A Survey on Fault Tolerance Techniques in Wireless ... - IEEE Xplore

A Survey on Fault Tolerance Techniques in Wireless Sensor Networks GHOLAMREZA KAKAMANSHADI SAVITA GUPTA Dept. of Computer Science & Eng. Dept. of Computer Science & Eng. UIET, Panjab University UIET, Panjab University Chandigarh, India Chandigarh, India [email protected] [email protected]

addition, faulty sensor nodes cannot perform any monitoring task properly. Therefore, these faulty SNs should be recognized at the right time and removed from the data collection process to guarantee the overall data quality. In large-scale WSNs, it is not possible for the BS to gather data from each SN and detect faulty SNs in a centralized manner. Therefore, localized and distributed SN fault detection algorithms are highly preferred [7]. In a WSN, it is significant to have continuous connectivity, when the SNs are being deployed in a particular area. However, WSNs are restricted by the limited power, bandwidth, networks without fixed infrastructure and different types of problems such as node failure, link or path failure, vulnerability to attack, etc. To address these problems, there is a need for WSNs to be self-configuring and self-organizing so as to enhance performance, increase power efficiency, reduce data transmission and save network resources [8]. The rest of this paper is organized as follows: Section II is about Fault Tolerance (FT) in WSN. Section III presents a review of existing techniques for FT in WSN. Finally, we conclude the paper and give future directions in section IV.

Abstract- Fault tolerance is one of the critical issues in Wireless Sensor Network (WSN) applications. The problem of missing sensor node, communication link and data are inevitable in wireless sensor networks. WSNs experience failure problems due to various factors such as power depletion, environmental impact, radio interference, asymmetric communication links, dislocation of sensor node and collision. Many researchers have proposed fault tolerant mechanisms that are able to achieve higher data reliability, accuracy, energy saving, enhance network lifetime and minimize failure of components of wireless sensor network. This paper presents a critical analysis of various fault tolerance mechanisms in wireless sensor networks such as redundancy based mechanisms, clustering based mechanisms and deployment based mechanisms to identify the strengths and weaknesses of each one of these mechanisms. Finally, the paper presents conclusion and suggests some future research directions that will be helpful for researchers who are working in this field. Keywords: Wireless Sensor Networks; Fault Tolerance; Cluster Head Failure; Node Failure; Link Failure

I.

INTRODUCTION

Recently, one of the most active research areas in the field of networking is Wireless Sensor Network (WSN). There are numerous applications, where sensor nodes (SNs) are able to sense the physical world to perform data collection and task monitoring. A WSN is a self-organized network and it consists of a collection of tiny, low powered SNs with limited transmission range and sink node [1]. A WSN can be classified into two types: homogeneous WSN and heterogeneous WSN. In homogeneous wireless sensor networks, all of the devices possess the same communication range and computing capability. However, in heterogeneous wireless sensor network, all of the SNs have distinct capabilities like distinct processing and computational power, different communication ranges and sensor types [2]. Sensor nodes collaborate with each other to accomplish data sensing, data processing and data communication via either single-hop or multi-hop wireless links [3]. Sensor nodes can be deployed on the ground, in the air, in vehicles, under water, on bodies, and inside buildings to achieve monitoring task [4, 5]. Due to their deployment in hostile or harsh environments, unknown area, etc. SNs tend to failure [6]. Faulty SNs are likely to report arbitrary readings that do not send the reality of observed physical process to the Base Station (BS). In

c 978-1-4673-7910-6/15/$31.00 2015 IEEE

SUKHWINDER SINGH Dept. of Computer Science & Eng. UIET, Panjab University Chandigarh, India [email protected]

II.

