Abstract
Source location privacy, one of the core contents of Wireless Sensor Network (WSN) security, has a significant impact on extensive application of WSNs. In this paper, a novel location privacy protection routing scheme called Energy Balanced Branch Tree (EBBT) is proposed by using multibranch and fake sources. This scheme has three phases. In the first place, the data of the source are randomly sent to a certain intermediate node. Then, a minimum hop routing (MHR) from the intermediate node to the base station is formed. Then, branch paths with fake sources are generated dynamically from some nodes on the MHR path. Finally, a tree-shaped structure from real source nodes and fake source nodes to the base station is achieved. In difference to the previous schemes, the location of the real source in the EBBT scheme does not affect the location and the number of fake sources. During the formation of the tree-shaped multibranch paths, the residual energy of nodes is considered sufficiently, and the control of the direction of each branch path is also involved. The influence of the number and length of branches on the network lifetime and network security is also investigated. Experimental results show that the proposed algorithm has the advantages of long network security period and lifetime, as well as high path diversity. Our simulation further illustrates that the EBBT scheme has favorable privacy of the source location without changing the network lifetime.
1. Introduction
WSNs (Wireless Sensor Networks) as the main component of Internet of Things (IoT) applications have been used widely in monitoring field [1, 2], intelligent agriculture [3], medical health [4, 5], and so on. With the continuous expansion of WSN application scenarios, its securities, such as the authentication and encryption mechanism [6, 7], have attracted extensive attentions. The security and privacy of WSNs are usually affected by the wireless communication media, dynamic topology, and limited resources. Specifically, wireless communication media used in WSNs is easily monitored, intercepted, or modified by an attacker. An attacker may infer the location information by listening to the transmission of network data and analyzing the network traffic [8].
In the monitoring events or object tracking applications, networks consist of some monitoring sensor nodes. The node that senses the monitored object is called the source node or source location. In general, the source node is responsible for collecting and sending data to the sink node. Considering that physical locations of source nodes are closely related to the identity information of monitored objects, the location information of these nodes is important to the whole system. For instance, in the pander-hunter game model [9], the concept of the source location privacy (SLP) is presented. After sensing the panda, the source node sends the collected data to the base station periodically. The hunter acts as an attacker, and the hunter tries to catch the panda by backtracking. Keeping the location of the source node from being discovered by attackers is the goal of SLP.
The researches on SLP protection mechanism put forward different strategies to cope with different attacks. Random walk is one of the most classical methods to resist the hop-by-hop backtracking attacks. And the fake packet injection is another well-known method against backtracking attacks. This method embeds fake traffic into real data making it difficult for attackers to distinguish real traffic from fake data and thus guarantees the security of source location [10]. Due to the reason that the fake source is generated statically, attackers can easily find the location of the fake source and thus discard it to find the real one [11]. The technology of multibranch path for forwarding fake packet would make the opponent deviate from the delivery path of the real packets and thus protect the source location when the attacker traces the source through hop-by-hop backtracking packets. This source location privacy protection can be implemented by building a tree structure route [12], where the backbone route and the transfer route are established. As the number of branches or false data traffic increases, the intensity of privacy protection will increase. Moreover, the energy consumption of network is also higher than before with the increase of branches.
The nodes that have balanced energy consumption in WSNs can greatly improve the lifetime of network. On the contrary, uneven energy consumption in network will incur hotspots, where the data traffic load is heavy. For instance, the nodes near the base station are responsible for more data packet forwarding. In addition, if a dead node appears in the network, the node energy of the nonhotspot area is about 90% of the remaining energy [13, 14]. Moreover, repeated usage of the same paths and nodes will lead to hotspots to reduce network lifetime. Therefore, how to improve the SLP with high energy efficiency appears to be a difficult task in WSN security.
