Integrated social network reputation inspired routing for effective data forwarding
© Shobana and Narayanasamy; licensee Springer. 2014
Received: 26 June 2014
Accepted: 19 August 2014
Published: 28 August 2014
Wireless sensor networks face many threats which drain the energy. The performance of sensor network routing is much affected in the presence of selfish nodes with messages being delivered with a longer delay. Social network routing is a method in which the messages are selectively forwarded through the nodes where the encounters between these nodes are more likely to occur. Network reputations clearly speak about the quality of nodes involved in data forwarding. The idea is to utilise social network reputations of source or destinations for effective data forwarding in farmland sensor networks.
In wireless sensor networks (WSNs), nodes do not physically move in many directions. Instead, the nodes are at least temporarily static and are attached to a physical location until the task is done . In case of any malfunctioning only, the sensor nodes disintegrate themselves from the network for a while until the problem is attended. After it is done, the sensor nodes generally reappear again bounded to the same previous physical location. If any such topological changes do exist, it shall be attempted by configuring themselves into the network .
However, the configuration of a new node involves registration in all necessary locations which would be time-consuming. The only solution is to insert the node at a desirable location and then allowing the node to communicate with the network on its own! This communication has to happen casually within fraction of seconds after the new node joins the network. The same is applicable when the old node rejoins the network. These requirements seemed to have been inspired from social network behaviour where there are no constraints or limitations in general on their collaborative or social behaviour except for security constraints. This social behaviour of WSNs shall be modelled via opportunistic networks.
But the main problem which could be seen in this type of routing is that the nodes in the friends list might become malicious or selfish in the future. Hence, it is necessary to monitor the nodes in the friends list for finding the misbehaving nodes.
In this paper, Integrated Social Network Routing (SoNR) protocol is modelled behind the movement of cattle in farmland sensor networks. In Figure 1, S could be seen as a calf and D could be seen as it's mother. S might require the location or movement pattern of S for successive time period so that S could rejoin D for some desired purpose. And it is always easier if such S-D pairs are always together, while harvesting the healthcare data [5–7]. In this context, the proposed SoNR protocol would monitor the friends list for misbehaving nodes and it would maintain the friends list only with good reputed friend nodes. Reputation is calculated by (1) Acknowledgement system and (2) Message delivery-based reputation system.
The friends list is periodically updated based on the calculated reputation values. Thus, the updated friends list consists of only good reputed friend nodes, and the message delivery increases gradually as all the nodes in the friends list helps in data forwarding.
2 Related work
Parris and Henderson  proposed privacy-enhanced social routing. In this work, two schemes namely statisticulated Social Network Routing (SSNR) and obfuscated Social Network Routing (OSNR) were used. In SSNR, the friends list is modified by adding or removing nodes for each message transmission. Hence, it is not easy for a node to identify the original friends list of a sender by just interpreting a single message. In OSNR, the friends list of source node is embedded in a bloom filter.
Lilien et al.  discuss security and privacy challenges in opportunistic networks. The various privacy challenges that need to be considered in oppnet are helper privacy, oppnet privacy, and data privacy. This work discusses the use of trusted devices for more critical tasks, and security routing is enabled by selecting a route that passes through only trusted devices. Alternatively, opportunistic feeding and routing  has also been experimented.
Li et al.  designed a trust-based framework for data forwarding in opportunistic networks. A watchdog component is included in the trust framework to monitor the behaviour of the forwarding node. The Positive Feedback Message (PFM) is generated by the receiving node to the source to inform the behaviour of the forwarding node. Based on the received PFM, trust to the forwarding behaviour of a node is calculated. The forwarding decision of a node is taken based on the trust and the forwarding ability of a node. Trust-based secured routing models [12, 13] concentrating QoS parameters  involving artificial intelligence techniques  are well addressed in the WSN literature.