FAULT TOLERANCE IN WSN

A fundamental aspect in the design of WSNs is to keep SNs functional as long as possible. Failures are likely to be caused by conditions out of the control of the designers [9]. Failure of SN may be caused by different reasons containing transmission link instability, environmental impact, failure of hardware component, radio interference, dislocation of SNs and battery depletion [10, 11]. In addition to sensor node, BS may also fail due to different reasons such as hardware failure, software failure, intentional attacks [12, 13]. Moreover, in heterogeneous WSN, when different devices are used, like Relay Node (RN) [26] or Super Node [38] can also fails due to hardware failure, software failure, etc. [40]. Hence, it is important that the WSN as a whole should be able to fault tolerate. Therefore, there is a need of fault tolerance system, which exhibits a low probability of false alarms and a high probability of fault identification [14]. Fault tolerance is a critical issue for reliable data delivery in WSN applications. It should ensure that a system is available for use during any interruption or presence of fault. Therefore, fault tolerance enhances the availability, reliability

168

also dependability of the wireless netwoork system [15]. Sensor node failures can be classified intoo two categories; single node failure, which indicates the loss of one node at a time and multi-node failures, which indiccates multi node failures at a time [10]. There are many strategies to achieve FT in WSNs, like Redundancy based, Deployment based, Clusstering based, etc. Classification of FT mechanisms is shown inn Figure 1.

Node Reddundancy Redundancy based

Path Reddundancy Data Reddundancy Time Reddundancy

Fault Tolerant Mechanisms

Clustering based

Singlee hop Multipple hop Hierarrchical

Deployment based

Topolpgyy control Topologgy design

Fig. 1: Fault tolerance mechanisms taxoonomy

III.

T IN WSN EXISTING TECHNIQUES FOR FT

Many mechanisms have been proposed for f FT in WSN to achieve reliability assurance, enhancing lifetime, energy saving and so on. All these methods can be b classified into three categories; Redundancy based mechannisms, Clustering based mechanisms and Deployment based mechanisms. m

A. Redundancy based mechanisms and Faullt tolerance Many researchers have investigated variious mechanisms based on redundancy in WSNs, such as pathh redundancy [17, 20], time redundancy [18] or temporal redunndancy [16], data redundancy [18], node redundancy or physical redundancy [21]. These techniques are used to achieve energy e efficiency, ensure the reliability, security and FT in WS SNs. For instance, in dynamic sensing environments, time reddundancy is used whenever parameters like weather conditionns quickly change in time [16]. Furthermore, in this case, thee computation or data sending is repeated and the result is coompared with the previous output in order to increase reliability [9]. Liang in [17] presented an Energy Eff fficient Multipath routing and Fault-tolerant (FEEM) method. FEEM is a cross layer approach for WSNs. In addition, autthor presented an Energy and Mobility-aware Geographical Multipath M Routing (EM-GMR) scheme for multi path selectioon. The mobility, remaining energy and distance to the target node n of candidate SNs in the local communication range were used u for next hop RN election and a fuzzy logic approach was used for decision making. Results demonstrated that FEEM M could tolerate

some link failures and saves energy. The author concluded that most energy consumptionn was related to communication of SNs and the multiple pathhs in FEEM could achieve the higher performance in terms of o link failure. Lee and Choi in [18] proposed a distributed fault detection algorithm where thhe faulty SNs were detected by comparing the neighboring SN Ns and the decision was made at every SN in the network. To T tolerate transient failures in communication link and sennsing process, time redundancy was used. In order to rem move delay involved in time redundancy system, a slidingg window also employed with small memory for storage of earlier compared results. Simulation results revealed thhat for a large spectrum of fault rates, SNs with permanent faults f were detected with high accuracy, while most of the transient faults were tolerated with inconsequential degradattion of performance. Qiu et al. in [19] propoosed a new energy-aware FT method for WSN, named as Innformer Homed Routing (IHR). In IHR method, the collectoor node just sent data to the backup Cluster Head (CH) when it found that main CH failed, instead of transmittinng data to the main CH and backup CH at the same time.. The effectiveness of IHR was compared with Dual Homeed Routing (DHR) and LowEnergy Adaptive Clustering Hierarchy (LEACH) protocols. In addition, authors have conssidered power consumption and the number of dead SNs. Resuults revealed that the introduced protocol could significantly reduce r power consumption and decrease data loss rate. Halder et al. in [20] intrroduced a Fault Tolerant Load Balancing Scheme (FTLBS) to improve FT and lifetime of sensor network. The proposeed scheme organized the entire sensor network into groups and levels. A multipath data transmission technique wass devised for FT and the transmission load was balancced by varying group size. The method dynamically selects a route based on fitness of the nodes. The proposed approachh delivered data efficiently with minimum delay, even in faultyy network. The connectivity and coveerage are two important issues in WSN. The connectivity is determined d as the capability of each SNs to reach the BS S through a multi-hop route. However, the coverage is deffined as ensuring that how well the network monitors the areaa of interest [39]. As to address these issues with FT aspect, Korbi et al. in [21] proposed a new approach to ensure FT inn WSNs. The proposed method was used to maintain coveerage and connectivity in the network. In the case of node failure, the “up to fail” node is considered and replace beffore failure on the network. However, in the case where itt is impossible to replace of the “up to fail” node, a fast reroutting mechanism has proposed to forward the traffic initially roouted via the “up to fail” node. Performance evaluation of prresented approach revealed that the number of nodes potentiaally eligible for the “up to fail” node replacement depend onn parameters such as the node redundancy level threshold and the network density. In addition, authors compared proposed routing algorithm with a classical routing algorithm and proved that the proposed algorithm reduce the rate of paacket loss in the network.