Considering the limited energy of WSNs, balancing the energy consumption of nodes will help to improve the life of the network. Therefore, the residual energy will be an important factor in the selection of forwarding nodes. The diversity of paths can not only improve the security of the network but also balance the energy consumption of nodes. However, the number of dynamic false sources and the length of branches will be contradictory to the security and the network lifetime. The privacy protection mechanism in this paper will seek a dynamic balance between them. While strengthening the network security and fully improving the energy efficiency of nodes, we propose a novel energy balanced multibranch privacy protection scheme, called Energy Balanced Branch Tree (EBBT). Our contributions are summarized as follows: (1)A novel three-phase SLP protection scheme is proposed. This scheme uses constrained random walk, multibranch paths, and fake sources to achieve the protection of SLP(2)The selection of phantom source nodes is improved, and the constrained random walk is used. In the process of moving away from the real node, the phantom source moves towards the periphery of the network(3)The fake sources and multibranch paths are generated dynamically. The direction range of the branch path is determined to make the path more dispersive. And, the compromise of the length and number of branches are tackled to balance between the energy consumption and the security(4)Theoretical analysis and experiments are carried out to evaluate the security period, energy consumption, and the path diversity and data delay
The rest of this paper is organized as follows: Section 2 provides related work. The network structure and the adversary model are given in Section 3. The detailed EBBT scheme is described in Section 4. The properties of the EBBT scheme are discussed in Section 5. The simulations and performance evaluation are presented in Section 6. Section 7 concludes the results.
2. Related Work
In 2004, Ozturk et al. [9] firstly introduced the concept called privacy of source location and designed the panda-hunter model and the phantom routing (PR) protocol. The proposed protocol is a two-stage technique. In the first stage, data packets of the source node are randomly forwarded to the phantom source node. And the packets are forwarded to the sink node through flooding in the second stage. One year later, Kamat et al. [15] proposed an energy-efficient phantom single-path routing (PSPR) protocol in which the trade-off between energy consumption and privacy performance is considered. Later, Xi et al. [16] illustrated the drawbacks of the phantom flooding protocol, thus proposing greedy random walks (GROW) where both the source and the sink simultaneously performed random walk. This scheme enhances the strength of privacy protection but also increases the transmission delay. Wang et al. [17] introduced the concept of attacker listening range and developed the phantom routing with locational angle (PRLA). In their work, random walks are chosen according to the random angle. PRLA can improve the security period compared with the single path phantom protocol, but more computations are needed to recognize the geographic location information of nodes. To improve the phantom routing, Li and Ren proposed Routing to a Random Intermediate Node (RRIN) [18]. A minimum distance from the intermediate node, i.e., phantom node, to the source node holds so that it is difficult for the attacker to locate the real source. Similar to RRIN, the Sink Toroidal Region (STaR) [19] area is designed to ensure the proper location of the intermediate node.
Wang et al. [20] proposed a multipath routing named Weighted Random Stride (WRS) adopting two parameters, randomly forwarding angle and stride value, to build a multipath route. Although using multipath routing can confuse the adversary and make it take longer time to track the source node, the energy consumption also increases. To reduce the energy consumption in the multipath scheme, a constrained random routing (CRR) scheme was proposed [21]. The CRR technique uses the changed next-hop node to build the multiple path route. The selection domain is also designed for the selection of the forwarding node. The selection weight is calculated. The next-hop nodes are selected according to the weight.
Chen and Lou [22] proposed a forward random walk (FRW) scheme. According to hop counts of candidate nodes to sink, the next-hop node in the FRW scheme is selected. The nodes used to forward packets are randomly selected from the closer set. The FRW scheme forms a transmission path between the sink and the source, resulting in insufficient privacy protection. To hide the source by using seek strategy, Long et al. [12] proposed a tree-based (TB) routing, where a backbone route between the source and the sink was established to forward the real message. On the other hand, the diversionary route can send fake messages in the nonhotspot region to minimize the energy of the hotspot area. The bidirectional tree (BT) privacy protection scheme was presented [22]. Some branches are established in the route between the real source and the sink node. In the BT scheme, the real data arrived at the sink node along the shortest path so that the data transmission delay is reduced. With the increase of branches, the security period is greatly enhanced while the tree-shape route leads the additional energy overhead. Moreover, the staircase and branching scheme was introduced to prevent backtracking along the routing paths [10].