Novel reputation computation model based on subjective logic is introduced by Liu et al. . This model identifies and prevents selfish behaviours. This model consists of two phases namely reputation query and reputation computation. In the reputation query phase, a node that receives a service request accumulates the recommended opinions on the requester from their common neighbours who have interacted with it. In the reputation computation phase, node evaluates the reputation with the opinions accumulated in the reputation query phase.
Bigwood et al.  proposed a novel incentive mechanism in which self-reported social networks (SRSNs) are used to collect social network data. SRSNs can be obtained from an online social network. These SRSNs are used to provide reputation for nodes. When the network starts up, nodes assign higher trust values to nodes in their SRSN. Selfishness is detected by storing a history of encounter times and exchanging the histories during encounters. Once a node is detected as selfish, the detecting node decrements the value of selfish node by the behaviour constant x.
Packet dropping detection scheme and routing misbehaviour mitigation scheme is introduced by Qinghua et al. . In this work, the misbehaviour is monitored and verified by contact record scheme. Every node reports its encountered node with contact records. Any forging in contact records would be identified since there would be inconsistencies in the contact records within the network. The encountered node announces every contact records across the network to at least two witness nodes. Any witness node detecting the inconsistent contact record would report it, and therefore, the misbehaviour shall be identified.
Reputation-based protocol is introduced by Gianluca et al. in . Here, the node with highest reputation is chosen as the next forwarding node. The node list keeps the list of all nodes through which the message has passed through to reach the destination. A node adds itself only once, even though a message passes through that node many times. To avoid malicious nodes from increasing the reputation of other malicious nodes, a list of digital signatures is also attached. This protocol adapts to the changing conditions of DTN and has reduced overhead compared to the existing protocol.
In addition, few energy-preserving routing protocols do exist for opportunistic networks [20–26]. Location-based routing  and directional routing  is also experimented for WSNs. Innovative data forwarding methods based on computational geometry  have equally been researched so far. The following section briefs the methodology behind the proposed social network-based opportunistic routing for wireless sensor networks (SoNR).
Friends list is created among the good reputation nodes, thus avoiding intruder nodes. But an intruder node in the group can be a friend of another intruder node. A random number generator is used for generating the friends list. The created friends list is stored in a hashmap, where the source node's id is the key for storing its friends list. A node cannot be a friend of itself. An intruder node should not be a friend of a node with good reputation.
Once the friends list is created, messages should be forwarded through the friends in the friends list. For this, the source node first identifies its neighbour nodes. Then, these neighbour nodes are compared with the friends list to identify if any of the neighbour nodes is in the friends list of a source node. If the neighbour node is a friend node of source node, then the messages to the destination is forwarded through this friend node. If the source node meets more than one neighbour friend node, then it forwards the message to all the friend nodes. Hence, the chance of delivering the message to the destination increases.
4 Social network routing
One of the routing techniques available for opportunistic networks is a routing method called epidemic routing . In epidemic routing, messages are routed by flooding the network with copies of messages. Therefore, if a path exists between source and destination, message will certainly be delivered via that path. But sending large numbers of redundant messages is wasteful and will drain the batteries of the sensor nodes rapidly. Though many sensor nodes operate under solar power, a methodology which would not consume much of the energy is ever advisable.
Another main disadvantage of this routing is that the messages are flooded between intruder nodes also. An intruder node will not forward the incoming messages and hence the messages will not reach the destination. Social network routing is one which provides solution to the abovementioned problem.
In social routing, the messages are forwarded through friends of source node or destination node. Friends list is formed only by using good reputation nodes. Hence, there is no possibility for messages forwarded through intruder nodes. The source node first identifies its neighbour nodes. Then, these neighbour nodes are compared with the friends list to identify if any of the neighbour nodes is in the friends list of a source node. If the neighbour node is a friend node of source node, then the messages to the destination is forwarded through this friend node.
If none of the neighbour node is in the friends list of source node, then the data is stored in the buffer of the source node until it meets the destination node and delivers it once the source node and destination node meets each other. Once the messages in the buffer reach the maximum time to live (TTL) value, they are dropped by the source node. If the incoming message for a node is one which is already in the node's buffer or if there is not enough space for the incoming message, then this node will reject the incoming message. The main advantage of this type of routing is the probability of delivering the data to destination is high.