2015 International Conference on Green Computing and Internet of Things (ICGCIoT)

169

We briefly summarize the existing approaches of fault tolerant redundancy based for WSN in Table 1. TABLE 1. Fault tolerant Redundancy based Mechanisms

2

Distributed fault detection method [18]

3

IHR [19]

Path

FEEM, EM-GMR [17]

* * *

4

FTLBS [20]

5

Coverage and Connectivity based fault tolerance [21]

Time

1

Data

Mechanism(s)

Redundancy types Node

S. No.

* *

Major features achieved 1) Energy efficiency 2)Tolerant link failure 1)High fault detection accuracy 1)Energy efficiency 2)Data loss rate 1)Data accuracy with minimum delay 1)Packet loss rate

B. Clustering based mechanisms and Fault tolerance One of the efficient mechanisms for improving FT in WSNs is clustering mechanism. Scalable control of WSNs can be done by clustering, uses of clustering mechanism effect on energy saving and control over the network. This type of mechanism is useful to achieve local communication in which each CH receives data from cluster members and sends aggregated data to the BS named as single hop communication. On the other hand, data gathered can be transferred to the BS via multiple hop communications. Failure of CH is inevitable in WSNs; therefore, that is a need for adaptive fault tolerant clustering mechanism. Lai and Chen in [22] proposed a distributed fault-tolerant method named CMATO (Cluster-Member-based fAultTOlerant) for WSNs. It viewed the cluster as a discrete whole and utilized the monitoring of cluster members to find and recover from the failed node in a fast and energy efficient way. Due to the incorporation of the local information of the network, proposed method is flexible and can be combined with different available clustering methods in WSNs. Moreover, CMATO was capable to tackle multiple cluster heads failures, so CMATO impressively recovered the sensor nodes from multiple CHs failure and failure of communication links inside the cluster, obtaining faulttolerant networks. Simulation results demonstrated that proposed method outperformed the earlier fault-tolerant methods in terms of power consumption and fault coverage. Bansal et al. in [23] presented a Fault Tolerant Election Protocol (FTEP) that is distributed and dynamic new CH selection with FT capabilities. The proposed method was based on two level clustering schemes. In this protocol, if a CH fails or energy level of CH decreases below threshold, then selection process is started based on power levels. To handle CH failure, selection process chooses a CH node and a back-up node. Back-up node automatically performs the duty of CH once the failure of present CH is identified. Simulation results showed significant power consumption when compared with existing clustering methods.