The stochastic and diffusive routing using multiple virtual source (SDR-m) technology was proposed [23]. Similar to the TB method, the SDR-m scheme also exploits the excess energy in the nonhotspot area. It is a two-phase scheme. The packet of the source node first reaches the virtual source through dispersive routes. The concept of escape-angle, sequitur-angle is then introduced. In order to avoid a decrease in the security period due to the expansion path loop, a path extension method (PEM) was proposed, in which the fake source is dynamically generated and the principle of the extended path combination is specified [24]. A method of mixing the permanent and temporary fake sources was proposed [25]. The selection of fake sources is designed as a scheduling problem. By dispatching the temporary and permanent fake sources to send packets, the security of the source location is improved, but the energy consumption is still high. Zhou and Wen [26] came up with an energy-efficient scheme by applying ant colony optimization, in which the information of distance, pheromones, and remaining energy are used to figure out the next-hop node.
Different from the above literatures, a three-stage source privacy preservation using a random routing scheme (SLP-R) was proposed [27], which does not use any fake source or multiple paths. The reverse random walk is moving away from the source node. The same depth route increases the path angle between the sink node and the source node. The minimum hop route ensures fast convergence and low latency. This scheme provides better privacy protection but increases the path length and delay of packet transmission. Han et al. [28] proposed the DRBR algorithm. At the mixed link point, the probability is used to generate branches, and the attacker is induced to arrive at the fake source to improve the privacy protection intensity. Meanwhile, the transmission of fake data packets consumes extra energy. Wang et al. [29] proposed the SPAC scheme. The original packet is mapped into a set of shares and transmitted with minimal energy consumption. The fake source sends fake packets to the sink node at the edge of the anonymous cloud to confuse the attacker. The sink node can recover the original message after receiving at least shares. This scheme protects the privacy of source location, improves the confidentiality of data, but reduces the accuracy of data.
3. System Model
The network structure and the properties of application scenario are abstracted, and the corresponding adversary characteristics and attack methods are also given in this section.
3.1. Network Model
The network has sensor nodes and a base station (BS). When the sensor node monitors an object, the collected data are forwarded to the base station in a hop-by-hop manner. In this scenario, the sensor node becomes the source. However, if a monitored object cannot be detected for some time, the source node would stop sending packets. Some settings about the network model are given. (1)The sensor nodes are uniformly distributed. All the nodes have the same configuration including calculation ability, communication radius, storage, and initial energy(2)The packet is encrypted. The attacker cannot directly decipher the data packet to distinguish the real data packet and fake data packet(3)The deep configuration information is initiated by the BS. Deep configuration information is broadcast. After the deep configuration phase, all nodes finally get the hop value to the BS(4)Nodes can communicate with each other if their distance is less than the communication radius. They can obtain the information of neighbor nodes, i.e., the hops, ID of neighbor node, or other interesting information(5)The base station is secure enough that the attackers cannot get any information from it
3.2. Adversary Model
Getting the source location is of great value to the attacker. The attackers trace the source location in terms of temporal correlation of transmitted data and traffic analysis among nodes. The attacker has the following settings. (1)The attacker has complicated listening equipment, enough energy, enough storage, and strong computational power(2)The attacker monitors the local communication situation. All the packets in the listening range will be found. But the attacker cannot listen to the packets out of the eavesdropping range. Attackers use a combination of eavesdropping and hop-by-hop tracking(3)The attacker performs passive attack. It means that the attacker will not modify routing paths and data packets or destroy sensor nodes, etc., assuming that the attacker cannot decrypt the encrypted information. Therefore, the attacker intercepts and analyzes the data traffic to obtain useful information(4)The attacker initially listens to the network near the base station until it overheard a packet sent by some nodes. The attacker analyzes traffic to determine the last node sending information. If there is no wireless signal monitored in the specified time during the tracking, it is considered that the tracking has failed, and the attacker quickly go back to the base station to start listening again(5)The movement speed of the attacker is lower than the speed of packet transmission. Therefore, the attacker moves a node distance each time
4. The Proposed EBBT
This section describes the details of the proposed EBBT scheme. As shown in Figure 1, the EBBT scheme includes three phases: the random walk (RW), the minimum hop routing (MHR), and the creation of branch paths. In the RW phase, the real data packet of the source node is firstly forwarded hops to reach a certain intermediate node . Then, the real packet is transmitted from to BS along the MHR path. The node that generates the branch path on the MHR path is selected with a certain probability . The node at the end of (green node in Figure 1) will play the role of the fake source. The fake source generates fake packets. The transmission path of fake packets consists of . The transmission paths of the real and fake data form a tree-shaped routing structure rooted at BS.