Social network routing is a protocol in which the messages are forwarded through the friends list of source node or destination node. There is a possibility that the nodes in the friends list might become selfish or malicious in the future. If a friend becomes malicious, then it would drop all the incoming messages to it except for its own destined message. If a friend node becomes selfish, then it would forward the messages only for a short period of time, and after sometime, it would start dropping all the incoming messages except for its own destined message. Hence, it is necessary to monitor all the nodes in the friends list.
In this paper, we propose the integrated social network routing protocol in which the friends list is monitored periodically for identifying the misbehaving nodes in the friends list. Once the misbehaving nodes are identified, they are removed from the friends list and the friends list is updated only with good reputed friend nodes.
The integrated social network routing protocol calculates the reputation of all the nodes in the network and maintains the reputation values in a reputation table. It uses two methodologies for calculating the reputation of nodes in the network, namely (1) acknowledgement system and (2) message delivery-based reputation system.
In the acknowledgement system, the destination node creates the ACK messages for each helper node which helped for transmitting its own destined message to it. It sends the created ACK to the source node of the message, and if the source node receives the ACK, it would increment the reputation of the helper nodes. In message delivery-based reputation system, the destination node will directly increment the reputation of helper nodes in the reputation table.
The reputation values in the reputation table are periodically analyzed, and the reputation threshold (RT) is calculated. Then the reputation (R) of all the nodes is compared with this reputation threshold, and if any node's reputation value is less than the reputation threshold, then the node is identified as misbehaving node. However, the misbehaving node thus identified is not immediately removed from the friends list, but is added to a list of suspicious nodes. And if a node remains in the suspicious nodes list for a long time, it is removed from the friends list. Then, the updated friends list consists of only good reputed friend nodes and ensures good delivery rate of messages.
5 Reputation calculation using acknowledgement system
6 Reputation calculation using message delivery-based reputation system
Initialize the reputation values of all the nodes in the network
- 3.If the message (M) reaches the destination node (D), then 'D’
Retrieves the path information of the received message
Identifies the helper nodes from path  of the message M)
Increments the reputation values of the helper nodes
7 Threshold-based updation of friends list
The following parameters are considered for the assessment of integrated social network routing protocol.
Delivery probability - maximum probability to deliver the message successfully.
Overhead - additional bytes relayed to ensure packet delivery with maximum probability.
Average delay - the duration between the message's generation time and the message's delivery time.
False negative - if a misbehaving node is classified as a good node, then it a false negative.
False negative - if a misbehaving node is classified as a good node, then it a false negative.
Classification accuracy - accuracy of a measurement system is the degree of closeness of measurements of a quality to that quantity's actual (true) value.
Evaluation of the proposed protocol with AODV and DSR – with 50% malicious nodes
Proposed SoNR (Ack)
Proposed SoNR (Rep)
Packet Delivery Ratio
Routing Overhead %
The reputation-based SoNR shows significant reduction in routing overhead since AODV and DSR are less efficient in identifying the presence of malicious nodes and are less capable for malicious node-aware effective data forwarding. Subbaraj et al  support the above claim that the standard AODV and DSR protocols require trust and reputation support for handling routing in the presence of malicious nodes.
This research mainly focused on integrated social network routing (SoNR) protocol which routes the messages only through good reputed friend nodes. Simulation results show that the delivery probability of message delivery-based reputation system of integrated social network routing protocol is 30% better when compared to that of the acknowledgement system of integrated social network routing protocol. The percentage of classifying bad nodes as good nodes is very less. However, the present work does not consider the collusion of malicious and selfish nodes.