170

Kaur and Sharma in [24] proposed Fault Tolerant Twolevel Clustering Protocol (FTTCP) for WSN. FTTCP periodically checked CH for failure. The aims of proposed method are to accurately identify CH failure and prevent unwanted power consumption caused by an incorrect identification process. In this method, each cluster member could identify its CH failure separately. To reduce the power consumption, distributed identification process was used at every cluster member; it made use of heartbeat messages sent in a repeated manner by a CH for fault identification. To recover from faulty CH, back up node was selected as new CH and new back up node was selected locally. Selection of CH and back up node were based on remaining power of SNs. Simulation results revealed the proposed protocol provided high fault identification accuracy in harsh area and consumed little bit more energy than FTEP. Karim et al. in [25] presented a power efficient and Fault Tolerant Dynamic Static Clustering (FT-DSC) algorithms for WSN. The main objective of proposed protocol was to increase the performance of the DSC protocol by introducing FT mechanism. The performance of the presented protocol has been tested and compared with the DSC protocol. Simulation results showed that the FT-DSC protocol yield higher performance as compared with the DSC protocol in terms of reliability and power efficiency. Bari et al. in [26] proposed a new integrated Integer Linear Program (ILP); the presented method incorporate fault tolerance by properly selecting the minimum number of RN locations in the given network also assigned the SNs to the clusters and determined a load-balanced routing scheme. At the upper and lower ties of the network, the desired levels of fault tolerance for RNs and SNs were specified by two parameters. The developer determined actual values of these parameters and values were given as inputs to the ILP. The presented method met indicated performance guarantees with regard to lifetime of the network by limiting the maximum power consumption of the RNs. R. Kumar and U. Kumar in [27] presented a novel flexible, hierarchical clustering method for homogeneous WSN. The aim of their method was to address five issues; energy efficiency, scalability, FT, multi-level clustering and load balancing. In their method, Function Delegation protocol was used to provide FT. The evaluation of the presented method was done from different angles. Results revealed performance and power efficiency of proposed algorithm were promising, particularly in bridge topologies. Chang and Huang in [28] developed a FT method with the cluster-based structure and introduced a novel fault model based on behavior of SNs in heterogeneous WSNs. The proposed model detected the probably faulty SNs. This method also adapted the Bayesian method to monitor and estimate SNs reputation respecting message forwarding. Implementation of the proposed FT method can be done to analyze its performance in terms of data accuracy and reliability in WSNs. Brust et al. in [29] proposed the clustering coefficient as a local factor for FT in WSNs and described how to enhance


the clustering coefficient in a WSN topology by exclusively adding and removing communication links. Changes in the topology could lead the network to failures. The clustering coefficient metrics were analyzed to measure the degree of FT. The results indicated that the clustering coefficient in the WSN is highly correlated to the FT. The proposed method can be extended to analyze the additional clustering metrics to infer their correlation with FT. Rong Duh et al. in [30] presented a distributed fault tolerant algorithm to detect event region in WSNs. The design was based on fault-event disambiguation problem. The proposed strategy could identify the faulty and fault free SNs. In the proposed algorithm, CHs sent event report to the BS. The base station identified the abnormal region and sent out false alarm. Simulation results revealed that event detection accuracy was greater than 99%, fault detection accuracy was greater than 92% and false alarm rate was near 0% under uniform distribution. Furthermore, under random distribution; event detection accuracy was greater than 88%, fault detection accuracy was greater than 92% and false alarm rate was less than 1.2% when sensor fault probability was less than 0.3. Table 2 shows summarization of the existing approaches of fault tolerant clustering based for wireless sensor network. TABLE 2. Fault tolerant Clustering based Mechanisms Clustering types

1

CMATO [22]

*

2

FTEP [23]

*

3

FTTCP [24]

4

FT-DSC [25]

5

ILP [26]

6 7 8

9

Hierarchical

Multiple hop

Mechanism(s)

Single hop

S. No.

* * *

Novel flexible, hierarchical clustering method[27] Fuzzy knowledge based Fault tolerance [28] Clustering coefficient and Fault tolerance [29] Distributed fault tolerant algorithm[30]

* * *

*

Major features achieved

1)Energy efficiency 2)Fault coverage 1-Energy efficiency 1)High fault detection accuracy 1)Energy efficiency 2)Reliability 1)Energy efficiency 2)lifetime 1)Energy efficiency 1)Reliability 1)Degree of fault tolerance 1)Event detection accuracy 2)Fault detection accuracy 3)False alarm rate