4.1. Random Walk
The node knows the remaining energy and the hop value of its neighbor nodes. Neighbor nodes are divided into three sets ,, and according to the hop value. ,, and are represented, respectively, as the set of neighbors with equal hop, smaller hop, and larger hop than node has. In the RW phase, and constitute the forwarding candidate set of . Then, selects randomly the nodes with the most remaining energy from the candidate set as the next hop. represents the communication radius of the sensor node and defines the maximum hop of the network as
The number of forwarding is controlled by the parameter . The value of is defined by
For a packet is forwarded, is reduced by 1. If is reduced to zero, the RW phase terminates. The last node is the intermediate node. The nodes that existed in the RW path are deleted from the candidate set to avoid the case that the forwarding path forms a loop. The construction of the RW path and building the candidate set are described in Algorithm 1.
| ||||||||||||||||||||||||||||||||||
Algorithm 1: RW Scheme (Node ) | ||||||||||||||||||||||||||||||||||
4.2. Minimum Hop Routing
A minimum hop routing (MHR) will be built from to the BS. In the construction process of MHR, the candidate set of the next-hop node is . There will be a situation that the RW path crosses the MHR path. As shown in Figure 1, selects as the next-hop node from the set and in the RW phase. The set of contains and , and may appear in both RW and MHR paths. If is the next-hop node of in MHR, the RW path intersects with the MHR path. Therefore, in order to avoid the intersection of the two paths, the range of the node is defined. As shown in Figure 2, the range [] represent the scope of the node in the RW path.
The visible area is introduced [17] to improve the security of the source location. A certain range around the source node is defined as the visible area. The circle around the source node represents this area in Figure 2. The tangential range of the nodes in the visible area and is represented by the range .
The geographic region of the nodes of the RW phase and within the visible range of the source is represented by the angle , which satisfies the relation as follows:
The range of candidate node is beyond . In addition to the geographical distribution of nodes, we also need to consider the remaining energy of the node. Algorithm 2 gives the construction process of MHR and the selection method of the initial node of branches.
| ||||||||||||||||||||||||||||||||||||||||||||||
Algorithm 2: MHR Scheme (Node ) | ||||||||||||||||||||||||||||||||||||||||||||||
4.3. MHR with Multibranch
The node on the MHR path becomes the initial node of , , with the probability . Here, the parameter controls the number of branch paths. But a branch path generated at is not controlled by probability . As shown in Figure 1, the branch path is generated at , i.e., is the initial node in . The initial node in selects the next-hop node among its candidate set with most energy. The construction process of the branch path is shown in Algorithm 3. The node at the end of the branch path is the fake source, and consequently, the node set forms the branch path . represents the length of this branch path.
In the process of constructing paths, in order to obtain good path dispersion, the following mechanisms are specified for the selection of the next-hop node. As shown in Figure 3, the -axis of the coordinate system is the connecting line from the intermediate node to the sink node. The forwarding candidate set of is composed of nodes in and with the angle in the range. The direction range of a generated branch path is defined as where represents the angular deviation. is used to control the scattered path. If is in the range , is the candidate node. If there is no matching node, the range of angle will increase by . The node with the most remaining energy in the candidate sets becomes a new branch node. This process will repeat until the value of reduces to 0. The node at the end of the branch path is the fake source. The decentralized branch paths make the energy consumption more balanced and avoid to pull the attacker back to the real path. The construction of branch path is shown in Algorithm 4.
| ||||||||||||||||||||||||||||||
Algorithm 3: Scheme (Node , Hop ) | ||||||||||||||||||||||||||||||
| ||||||||||||||||||||||||||||||
Algorithm 4: Bran Scheme (Node , Hop ) | ||||||||||||||||||||||||||||||
4.4. Analysis of the Number and Length of Branch Paths
In the EBBT scheme, the energy consumed by the branch path is the additional energy overhead. Therefore, the number and length of branch paths affect the security period and the network lifetime.
