clearly a need for further improvements in the field of energy-constrained wireless sensor networks. [24]. This literature proposed a routing algorithm based on trustworthy core tree for WSN (RATCT). In this literature, the cluster head selection algorithm comprehensively considers the trust degree and residual energy of the node, so that the probability of the node with higher trust and more energy being elected as the cluster head is higher. All cluster heads are organized into a trust core tree, the base station is the root node, and then the core tree extension covers all nodes. At the same time, a trust model is used to evaluate trust and detect malicious nodes. Each node keeps a pair of public and private keys to encrypt and sign the data packets to be sent. The base station detects malicious attacks by analyzing encryption and data packets. Although the algorithm can effectively monitor malicious behavior to a certain extent, it uses several methods that require additional energy to build the core tree and calculate the trust value. [25]. This literature proposed a Trust-based Energy Efficient Routing Protocol (TEESR), which uses appropriate authorization and flooding mechanisms to restrict neighboring malicious nodes. Through the use of the node trust value, a coverage area network and multipath security route are established, then the cluster head and the base station node select a secure route. Although the protocol proposed in this literature can cope with convergence holes and wormhole attacks, it can't do anything against internal attacks. [26]. This literature proposed a low-energy secure routing algorithm based on the Cuckoo search algorithm, which uses clustering, trust awareness, and multipath routing to extend the network lifetime. The calculation of the trust degree is based on the forwarding rate factor, the packet consistency factor, and the residual energy. The cluster head is responsible for calculating the trust value of the common node in the cluster, and the trust value of the cluster head is calculated by the base station or the neighboring cluster head. The routing path of the algorithm only considers the trust score of one expectation. [27]. This literature proposed a trust-aware routing protocol with multi-attributes (TRPM). The trust model of the protocol includes communication, data, energy, etc., and adds trust evaluation attributes and a sliding time window with the function of measuring attack frequency. TRPM shows good performance when dealing with various routing targets and trust target attacks. Improves routing processing quality, extends network lifetime and reduces end-to-end latency.

Literature
Vulnerable to collusion attacks and unable to defend against internal attacks.
Literature [27] Establish reliable routing based on communication, data, energy and recommended trust evaluation.
The trust model consists of communication, data, energy, and recommended trust evaluation.
Works well when dealing with various routing and trusted node attacks.
Does not include the trust-aware cluster head selection mode.

Conclusion
In recent years, wireless sensor networks have been widely promoted in many fields, and security issues have become more and more concerned. The network layer is an extremely important part of wireless sensor networks, and its security directly affects the entire network. Due to various limitations of WSN, most of the research on WSN secure routing direction in recent years is based on trust evaluation. This paper summarizes a large number of secure routing protocols based on trust mechanism, then analyzes and compares these protocols. As can be seen from these papers, most of the future research on wireless sensor network security routing will focus on cluster-based routing protocols, and also explore the use of multiple security schemes to ensure the security of wireless sensor networks.