C. Deployment based mechanisms and Fault tolerance Careful deployment of SNs in WSNs can lead to effective design goals. There are different strategies of deployment of SNs, which are; pre deployment of SNs that is called design of the network, during deployment of SNs and after deployment of SNs. After deployment of nodes, topology control mechanism is needed; network topology may changes

due to dislocation of SNs, sensor nodes failure, or other conditions. Moreover, connectivity of SNs may also change due to noise, interference, etc. Therefore, topology control algorithms are required to increase the network lifetime. Moreover, topology control methods in fact decrease the degree of routing redundancy by decreasing the number of communication links in the WSN. Li and Hou in [31] proposed a topology control algorithm; the main objectives of this algorithm are to keep network connectivity, enhancing power efficiency and increasing capacity of the network. The derived topology was more prone to SN failures. The authors resolved this problem in the topology construction process by enforcing k-vertex connectivity. A localized method named Fault-tolerant Local Spanning Subgraph (FLSS) was proposed, that could preserve k-vertex connectivity and was min-max optimal between other methods. Results revealed that, compared with previous localized and distributed fault-tolerant topology control methods; FLSS yields better energy efficiency as well as higher network capacity. Furthermore, FLSS is robust with respect to position estimation errors. Chen et al. in [32] presented a fault tolerance and energy efficient topology control protocol (P-CDS). The presented mechanism scheduled active sensor nodes and backup sensor nodes in the backbone to tolerate failures. To recover network connectivity, the P-CDS protocol could adapt the radio ranges of coordinators. When SN failure was identified, both active and backup SNs adjusted their transmission ranges to keep network connectivity. Simulation results showed that P-CDS created a smaller connected dominating set than the existing fault tolerant topology control method. The P-CDS method also provided effective broadcast with relatively less power consumption. Sitanayah in [33] designed Emergency Response Medium Access Control (ER-MAC) which is a hybrid protocol for emergency response WSNs. The proposed protocol is able to switch from energy efficient operation in normal periodic monitoring to reliable and fast delivery for emergency monitoring, and vice versa. In addition, the author provided solutions for topology planning problems using redundancy method. Additional Relay Placement, Additional Backup Placement, Multiple Sink Placement, and Multiple Sink and Relay Placement were used to support FT by ensuring that there are alternative acceptable routes to data sinks when failure occurred. The ER-MAC protocol was used to evaluate the FT of each deployment resulted in multi-hop data gathering. In this work, the author mentioned that MAC protocols and topology planning algorithms can be used together to create fault tolerant WSNs for volatile environment. The proposed strategy could only guarantee robustness against single failure. Rong et al. in [34] introduced a novel fault-tolerant method named “Adaptively Fault-tolerant Topology Control (AFTC)” to assign network resources reasonably. The SNs with greater fault-tolerant degree were elected as backbone nodes, and every backbone node had the backup nodes for advance performances of WSNs, containing fault-tolerant


171

S. No.

Deployment types Mechanism(s)

Topology control

172

TABLE 3. Fault tolerant Deployment based Mechanisms

Topology design

capability and power efficiency. Results showed that the method is effective and reliable to fabricate and keep the fault-tolerant topology for WSNs. Rehena et al. in [35] presented a robust recovery mechanism of nodes failure in a certain area of the network during data transmission to achieve FT. This recovery mechanism was concentrated on multiple-sink partitioned network. The proposed technique dynamically discovered new node to route data from source to sink. In addition, authors presented an area handling mechanism. Results demonstrated that the data delivery to sink was still possible, when all nodes in a particular geographical area were dead. In addition, the proposed mechanism considered the active nodes in a specific partition when they are unable to transfer data to the sink. Therefore, the active nodes tried to be attached with other partitions. The outcomes revealed that proposed technique has better tolerability and energy efficiency for area failure. Sun et al. in [36] proposed a trust framework for data aggregation with FT in wireless multimedia sensor networks. In this approach, multilayer data aggregation structure is used which divide the network into different layers or levels. In the proposed framework, nodes are classified as SNs and data aggregators, which they play different roles in the data aggregation process. To increase the correctness and trustworthiness of gathered information jointly considered data aggregation, FT and information trust. The proposed framework could evaluate both continuous media streams and discrete data in a wireless multimedia sensor networks. Results revealed the proposed scheme could meaningfully enhance the quality of multimedia data as well as reliability of gathered data. Geeta et al. in [37] presented an Active node based Fault Tolerance applying Battery power and Interference model (AFTBI) in wireless sensor network. Fault tolerance against low energy was formulated via hand-off mechanism. The faulty SN selected the nearby SN containing the high level energy then transferred completely all services that were to be carried out to the elected nearby SN. Fault tolerance against interference was provided through dynamic energy level adjustment mechanism by assigning the time slot to the nearby SNs. If a candidate SN wanted to send the sensed data, it entered active state and sent the data with maximal energy. Otherwise, it entered into sleep state containing lowest energy that was enough to maintain the connectivity and to receive hello messages. The performance was tested via simulation in terms of data delivery ratio, memory overhead, control overhead and failure recovery delay. The results compared with Fault Detection in Wireless Sensor Networks (FDWSNs) for different performance measures. AFTBI outperformed the results of FDWSN. There is a problem to select node after failure. It only depends on the highest residual power of selected node to transfer all the services. Maybe the node selected is far away. We briefly summarize the existing approaches of fault tolerant deployment based for WSN in Table 3.

Major features achieved

1

Topology control algorithm [31]

*

1)Energy efficiency 2)Higher network capacity 3)Network connectivity

2

FT and energy efficient topology control protocol [32]

*

1)Energy efficiency

3

ER-MAC [33]

4

AFTC [34]

5 6

7

IV.

Robust recovery mechanism[35] Trust framework for data aggregation with FT [36] Active node based Fault Tolerance[37]

*

*

* *

*

1)Reliability and fast delivery 2)Robustness against single failure 1)Energy efficiency 2)Fault tolerant capability 3)Reliability 1)Energy efficiency for area failure 1)Data accuracy of multimedia data 2)Reliability 1)Data delivery ration 2)Memory overhead 3)Control overhead 4)Failure recovery delay

CONCLUSION AND FUTURE DIRECTIONS

This paper presents a review of various fault tolerant algorithms designed for WSN. After critical analysis, it has been observed that the different strategies such as deployment, redundancy, and clustering can be used in different applications with respect to the level of FT requirement. Adding few redundant components can increase level of FT and enhance data accuracy. Efficient clustering can improve energy consumption and increase lifetime of the system. Attention to the deployment phase can drastically save the energy, increase the lifetime and enhance the reliability of the network. Some future research directions are listed below. • Most of the researchers investigated considering only one base station in the network where in large-scale WSNs, we need to consider multiple BSs in order to save energy and to increase fault tolerant against BS failure. Moreover, the deployment of BSs should be done based on some significant parameters such as geographical area of the network, the average ratio of SNs subsequent to failure of BS, which represent the FT of the network, the average delay of the network due to congestion. • Area based failures can be considered, where failed devices are in close proximity to each other. • The presented mechanisms are limited to smaller networks, consideration for large-scale networks is required and distributed approach can be developed. • A WSN can be interrupted by malicious attacks. In order to achieve security goals for WSNs, there is a need to design and develop specific mechanism with respect to FT.


REFERENCES [1]

[2]

[3]

[4]

[5] [6]

[7]

[8]

[9] [10]

[11] [12]

[13]

[14]

[15]

[16]

[17]

[18] [19]

[20]

[21]

F. J. Atero, J. J. Vinagre, J. Ramiro, and M. Wilby, “A low energy and adaptive routing architecture for efficient field monitoring in heterogeneous wireless sensor networks,” In: 31st Int. Conf. on Distributed Computing Systems Workshops, IEEE, pp.172 – 181, 2011. C. Wu and Y. Chung, “Heterogeneous Wireless Sensor Network Deployment and Topology Control Based on Irregular Sensor Model,” Adv. Grid Pervasive Comput. Springer Berlin Heidelberg, pp. 78–88, 2007. G. Anastasi, M. Conti, M. Di Francesco, and A. Passarella, “Energy conservation in wireless sensor networks: A survey,” Ad Hoc Networks, vol. 7, no. 3, pp. 537–568, 2009. S. Climent, J. V. Capella, N. Meratnia, and J. J. Serrano, “Underwater sensor networks: A new energy efficient and robust architecture,” Sensors, vol. 12, no. 1, pp. 704–731, 2012. M. Sudip, S. Chandra Misra, I. Woungang, Guide to Wireless Sensor Networks. London: Springer, 2009. R. Ma, L. Xing, and H. Michel, “Fault-Intrusion Tolerant Techniques in Wireless Sensor Networks,” In: 2nd Int. Symposium on Dependable, Autonomic and Secure Computing, IEEE. pp. 85-94, 2006. C. Zhang, J. Ren, C. Gao, Z. Yan, and L. Li, “Sensor fault detection in wireless sensor networks,” IET Int. Commun. Conf. Wirel. Mob. Comput., pp. 66–69, 2009. B. Khelifa, H. Haffaf, M. Madjid, D. Llewellyn-Jones, “Monitoring Connectivity in Wireless Sensor Networks,” J. of Future Generation Communication and Networking, vol. 2, no. 2, pp. 507 – 512, 2009. E.Dubrova, Fault Tolerant Design: An Introduction. New York: Springer, 2013. M. Younis, I. F. Senturk, K. Akkaya, S. Lee, and F. Senel, “Topology management techniques for tolerating node failures in wireless sensor networks: A survey,” Comput. Networks, vol.58, no.1, pp.254-283, 2014. L. Moreira, H. Vogt, M. Beigl, A survey on fault tolerance in wireless sensor networks. braunschweig, Germany: Sap research, 2007. S. Kim, J. Ko, J. Yoon, and H. Lee, “Multiple-Objective Metric for Placing Multiple Base Stations in Wireless Sensor Networks,” In: 2nd Int. Symposium on Wireless Pervasive Computing, IEEE. pp. 627-631, 2007. R. Raj, M. Ramesh, S. Kumar, “Fault Tolerant Clustering Approaches in Wireless Sensor Network for Landslide Area Monitoring.” In: Int. Conf. on Wireless Networks, pp.107-113, 2008. H. Ammari, S. Das, “Fault tolerance measures for large-scale wireless sensor networks,” ACM Trans. Auton. Adapt. Syst., vol. 4, no. 1, pp. 1– 28, 2009. H. Alwan and A. Agarwal, “A Survey on Fault Tolerant Routing Techniques in Wireless Sensor Networks,” 2009 Third Int. Conf. Sens. Technol. Appl., IEEE, p. 366-371, 2009. D. Ioan Curiac, C. Volosencu, D. Pescaru, L. Jurca, A. Doboli, “Redundancy and its applications in wireless sensor networks: a survey,” WSEAS Trans. on Computers, vol. 8, no.4, pp.705-714, 2009. Q. Liang, “Fault-Tolerant and Energy Efficient Wireless Sensor Networks: A Cross-Layer Approach,” In: IEEE Military Communications Conference, IEEE. pp.1862-1868, 2005. M. Lee, Y, Choi, “Fault detection of wireless sensor networks,” J. of Computer Communications, vol.31, no.14, pp.3469–3475, 2008. M. Qiu, J. Liu, J. Li, Z. Fei, Z. Ming and E. Sha, “Informer homed routing fault tolerance mechanism for wireless sensor networks,” J. of Systems Architecture, vol.59, pp.260-270, 2013. S. Halder, M. Mazumdar, P. Chanak, I. Banerjee, “FTLBS: Fault Tolerant Load Balancing Scheme in Wireless Sensor Network,” J. of Advances in Computing and Inform. Technology, pp.621-631, 2012. I. Korbi, Y. Ghamri Doudane, R. Jazi, L. Azouz Saidane, “CoverageConnectivity based Fault Tolerance Procedure in Wireless Sensor Networks,” In: 9th Int. Conf. on Wireless Communication and Mobile Computing, IEEE. pp. 1540-1545, 2013.

[22] Y. Lai and H. Chen, “Energy-efficient fault-tolerant mechanism for clustered wireless sensor networks,” In: Int. Conf. Comput. Commun. Networks, IEEE, pp. 272–277, 2007. [23] N. Bansal, T. Sharma, M. Misra, R. Joshi, “FTEP: A Fault Tolerant Election Protocol for Multi-level Clustering in Homogeneous Wireless Sensor Networks,” In: 16th Int. Conf. on Networks, IEEE. pp.1-6, 2008. [24] A. Kaur and T. Sharma, “FTTCP: Fault Tolerant Two-level Clustering protocol for WSN,” Int. J. on Networking Security, vol.1, no.3: pp.2833, 2010. [25] L. Karim, N. Nasser, and T. Sheltami, “A Fault Tolerant Dynamic Clustering Protocol of Wireless Sensor Networks,” Glob. Telecommun. Conf., IEEE, pp. 1–6, 2009. [26] A. Bari, A. Jaekel, J. Jiang, and Y. Xu, “Design of fault tolerant wireless sensor networks satisfying survivability and lifetime requirements,” Comput. Commun., vol.35, no.3, pp. 320-333, 2012. [27] R. Kumar and U. Kumar, “A Hierarchal cluster framework for wireless sensor network,” Int. Conf. Adv. Comput. Commun, IEEE. pp. 46-50, 2012. [28] S. H. Chang and T. S. Huang, “A Fuzzy Knowledge Based Fault Tolerance Algorithm in Wireless Sensor Networks,” In: 26th Int. Conf. Adv. Inf. Netw. Appl., pp. 891-896, 2012. [29] M. R. Brust, D. Turgut, C. H. C. Ribeiro, and M. Kaiser, “Is the clustering coefficient a measure for fault tolerance in wireless sensor networks?,” In: IEEE Int. Conf. Commun., pp. 183-187, 2012. [30] D. Rong Duh, S. Pei Li, V. Cheng, “Distributed Fault-Tolerant Event Region Detection of Wireless Sensor Networks,” Int. J. of Distributed Sensor Networks, pp.1-8, 2013. [31] N. Li and J. Hou, “Localized fault-tolerant topology control in wireless ad hoc networks,” IEEE Trans. Parallel Distrib. Syst., vol.17, no.4, pp. 307-320, 2006. [32] C. Chen, K. Feng Ssu and H. Christine Jiau, “Fault Tolerant Topology Control with Adjustable Transmission Ranges in Wireless Sensor Networks,” In: 13th Int. Symposium on Pacific Rim Dependable Computing, IEEE. pp. 131-138, 2007. [33] Sitanayah L. Planning the deployment of fault-tolerant wireless sensor networks. PhD, National University of Ireland, Cork, Ireland, 2013. [34] Y. Rong-rong, L. Bin, L. Ya-qian and H. Xiao-chen, “Adaptively faulttolerant topology control algorithm for wireless sensor networks,” J. of China Universities of Posts and Telecommunications, vol.19, no.2, pp.13-18, 2012. [35] Z. Rehena and S. Roy, “Handling Area Fault in Multiple-Sink Wireless Sensor Networks,” In: IEEE 3rd Int. Conf. on Advance Computing, IEEE. pp. 458-464, 2013. [36] Y. Sun, H. Luo, and S. K. Das, “A trust-based framework for faulttolerant data aggregation in wireless multimedia sensor networks,” IEEE Trans. Dependable Secur. Comput., vol.9, no.6, pp.785-797, 2012. [37] D. D. Geeta, N. Nalini, and R. C. Biradar, “Fault tolerance in wireless sensor network using hand-off and dynamic power adjustment approach,” J. Netw. Comput. Appl., vol.36, no.4, pp.1174-1185, 2013. [38] M. Cardei, S. Yang, and J. Wu, “Fault-Tolerant Topology Control for Heterogeneous Wireless Sensor Networks,” In: IEEE Int. Conf. Mob. Adhoc Sens. Syst., pp. 1–9, 2007. [39] S. Chouikhi, I. El Korbi, Y. Ghamri-Doudane, and L. Azouz Saidane, “A survey on fault tolerance in small and large scale wireless sensor networks,” Comput. Commun., pp. 1–16, 2015. [40] R. H. Abedi, N. Aslam, and S. Ghani, “Fault tolerance analysis of heterogeneous wireless sensor network,” 24th Can. Conf. Electr. Comput. Eng., pp.175-179, 2011.


173