- Li M, Zhenjiang L, Vasilakos AV: A survey on topology control in wireless sensor networks: taxonomy, comparative study, and open issues. Proc. IEEE 2013, 101(12):1-20.View ArticleGoogle Scholar
- Vasilakos A, Saltouros MP, Atlassis AF, Pedrycz W IEEE Transactions on Systems, Man, and Cybernetics, Part C: Applications and Reviews, 33. Optimizing QoS routing in hierarchical ATM networks using computational intelligence techniques 2003, 3: 297-312.Google Scholar
- Spyropoulos T, Rais RN, Turletti T, Obraczka K, Vasilakos A: Routing for disruption tolerant networks: taxonomy and design. Wirel. Netw 2010, 16(8):2349-2370. 10.1007/s11276-010-0276-9View ArticleGoogle Scholar
- Vasilakos AV, Zhang Y, Spyropoulos T (Eds): Delay Tolerant Networks: Protocols and Applications. CRC Press; 2012.Google Scholar
- Yao Y, Cao Q, Vasilakos AV IEEE 10th International Conference on Mobile Ad-Hoc and Sensor Systems (MASS). EDAL: An Energy-Efficient, Delay-Aware, and Lifetime-Balancing Data Collection Protocol for Wireless Sensor Networks 2013, 182-190.Google Scholar
- Han K, Luo J, Liu Y, Vasilakos AV: Algorithm design for data communications in duty-cycled wireless sensor networks: a survey, communications magazine. IEEE 2013, 7: 51.Google Scholar
- Acampora G, Cook DJ, Rashidi P, Vasilakos AV: A survey on ambient intelligence in healthcare. Proc. IEEE 2013, 101(12):1-25.View ArticleGoogle Scholar
- Parris I, Henderson T: Privacy-Enhanced Social Network Routing, Computer Communications. Volume 1. 35th edition. Elsevier; 2012:62-74.Google Scholar
- Lilien L, Kamal ZH, Bhuse V, Gupta A: The concept of opportunistic networks and their research challenges in privacy and security. Mobile Wireless Network Secur. Privacy 2007, 85-117.View ArticleGoogle Scholar
- Li P, Guo S, Yu S, Vasilakos AV: CodePipe: an opportunistic feeding and routing protocol for reliable multicast with pipelined network coding, INFOCOM. Proc. IEEE 2012, 108: 100-108.Google Scholar
- Li N, Das SK: A trust-based framework for data forwarding in opportunistic networks. Ad Hoc Netw. 2013, 11(4):1497-1509. 10.1016/j.adhoc.2011.01.018View ArticleGoogle Scholar
- He D, Chen C, Chan S, Bu J, Vasilakos AV: ReTrust: attack-resistant and lightweight trust management for medical sensor networks. IEEE Trans. Inf. Technol. Biomed. 2012, 16(4):623-632.View ArticleGoogle Scholar
- He D, Chen C, Chan S, Bu J, Vasilakos AV: A distributed trust evaluation model and its application scenarios for medical sensor networks. Inf. Technol. Biomed. IEEE Trans. 2012, 16(6):1164-1175.View ArticleGoogle Scholar
- Zhengguo S, Yang S, Yu Y, Vasilakos AV, McCann JA, Leung KK: A survey on the ietf protocol suite for the internet of things: standards, challenges, and opportunities. Wireless Commun. IEEE 2013, 20(6):91-98.View ArticleGoogle Scholar
- Moustafa Y, Ibrahim M, Abdelatif M, Chen L, Vasilakos A: Routing Metrics of Cognitive Radio Networks: A Survey. 2013, 1-18.Google Scholar
- Liu Y, Li K, Jin Y, Zhang Y, Qu W: A novel reputation computation model based on subjective logic for mobile ad hoc networks. Futur. Gener. Comput. Syst. 2011, 27(5):547-554. 10.1016/j.future.2010.03.006View ArticleGoogle Scholar
- Bigwood G, Henderson T 2011 IEEE third international conference on and 2011 ieee third international conference on social computing (socialcom). In IRONMAN: Using Social Networks to Add Incentives and Reputation to Opportunistic Networks, Rivacy, Security, Risk And Trust (Passat). IEEE, MIT, Boston, USA; 2011:65-72.Google Scholar