The nodes on the MHR path generates a branch path with probability . Then, the length of the MHR path and probability have influence on the number of branch paths. The length of the MHR path is defined as the hops between the and the BS, i.e., . The probability is dynamically changed and is controlled using a threshold value . Otherwise, the branch path should be far from the BS because the traffic load of the node near the BS is heavy. Therefore, the relationship among , and is defined as
is the hop value between the BS and .The maximum value of is . The maximum value of is . Equation (5) is simplified as follows:
It can be concluded from Equation (6) that is related to, , and ω. The larger the value, the larger the probability is. The node distribution exhibits a decreasing trend from to the BS according to . The relationship between the number of branches and the hops from to the BS under different is shown in Figure 4.
Choosing an appropriate branch length is important because it affects energy consumption and security period. The length of the path is set to be the same as the length of the RW phase, that is,. represents the length of each , where is related to and . is consistent with the change of . The maximum length of the branch path is . The definition of is defined as where is the adjustment parameter of the path length.
5. Performance Analysis
The average delay of data transmission, the energy consumption, and the security performance of this scheme are analyzed in this section.
5.1. The Average Hops
The forwarding path of real data in the EBBT scheme consists of two phases: RW and MHR. The average hops for MHR are shown as
As shown in Figure 1, is the hops between and the BS. is the hops between the source and the BS. The maximum hop is given in (1). According to the constraint of the random walk phase, the hop range of is between and . The candidate region of is in the area between the orange line and the edge of the network shown in Figure 1. The hops in the random walk phase are controlled by . Therefore, the total delay for real data forwarding is described as
5.2. Energy Consumption
The main energy consumption of the EBBT scheme is to receive and forward real and fake data. Next, we analyze the energy overhead. Assuming that the frequency of source reporting event is equal to the frequency that each of the fake sources generates a fake data packet. When there is no real data, the fake sources stop generating the fake data packet. The path for transmitting a real packet includes RW and MHR. The RW path has hops, the energy consumed by the single-hop propagation packet is set to be , and the energy consumption is . is the length of the MHR path. Then, the energy consumption of the MHR path is . Therefore, the energy consumption for transmitting a real data packet is expressed as
When the source node sends real data packets to the base station, fake packets are generated together. The fake packets are forwarded along . The length of is calculated by Equation (7). Therefore, the energy consumption of the fake packets is represented as
The total energy consumption is calculated by
The energy consumption of the node is mainly divided into two parts: forwarding data packets coming from other nodes and forwarding data packets from to BS. The energy consumption while forwards the data packet in the random walk phase is . The hop count between the intermediate node and is, and the energy consumption is . The energy consumption while forwards the packet to the BS is . only receives the fake packet coming from but not forwards it. The energy consumption of is measured by the packet transmission on the average hop count. Therefore, the total energy consumption of is represented as
The probability of repeated usage is reduced because the residual energy of the node is considered when selecting the next hop. And the energy of the node is reduced after a round of data transmission. In the next round, the node has no advantage of becoming the next hop. For example, in the candidate set of the node , is selected due to its most remaining energy. So, the remaining energy of is reduced. Compared with other neighbor nodes, it would not be the next-hop node in the next round. At the same time, the next-hop node is also far away from the base station. Therefore, the branch path will extend to the edge of the network. In summary, the dispersed paths and the factor of the node remaining energy make the energy consumption to be uniform and prolong the lifetime of the network.
5.3. Security Period
In this section, we give the security analysis of the EBBT scheme. Assume that the monitored objects are randomly strolling in the protected area. The monitored object moves to another location after a period. Define the dwell time . The source node sends the information to the BS if it senses the object. The information is sent at a rate of messages per second. The security period refers to the number of data packet sent by the source node before being captured by the attacker. If , it means that the monitored object has left before the attacker arrives. Otherwise, the attacker may track back to the source node if the routing path is unique. A routing strategy with multibranch paths could cause an attacker to deviate from the real source and increase the network security period.
The total probability that the attacker chooses the branch path and the attacker chooses the correct path is 1. Since the attacker goes back to the base station, the probability that the attacker selects the true path or selects the branch path is all 1/2. The probability that an attacker always chooses the correct path is where represents the number of branches. generates a branch path. The security period becomes long because the attacker has a higher probability of selecting the branch path and is trapped near the fake source node.
The number and length of branch paths in the EBBT scheme have a significant impact on the security period. As mentioned in Section 4, the length of the is described as . The length of path is . Therefore, the average of is defined as
When the attacker is trapped in the branch path, it takes the attacker 2 in the branch path and 2 in the MHR path. According to formula (14), the real path is one of paths. Therefore, the safety period is defined theoretically as
The safety period is represented by the backtracking distance of the attacker. If the movement rate of the attacker is , then the security period is defined as
When the source node is close to the BS, the network security period will reduce. The hop value between and the BS can be increased by increasing the value of in order to increase the security period. In addition, the threshold can be increased. In the EBBT scheme, the branch path, the fake sources, and the intermediate nodes are dynamically changed. The shortest path that is from the intermediate node to the base station is also generated dynamically. Therefore, there are enough paths to confuse the attacker. So, the source location is fully protected.
6. Simulation Experiment and Performance Evaluation
The proposed EBBT scheme in this paper is based on random walk and the routing of multibranch tree structure. Therefore, in the simulation, we use the schemes SLP-R [27], BT [22], and FRW [22] for comparison. The simulation experiment is performed on the MatlabR2014a simulation platform. Experimental environment settings are as follows: network, 11000 sensor nodes are randomly and uniformly deployed. The purpose of this setting is to make the average number of neighbors per node up to about 11. The communication radius of the sensor node and attacker is all 25 m. The node power and the communication radius could be adjusted when existing no suitable neighbor node, assuming that the parameters of EBBT are , , , and =1. The settings of related parameters are verified by the following simulation experiments. Under these settings, the system obtains a better network lifetime and security period. The initial energy of the node is 3 J. The data transmission energy consumption model used in the simulation experiment is referenced to the model in [30]. The length of the packet is 4000 bits. The simulation results given here are the average results of 25 trials, and the source node sends 500 packets in each round.
6.1. EBBT Algorithm Parameter Evaluation
The influence of parameter on the network lifetime is shown in Figure 5, and the influence of parameter on the security period is shown in Figure 6. The coordinate of the source node is (700, 650). As the value increases, the lifetime of network decreases slightly while the security period increases. Since is the threshold value of the control probability , we can infer that increases with the value of increases. Probability controls the number of branches. The larger the value of is, the more branches are created. In the analysis of Section 5, the more the number of branches is, the longer the security period has gained, which is consistent with the experimental results.
Figure 7 shows the effect of parameter on the total number of nodes in branches. Since is equal to , the length of and is positively correlated according to Equation (7). It indicates that there are more nodes in branches with the increment of . It is consistent with the overall trend shown in Figure 7. The number of nodes is about twice times when the distance of the source to BS changes from 500 m to 1000 m. At the same time, with the increase of the distance from the source node to the base station, the number of nodes involved in the branch paths also increases. Therefore, selecting the appropriate number and length of branch path must be compromised between the network lifetime and security period. Figure 8 shows the effect of the lengths of the RW phase on the network security period. The larger the value is, the longer the security period can be obtained.
6.2. Security Period
The simulation comparison of the source location privacy protection security period of four schemes FRW, SLP-R, BT, and EBBT are shown in Figure 9. The source node location in this scenario is between 500 m and 1000 m. The ordinate of the source node is 650 in -axis. The security period tends to increase as the distance between the source node and the BS increases. In the FRW and SLP-R, there is only one path between the BS and the source. Therefore, the network has a shorter security period. An attacker could trace back to the source by a hop-by-hop manner. The security period of the BT method is higher than FRW and SLP-R due to the increasing branch path, but lower than the EBBT method. Because the node in the branch path may have a looped routing in the BT scheme, the attacker can directly skip the loop to reach the upstream node. The attacker’s backtracking time is reduced, so the security period is lower than EBBT.
6.3. Network Lifetime
Figure 10 shows the effect of different source-to-BS distances on network lifetime. In the four schemes, the network lifetime of EBBT is the longest. The parameter in the EBBT controls the direction of the branch path such that the branch paths are relatively dispersed. When selecting the nodes in the path, the node with the most remaining energy becomes the next-hop node. Therefore, since the remaining energy of the node is reduced in the next selection, there will be no advantage of becoming the most residual energy of the node. This guarantees the participation of all nodes. The lifetime of the network is increased by dispersing branch path load traffic to more sensor nodes.
BT and FRW have no obvious change of the network lifetime. It is because the routing path of the BT and FRW is limited in a narrow scope. Therefore, nodes may be reused. Heavy-duty nodes consume more energy, and they will become the bottleneck of the network lifetime. The number of nodes participating in the routing in the SLP-R scheme is lower than in the BT. And the SLP-R scheme selects the next-hop node in the overall network randomly. So, the lifetime is longer than FRW’s and BT’s. Although the number of nodes involved in transmission is increased in the EBBT scheme compared with the SLP-R scheme, the balanced consumption of communication energy in the network plays a positive role in prolonging the network lifetime.
6.4. Path Diversity
The path diversity is evaluated by entropy. Entropy, a mathematical tool, is widely used as an indicator to measure the uncertainty of information. The entropy is defined as
is referenced to the model in [27]. The higher value of implies that the degree of the path diversity is better.
Figure 11 compares the path diversity of the four schemes using entropy. The source node sends 100 packets to the BS and computes to measure the degree of the path diversity. As shown in Figure 11, the path diversity of the EBBT scheme is the highest. It indicates that there are more nodes involved in the path from the source to the BS in the EBBT scheme. The average load on each sensor node is low. Thereby, the network lifetime is improved.
6.5. Average Delay
The average delay refers to the time that packets are transmitted from the source node to BS. In the EBBT scheme, this time is described by the number of hops that data packets are forwarding on the route. Figure 12 shows the average delay of the four schemes. The next-hop node in the FRW is close to the base station. The data packets in BT are sent to the BS through the shortest path, so the average delay in BT and FRW is the least. The SLP-R technology has the largest delay, and the SLP-R algorithm used a three-phase routing path; its routing path is long, and the delay increases. The routing path of the real data in SLP-R is longer than BT and EBBT. The delay of the EBBT scheme is mainly originated by random walk. The packets are sent to a certain intermediate node firstly, which increases the delay. The EBBT solution improves the degree of privacy protection at the cost of relatively less delay.
7. Conclusions
In this paper, an energy-efficient source location privacy protection scheme (EBBT) was proposed which is mainly based on random walk and fake sources. In EBBT, multiple branches are used to transmit fake data streams. The real data flow and the fake traffic flow are mixed, together with the sources forming a tree-like routing structure. With the dynamic change of intermediate nodes, the whole tree structure also changes dynamically. The branch path control angle and the remaining energy of nodes are fully considered which greatly improve the utilization of energy. Theoretical analysis and the results of simulation experiment show that the proposed scheme has a higher network security period and network lifetime than the related scheme, and it has high path diversity. Our future work is to design the data privacy protection method while protecting the location of the source node.
Data Availability
The data used to support the findings of this study are available from the corresponding author upon request.
Conflicts of Interest
The authors declare that there are no conflicts of interest regarding the publication of this paper.
Acknowledgments
The work is supported by the National Natural Science Foundation of China (61862011), the Natural Science Foundation of Guangxi (2019GXNSFGA245004 and 2019GXNSFAA245053), the Science and Technology Major Project of Guangxi (AA19254016 and AA18118025), and the Project of Guangxi Key Laboratory of Trusted Software (KX202056).