- Li Q, Cao G: Mitigating routing misbehavior in disruption tolerant networks. Inf. Forensics Secur. IEEE Trans. 2012, 7(2):64-675.Google Scholar
- Dini G, Lo Duca A: Towards a reputation-based routing protocol to contrast blackholes in a delay tolerant network. Ad Hoc Netw. 2012, 10(7):1167-1178. 10.1016/j.adhoc.2012.03.003View ArticleGoogle Scholar
- Ahmedy I, Ngadi MA, Omar SN: Using store-forward technique to conserve energy in wireless sensor networks: initial step for routing mechanism, Computing Technology and Information Management (ICCM). 8th Int Conf 2012, 2(1):671-676.Google Scholar
- Chelloug S, Benmohammed M: Simulated Annealing for Maximizing the Lifetime of Sensor Networks under Opportunistic Routing, Proceedings of the 2012 Seventh International Conference on Broadband, Wireless Computing, Communication and Applications. IEEE Computer Society, Canada; 2012:14-19.Google Scholar
- Wei C, Zhi C, Fan P, Letaief KB: AsOR: an energy efficient multi-hop opportunistic routing protocol for wireless sensor networks over Rayleigh fading channels. Wireless Commun. IEEE Trans. 2009, 8(5):2452-2463.View ArticleGoogle Scholar
- Kaliszan M, Stanczak S: Maximizing Lifetime in Wireless Sensor Networks Under Opportunistic Routing, Signals, Systems and Computers (ASILOMAR), 2010 Conference Record of the Forty Fourth Asilomar Conference. IEEE, California, USA; 2010:1913-1917.View ArticleGoogle Scholar
- Rusli ME, Harris R, Punchihewa A: Quality aware opportunistic routing protocol with adaptive coordination scheme for wireless sensor networks. In Computational Intelligence, Modelling and Simulation (CIMSiM), 2012 Fourth International Conference. IEEE, Malaysia; 2012:434-439.View ArticleGoogle Scholar
- Rusli ME, Harris R, Punchihewa A: Performance analysis of implicit acknowledgement coordination scheme for opportunistic routing in wireless sensor networks. In Telecommunication Technologies (ISTT), 2012 International Symposium. IEEE, Malaysia; 2012:131-136.View ArticleGoogle Scholar
- Soares JM, Rocha RM: CHARON: routing in low-density opportunistic wireless sensor networks. In Wireless Days (WD), 2009 2nd IFIP. IEEE, Paris; 2009:1-5.View ArticleGoogle Scholar
- Han Y, Lin Z: A geographically opportunistic routing protocol used in mobile wireless sensor networks. In Networking, Sensing and Control (ICNSC), 2012 9th IEEE International Conference. IEEE, Beijing; 2012:216-221.View ArticleGoogle Scholar
- Zeng Y, Xiang K, Li D, Vasilakos AV: Directional routing and scheduling for green vehicular delay tolerant networks. Wirel. Netw 2013, 19(2):161-173. 10.1007/s11276-012-0457-9View ArticleGoogle Scholar
- Luo J, Cai Y: A data forwarding scheme based on delaunay triangulation for CPSs. In Wireless Communications, Networking and Mobile Computing (WiCOM), 2012 8th International Conference. IEEE, China; 2012:1-6.Google Scholar
- Tan SK, Munro A: Adaptive probabilistic epidemic protocol for wireless sensor networks in an urban environment. Proceedings of 16th International Conference. In Computer Communications and Networks, 2007. ICCCN 2007. IEEE; 2007:1105-1110.Google Scholar
- Subbaraj S, Prakash S: EigenTrust-based non-cooperative game model assisting ACO look-ahead secure routing against selfishness. Eurasip J. Wirel. Commun. Netw. 2014, 2014(1):78. 10.1186/1687-1499-2014-78View ArticleGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited.