Table of Contents Author Guidelines Submit a Manuscript
Mobile Information Systems
Volume 2018 (2018), Article ID 6462826, 8 pages
https://doi.org/10.1155/2018/6462826
Research Article

EpSoc: Social-Based Epidemic-Based Routing Protocol in Opportunistic Mobile Social Network

1Faculty of Computer Science and Information Technology, Universiti Malaysia Sarawak, Kota Samarahan, Malaysia
2Faculty of Computer and Information Sciences, Jouf University, Jouf, Saudi Arabia

Correspondence should be addressed to Halikul Lenando; ym.saminu@looc

Received 17 October 2017; Revised 20 February 2018; Accepted 11 March 2018; Published 4 April 2018

Academic Editor: Fabio Gasparetti

Copyright © 2018 Halikul Lenando and Mohamad Alrfaay. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Abstract

In opportunistic networks, the nature of intermittent and disruptive connections degrades the efficiency of routing. Epidemic routing protocol is used as a benchmark for most of routing protocols in opportunistic mobile social networks (OMSNs) due to its high message delivery and latency. However, Epidemic incurs high cost in terms of overhead and hop count. In this paper, we propose a hybrid routing protocol called EpSoc which utilizes the Epidemic routing forwarding strategy and exploits an important social feature, that is, degree centrality. Two techniques are used in EpSoc. Messages’ TTL is adjusted based on the degree centrality of nodes, and the message blocking mechanism is used to control replication. Simulation results show that EpSoc increases the delivery ratio and decreases the overhead ratio, the average latency, and the hop counts as compared to Epidemic and Bubble Rap.

1. Introduction

Opportunistic mobile social network (OMSN) [14] is a promising networking model for data dissemination. In the OMSN, mobile nodes grab the opportunity of encountering the peer (they are in the communication range of each other) to forward the data. The OMSN incurs intermittent and disruptive connectivity due to node mobility. To tackle with this complex environment, the store-carry-forward scheme is applied in the OMSN. If no connection is available at a particular time, a mobile node stores data in its buffer and carries them until it encounters other mobile nodes to forward the data [58]. Various approaches have been proposed to address the information delivery problem in the OMSN such as in [912]. The main concerns of OMSN routing approaches are yielding high delivery ratio, low delay, and low overhead or cost on networks and nodes.

Flooding is one of the dominant schemes to disseminate data in the OMSN [13]. Each message will be flooded to every node in the network. Multiple copies of each message are generated and spread in the network. Epidemic [9] routing protocol is the flooding-based routing protocol. When two nodes encounter, they exchange all of their messages. This results in messages spread over the whole network by pairwise contacts between two nodes. If no buffer constraints are applied, Epidemic represents the upper bound in message delivery and latency. Epidemic routing is used as a benchmark and a reference for the most of other routing protocols in the opportunistic network. The main drawback of the Epidemic scheme is its high overhead. Many schemes are proposed to decrease the overhead in Epidemic-based approaches by limiting the number of message replicas [1416]. An effective scheme to control replication spread is the vaccine [17]. It applies the antipacket mechanism to control replica distribution in Epidemic-based routing. In [18], a new scheme is proposed to control the replication of epidemically distributed information. Based on the vaccine scheme, signal distribution is early controlled by the fully immunized vaccine. In addition, a partially immunized vaccine is initiated when there is a local-forwarding opportunity to vaccine more packets.

In the OMSN, mobile devices are portable by humans so that social features of people can be exploited for networking purposes [1921]. Social-based protocols utilize social properties of mobile users such as similarity, centrality, and friendship to improve routing efficiency in the opportunistic mobile social network. This is because social features are more stable and less changeable than other features like mobility patterns. HiBOp [22] and CiPRO [23] exploit the similarity social feature and the user’s context information to forward data. LASS [24] takes into account the difference of members’ activity within the node’s community for data dissemination. In ML-SOR [25], node centrality (different types of centralities), the similarity between communities, and social ties are all exploited to effectively select the forwarding node. MCAR [26] exploits the preferred communities of people during their daily lives for effective information delivery. Direct (inside one community) and indirect (via different communities) contacts are considered in MCAR. In SPRINT-SELF [27], network and node overhead is decreased by exploiting social information of mobile users. The authors consider the social community of nodes and utilize it to predict future behavior based on contact history. In addition, they proposed a new mechanism to avoid selfish nodes for more improvements.

Social features are utilized widely for buffer management. Liu et al. [28] utilized social features and the congestion level to develop the forwarding strategy that drops the message with the minimum social link rather than random dropping. In SRAMSW [29], the buffer management mechanism is combined with social features to enhance the spray-based routing. Expired messages are deleted, successfully delivered messages are acknowledged, and messages are prioritized according to their spray times and residence time. In addition, three social features: centrality, similarity, and friendship, are adopted for better forwarding decision and to avoid the dead-end problem.

We hypothesize that combining social features with the Epidemic-based forwarding scheme improves the efficiency of routing in the OMSN. In this paper, we present an Epidemic-based Social-based routing protocol (EpSoc) that combines the advantages of the forwarding strategy used in the Epidemic routing protocol with the positive impact of exploiting social features. EpSoc exploits the degree centrality social feature to adapt the time to live (TTL) of the routed message. If a message is forwarded to a node which has higher degree centrality (socially active node), its TTL value will be decreased. If these messages are dropped in the active node when TTL is zero, the blocking mechanism is used to reject receiving replications of these messages.

The rest of this paper is organized as follows: the next section reviews the related work. We describe in detail our proposed algorithm EpSoc in Section 3. In Section 4, we introduce the performance evaluation and the discussion results. Finally, Section 5 concludes the paper.

2. Related Works

Delivering data to the destination at the right time with minimum resources is an optimal condition for any given routing protocols. Epidemic has optimal performance in terms of delivery ratio and latency. However, it suffers from high overhead cost. One of the solutions is to exploit social information to improve Epidemic-based routing protocols in the OMSN. Degree centrality is a social feature that is exploited widely in the literature to improve routing in the mobile social networks. For example, Bubble Rap [11] utilizes the mobile node’s degree centrality to provide cost-effective routing as compared to Epidemic routing. Bubble Rap exploits two social and structural metrics, namely, centrality and community. It selects high centrality nodes and community members of destination as relays. In the Bubble Rap algorithm, nodes belong to different sizes of communities and have different levels of popularity (i.e., rank). Each node is assumed to have two rankings: global denotes the popularity (i.e., connectivity) of the node in the entire society and local denotes the popularity within its community. Messages are forwarded to nodes that have higher global ranking until a node in the destination’s community is found. Then, the messages are forwarded to nodes having a higher local ranking within the destination’s community.

Similar to Bubble Rap, CAOR [30] exploits degree centrality and similarity social features to improve routing. However, CAOR constructed autonomous communities based on common interest locations between mobile nodes. Members of a community with high centralities act as the home of this community. CAOR also has a mechanism to turn the routing between lots of nodes to the routing between a few community homes. It also applies the reverse Dijkstra algorithm to determine the optimal relays and compute the minimum expected delivery delay. In [31], cultural algorithm (CA), ant colony optimization (ACO), and social connectivity between users are combined to address the routing problem. Social metrics of nodes including degree and betweenness centralities are analyzed to support forwarding decision in the opportunistic network environment.

Besides exploited in the routing problem, social features were also applied to solve different type of problems. For example, works in [32] exploit social features (degree and betweenness centralities) to solve the throwbox placement problem based on the given graph. A user’s degree is equal to the total number of its neighbors, and betweenness is the total number of shortest paths passing through the node. The work also introduced the concept of location degree to measure how many mobile users have the location as one of their top visited locations and the location betweenness to show how important the location is for the entire social graph. Social metrics such as degree centrality, social activeness, and community acquaintance are also applied to enhance data delivery in VSNs in [33]. The social interactions between nodes where nodes of similar interests or nodes belonging to the same community have greater probability to encounter each other. Social centrality is also utilized for congestion control in the postdisaster environment [34].

Unlike the aforementioned works, instead of considering community or homing structuring, our proposed protocol is based on the flooding-based forwarding strategy of Epidemic. This is because we intended to design our forwarding protocol to have high delivery ratio and low delay as the Epidemic-based forwarding strategy. Moreover, we also aimed to avoid extra time required to form and maintain the community structure. Thus, to decrease the overhead, we utilized degree centrality and adapted the message’s TTL (Time to live) to control the forwarding nodes’ activity. The message’s TTL is considered in literature when developing efficient routing protocols. Miao et al. [33] proposed the adaptive multistep routing protocol for mobile delay-tolerant networks (MDTNs). Their aim is to get the balance between the delay and cost of message delivery. The time to live of the message is used in order to allocate the minimum number of copies necessary to achieve a given delivery probability. In [34], the authors considered the contact information of nodes and the time to live message’s property to make route decision and improve performance. They built a replica distribution criterion between two encountered nodes based on residual messages’ TTL. In our proposed solution, we do not limit the number of replicas but allow messages to be spread in the network and then we utilize degree centrality to decrease the TTL and consequently decrease the overhead. Also, the blocking mechanism is adopted in the active node to cancel receiving replications of the same message which is seen previously.

3. EpSoc Routing Protocol

EpSoc is the routing algorithm designed to decrease overhead in the Epidemic protocol by embedding social features in routing the message in the opportunistic network. To achieve this objective, we propose two mechanisms. First, the message’s TTL is adapted based on degree centrality of the nodes in the OMSN. The second one is the message blocking mechanism that is used to prevent receiving the replications of runout TTL messages in active nodes.

3.1. Degree Centrality

Node centrality indicates the popularity of the node in the network, whereby node centrality in a social network is the reflection of its social relative importance [35]. A higher node degree centrality means the node connects with many numbers of nodes in the network. A degree centrality for a given node can be calculated as follows:where is the number of nodes in the network and if a direct link exists between node and node and .

We adopt the CWindow [11] calculation algorithm to calculate degree centrality. CWindow divides the day into time windows and calculates the average of nodes’ degree over these windows to estimate the node’s centrality. To decrease the processing overhead results from processing any changes of degree centrality, CWindow calculates the node’s degree centrality at regular intervals instead of every time the centrality changes.

We pick the CWindow algorithm to calculate degree centrality because it takes into account the changes in node centrality over time and averages the node’s centralities for few window intervals which is suitable for people’s behavior in the OMSN. It is also used by other social-based protocols that consider the node’s degree centrality such as Bubble Rap [11], Dlife [36], and SCORP [37].

3.2. EpSoc Forwarding Strategy

Figure 1 depicts the forwarding process in EpSoc, where two mechanisms are applied.

Figure 1: EpSoc forwarding scheme.

In Figure 1(a), node encountered three nodes: , , and . has higher degree centrality (DC) than . The TTL value of the forwarded messages to is decreased by dividing it by the DC value of . We call these messages socially infected messages. Node registers the ID of these messages as seen messages in its blocking register. Node will drop the socially infected message when it expires. In Figure 1(b), the Epidemic forwarding strategy which we adopted in our algorithm causes replications of socially infected messages to be sent to node from other nodes such as . In this case, node will reject receiving the message with the reason that it is a seen message (its ID is stored in the block register).

Based on (1), the TTL value is adapted using the following equation:

The two combined mechanisms adopted in EpSoc enhance routing performance. If a node is socially active, it meets more nodes in the network and has a higher potential to deliver more messages. Decreasing messages’ TTL value in active nodes results in releasing space in their buffers and increases the ability to deliver more messages. This will increase the delivery ratio. The blocking mechanism controls the replications by preventing decreased TTL messages from hitting again the previously traversed actives nodes. The result is decreasing network overhead. Regarding average latency, the messages that are delivered by active nodes have a shorter end-to-end delay compared to other messages delivered by lower activity nodes, so average latency will be decreased too. We conclude that decreasing message life socially by combining the message blocking scheme affects positively the performance of the Epidemic routing scheme in the OMSN.

3.3. Pseudocode of the EpSoc Message Forwarding

We used CWindow to calculate the centrality of the node. Each node records the encountered nodes in the network. When node encounters , the degree centrality values are calculated using CWindow. Then, centrality values are exchanged between both nodes. compares its centrality value with the centrality value . If is greater than , which means that is socially more active than , then for each message in the buffer , the TTL value is decreased by dividing it by the node centrality value . The ID of each socially infected message is registered in the blocking register . The time complexity of the EpSoc forwarding algorithm is , where is the number of messages carried in the node’s buffer and needed to be sent or forwarded.

Algorithm 1 shows the complete pseudocode of EpSoc.

Algorithm 1: Pseudocode of the forwarding strategy in EpSoc.

4. Performance Evaluations

4.1. Data Set

We adopt the Cambridge experimental data set of haggle [38]. This data set includes traces of Bluetooth sightings by groups of users carrying small devices (iMotes) for a number of days in campus environments. The experiments are conducted in the computer laboratory that includes the undergraduate first-year and second-year students and also some Ph.D. and postgraduate students which lasted for 11 days.

4.2. Simulation Setup

We use the opportunistic network environment (ONE) [39] simulator to evaluate our algorithm. Also, comparison with Epidemic and Bubble Rap routing protocols is included to measure the performance of our proposed algorithm EpSoc. We want to justify that the Epidemic algorithm performs better when considering social features in forwarding messages. The simulator settings are tabulated in Table 1.

Table 1: Simulation settings.

In each experiment, we compare the performance of the protocols EpSoc, Epidemic, and Bubble Rap based on the following metrics.

4.2.1. Successful Delivery Ratio

It is the ratio between the number of delivered messages and the total number of created messages. The ideal value of the successful delivery ratio is 1.0 when all created messages are delivered to their destinations.

4.2.2. Overhead Ratio

It is the additional bytes that are sent for successfully delivering a message to a destination.

4.2.3. Average Latency

It is the average of the time elapsed between message creation and delivery.

4.2.4. Average Hop Count

It is the average of the number of hops that messages must take in order to reach the destination.

4.3. Experiments and Discussion

To evaluate our work, we will consider two features: buffer size and message TTL value. These two features have a high impact on routing performance in the OMSN. In our work, we affect these two features. We adjust the TTL value socially and use the blocking mechanism to manage buffer storing. Consequently, our experiments are to measure the performance of EpSoc when varying the TTL and buffer size. The comparison will be with Epidemic and Bubble Rap protocols.

4.3.1. Varying Buffer Size

For varying the buffer size, we fixed the value of TTL to 2.5 d. Figures 25 show the performance comparisons between EpSoc, Epidemic, and Bubble Rap in terms of delivery ratio, overhead ratio, average latency, and average hop count, respectively.

Figure 2: Delivery ratio versus buffer size.
Figure 3: Overhead ratio versus buffer size.
Figure 4: Average latency versus buffer size.
Figure 5: Average hop count versus buffer size.

Figure 2 shows the delivery ratio with buffer size. Generally, for Epidemic, Bubble Rap, and EpSoc, increasing the buffer size will increase the delivery ratio. This is because more messages can be carried by intermediate nodes which consequently deliver more messages to destinations. Changing the number of delivered messages affects the delivery ratio, overhead ratio, average latency, and average hop count. Regarding the delivery ratio, it is clear that delivering more messages results in a higher value, while the decrease in delivered messages results in a lower value. In the lower buffer size scenario (1–25 MB), the delivery ratio of Epidemic is the lowest because of the redundancy. Bubble Rap and EpSoc outperform Epidemic due to utilizing social features. In higher buffer size scenarios (35–55 MB), Epidemic achieves higher delivery ratio than Bubble Rap. Larger buffer size alleviates the negative impact of dropping messages because of replication and enables Epidemic to deliver more messages. Our protocol EpSoc outperforms both Epidemic and Bubble Rap. Blocking runout TTL messages from being received by active nodes results in more space in their buffer and therefore can carry more different messages when encountering other nodes and later deliver them to destinations. In addition, the decrease of TTL of the messages that are forwarded to the more active nodes results in better utilization of the buffer’s space. Decreased TTL message copies are dropped earlier enabling carrying more other messages. Active nodes deliver the message quickly. Therefore, the number of delivered messages is increased, and consequently, the delivery ratio is increased.

The relation of overhead ratio with buffer size is shown in Figure 3. When increasing the buffer size, the overhead is decreased for all algorithms. Regarding Bubble Rap, overhead is decreased with buffer increase because no replication exists. We observe from Figure 3 that both protocols Bubble Rap and EpSoc outperform Epidemic significantly and EpSoc outperforms Epidemic. We mentioned formerly that applying the strategy of blocking messages and decreasing TTL increase the number of delivered messages. In addition, the blocking message to be resent to active nodes decreases the number of replications in the network and hence decreases the forwardings. As a result, decreasing the forwardings and increasing the delivered messages decrease the overhead ratio in the network. Bubble Rap achieves lower overhead than EpSoc for lower buffer size scenarios (1, 2, 5, and 15 MB). This is because of two reasons: first, the replication and Epidemic strategy adopted in EpSoc, and second, low buffer size causes dropping relayed messages early due to buffer overflow which in turn decreases the efficiency of our proposed algorithm compared to Bubble Rap in terms of overhead. However, for larger buffer size scenarios (25, 35, 45, and 55 MB), EpSoc is more efficient and its overhead ratio is very close to that of Bubble Rap (for 35, 45, and 55 MB, it is slightly better than Bubble Rap).

Figure 4 shows that average latency for all the three protocols, which increases when the buffer size is increased. A low-sized buffer only relays low-latency messages which will reach their destinations quickly. On the other hand, a higher-sized buffer allows messages to be carried for a longer time which contribute to a higher average latency. EpSoc decreases the average latency significantly. EpSoc optimizes the buffer usage where messages forwarded to active nodes have low TTL. With the low TTL, a relay node has more free space for new messages in its buffer as the buffer quickly dropped the old messages.

In Figure 5, the average hop count is recorded. The average hop count increases when the buffer size is increased. Epidemic has more hop count which indicates more nodes experiencing the duplicated messages, whereas Bubble Rap has lower hop count because it prevents replication of messages. For EpSoc, the number of hop count is contributed by allowing replication as in Epidemic.

The worst achievement is because of redundancy. Bubble Rap does not apply replication so that it outperforms EpSoc. EpSoc achieves better average hop count than Epidemic. This is because of the message blocking strategy which decreases the number of replications in the network and hence decreases the forwardings.

4.3.2. Varying Initial TTL

TTL determines the life of the message in the network. A high value of TTL gives high chances of the message delivered to the target destination and vice versa.

In Figure 6, the relation of delivery ratio with TTL is depicted. For lower TTL (10 m–3 h), the delivery ratio of Epidemic is slightly higher than that of Bubble Rap and EpSoc. The reason is that, for short TTL messages, exploiting social features will not be very effective due to quickly messages dropping. In addition, the higher replication occurred in Epidemic increases the number of delivered messages. When TTL is increased, Bubble Rap and EpSoc outperform Epidemic due to utilizing the social features. In high TTL scenarios (1.5 d–1 w), EpSoc outperforms Bubble Rap. This indicates an efficiency of EpSoc in its social feature selection.

Figure 6: Delivery ratio versus TTL.

Shortening the life of messages in the active nodes and blocking runout TTL messages from hitting again the active nodes cause an increment in the delivered messages and decrement in forwardings. Therefore, delivery ratio grows up.

Figure 7 compares the performance of the protocols with different values of TTL in terms of overhead ratio. When TTL is very low (10 m–1 h), Epidemic, EpSoc, and Bubble Rap achieve low overhead. The reason is that messages are dropped quickly. For high TTL values, Bubble Rap achieves the best and Epidemic the worst.

Figure 7: Overhead ratio versus TTL.

Our protocol EpSoc manages to decrease the overhead better than Epidemic. This is because EpSoc is a combination of the flooding-based forward strategy with social features.

From Figure 8, in terms of average latency, EpSoc appears to be outperforming others especially with high TTL values (12 h–1 w). This is because our algorithm always enables the active node to carry messages with lower TTL.

Figure 8: Average latency versus TTL.

From Figure 9, we observed that if TTL is very low (10 m–1 h), all routing protocols in the experiment have low average hop count. This is due to the low number of forwardings between nodes as message life exhausted quickly. Bubble Rap and EpSoc have almost stable performance when TTL is increased. Exploited social features in Bubble Rap and EpSoc and applying the blocking mechanism in EpSoc decrease the effect of changing the value of TTL on the number of traversed nodes to the destination. On the contrary, when the TTL is increased, the average hop count of Epidemic increases significantly because of the flooding-based forwarding strategy. For EpSoc, the average hop count is decreased significantly compared to Epidemic because of the social effect of active nodes.

Figure 9: Average hop count versus TTL.

5. Conclusion

In this paper, we investigate the flooding-based forwarding strategy, that is, Epidemic with social features to improve the routing performance in the opportunistic mobile social network (OMSN). Inspired by the advantages of Epidemic routing protocols in terms of delivery ratio and delivery delay and by exploiting the social activity of nodes, we formulated a flooding-based social-based routing protocol named as EpSoc. Simulation experiments using real data sets are conducted to evaluate our protocol performance. From the presented results, our approach increases the delivery ratio and decreases the delivery overhead, average latency, and average hop count as compared to the Epidemic protocol. As for benchmark social-based routing protocol performance (Bubble Rap), our protocol decreases the average latency significantly with a better delivery ratio in high buffer size and low TTL scenarios. Generally, we manage to exploit the advantage of Epidemic and Bubble Rap to improve the data dissemination in the OMSN.

Additional Points

Significance. Social features of a node can be utilized to have an effective role in the routing protocol in the OMSN network. This paper presents the routing protocol that exploits degree centrality to increase the delivery ratio and decrease the overhead and latency. Moreover, exploiting other social features such as similarity and community may lead to being more efficient routing protocol. Therefore, combining social features with other forwarding techniques such as the Epidemic-based strategy is significant to have an efficient forwarding strategy in the OMSN to support green technologies.

Conflicts of Interest

The authors declare that there are no conflicts of interest regarding the publication of this paper.

References

  1. A. Chaintreau, A. Mtibaa, L. Massoulie, and C. Diot, “The diameter of opportunistic mobile networks,” in Proceedings of the 2007 ACM CoNEXT Conference, p. 12, ACM, New York, NY, USA, December 2007.
  2. M. Conti, S. Giordano, M. May, and A. Passarella, “From opportunistic networks to opportunistic computing,” IEEE Communications Magazine, vol. 48, no. 9, pp. 126–139, 2010. View at Publisher · View at Google Scholar · View at Scopus
  3. K. Zhu, W. Li, X. Fu, and L. Zhang, “Data routing strategies in opportunistic mobile social networks: taxonomy and open challenges,” Computer Networks, vol. 93, pp. 183–198, 2015. View at Publisher · View at Google Scholar · View at Scopus
  4. Y. Liu, Z. Yang, T. Ning, and H. Wu, “Efficient quality-of-service (QoS) support in mobile opportunistic networks,” IEEE Transactions on Vehicular Technology, vol. 63, no. 9, pp. 4574–4584, 2014. View at Publisher · View at Google Scholar · View at Scopus
  5. L. Pelusi, A. Passarella, and M. Conti, “Opportunistic networking: data forwarding in disconnected mobile ad hoc networks,” IEEE Communications Magazine, vol. 44, no. 11, pp. 134–141, 2006. View at Publisher · View at Google Scholar · View at Scopus
  6. T. Le and M. Gerla, “Social-distance based anycast routing in delay tolerant networks,” in Proceedings of the 2016 Mediterranean Ad Hoc Networking Workshop (Med-Hoc-Net), Vilanova i la Geltru, Spain, June 2016.
  7. K. Fall, “A delay-tolerant network architecture for challenged internets,” in Proceedings of the 2003 Conference on Applications, Technologies, Architectures, and Protocols for Computer Communications, New York, NY, USA, August 2003.
  8. M. Alajeely, R. Doss, and A. Ahmad, “Routing protocols in opportunistic networks–a survey,” IETE Technical Review, pp. 1–19, 2017. View at Publisher · View at Google Scholar · View at Scopus
  9. A. Vahdat and D. Becker, “Epidemic routing for partially-connected ad hoc networks,” Tech. Rep. CS-200006, Duke University, Durham, NC, USA, 2000. View at Google Scholar
  10. A. Lindgren, A. Doria, and O. Schelen, Probabilistic Routing in Intermittently Connected Networks, Springer, Berlin, Germany, 2004.
  11. P. Hui, J. Crowcroft, and E. Yoneki, “BUBBLE Rap: social-based forwarding in delay-tolerant networks,” IEEE Transactions on Mobile Computing, vol. 10, no. 11, pp. 1576–1589, 2011. View at Publisher · View at Google Scholar · View at Scopus
  12. P. Pholpabu and L. L. Yang, “Routing protocols for mobile social networks achieving trade-off among energy consumption, delivery ratio and delay,” in Proceedings of the 2015 IEEE/CIC International Conference on Communications in China (ICCC), Shenzhen, China, November 2015.
  13. M. Nikoo, F. Ramezani, M. Hadzima-Nyarko, E. K. Nyarko, and M. Nikoo, “Flood-routing modeling with neural network optimized by social-based algorithm,” Natural Hazards, vol. 82, no. 1, pp. 1–24, 2016. View at Publisher · View at Google Scholar · View at Scopus
  14. E. Bulut, Z. Wang, and B. K. Szymanski, “Cost-effective multiperiod spraying for routing in delay-tolerant networks,” IEEE/ACM Transactions on Networking (TON), vol. 18, no. 5, pp. 1530–1543, 2010. View at Publisher · View at Google Scholar · View at Scopus
  15. T. Matsuda and T. Takine, “(p,q)-epidemic routing for sparsely populated mobile ad hoc networks,” IEEE Journal on Selected Areas in Communications, vol. 26, no. 5, pp. 783–793, 2008. View at Publisher · View at Google Scholar · View at Scopus
  16. P. Mundur, M. Seligman, and J. N. Lee, “Immunity-based epidemic routing in intermittent networks,” in Proceedings of the 2008 5th Annual IEEE Communications Society Conference on Sensor, Mesh and Ad Hoc Communications and Networks, San Francisco, CA, USA, June 2008.
  17. P. Y. Chen, S. M. Cheng, and K. C. Chen, “Optimal control of epidemic information dissemination over networks,” IEEE Transactions on Cybernetics, vol. 44, no. 12, pp. 2316–2328, 2014. View at Publisher · View at Google Scholar · View at Scopus
  18. C. Y. Aung, I. W. H. Ho, and P. H. J. Chong, “Store-carry-cooperative forward routing with information epidemics control for data delivery in opportunistic networks,” IEEE Access, vol. 5, pp. 6608–6625, 2017. View at Google Scholar
  19. Y. Zhu, B. Xu, X. Shi, and Y. Wang, “A survey of social-based routing in delay tolerant networks: positive and negative social effects,” IEEE Communications Surveys & Tutorials, vol. 15, no. 1, pp. 387–401, 2013. View at Publisher · View at Google Scholar · View at Scopus
  20. C. C. Sobin, V. Raychoudhury, G. Marfia, and A. Singla, “A survey of routing and data dissemination in delay tolerant networks,” Journal of Network and Computer Applications, vol. 67, pp. 128–146, 2016. View at Publisher · View at Google Scholar · View at Scopus
  21. P. Pholpabu and L.-L. Yang, “Social contact probability assisted routing protocol for mobile social networks,” in Proceedings of the IEEE 79th Vehicular Technology Conference (VTC Spring), Seoul, South Korea, May 2014.
  22. C. Boldrini, M. Conti, and A. Passarella, “Exploiting users’ social relations to forward data in opportunistic networks: the HiBOp solution,” Pervasive and Mobile Computing, vol. 4, no. 5, pp. 633–657, 2008. View at Publisher · View at Google Scholar · View at Scopus
  23. H. Nguyen and S. Giordano, “Context information prediction for social-based routing in opportunistic networks,” Ad Hoc Networks, vol. 10, no. 8, pp. 1557–1569, 2012. View at Publisher · View at Google Scholar · View at Scopus
  24. Z. Li, C. Wang, S. Yang, C. Jiang, and X. Li, “LASS: local-activity and social-similarity based data forwarding in mobile social networks,” IEEE Transactions on Parallel and Distributed Systems, vol. 26, pp. 174–184, 2015. View at Publisher · View at Google Scholar · View at Scopus
  25. A. Socievole, E. Yoneki, F. De Rango, and J. Crowcroft, “ML-SOR: message routing using multi-layer social networks in opportunistic communications,” Computer Networks, vol. 81, pp. 201–219, 2015. View at Publisher · View at Google Scholar · View at Scopus
  26. P. Pholpabu and L.-L. Yang, “A mutual-community-aware routing protocol for mobile social networks,” in Proceedings of the Global Communications Conference (GLOBECOM), Austin, TX, USA, December 2014.
  27. R. Ciobanu, C. Dobre, V. Cristea, F. Pop, and F. Xhafa, “SPRINT-SELF: social-based routing and selfish node detection in opportunistic networks,” Mobile Information Systems, vol. 2015, Article ID 596204, 12 pages, 2015. View at Publisher · View at Google Scholar · View at Scopus
  28. Y. Liu, K. Wang, H. Guo, Q. Lu, and Y. Sun, “Social-aware computing based congestion control in delay tolerant networks,” Mobile Networks and Applications, vol. 22, no. 2, pp. 1–12, 2016. View at Google Scholar
  29. J. Guan, Q. Chu, and I. You, “The social relationship based adaptive multi-spray-and-wait routing algorithm for disruption tolerant network,” Mobile Information Systems, vol. 2017, Article ID 1819495, 13 pages, 2017. View at Publisher · View at Google Scholar · View at Scopus
  30. M. Xiao, J. Wu, and L. Huang, “Community-aware opportunistic routing in mobile social networks,” IEEE Transactions on Computers, vol. 63, pp. 1682–1695, 2014. View at Publisher · View at Google Scholar · View at Scopus
  31. A. C. K. Vendramin, A. Munaretto, M. R. Delgado, M. Fonseca, and A. C. Viana, “A social-aware routing protocol for opportunistic networks,” Expert Systems with Applications, vol. 54, pp. 351–363, 2016. View at Publisher · View at Google Scholar · View at Scopus
  32. Y. Zhu, C. Zhang, X. Mao, and Y. Wang, “Social based throwbox placement schemes for large-scale mobile social delay tolerant networks,” Computer Communications, vol. 65, pp. 10–26, 2015. View at Publisher · View at Google Scholar · View at Scopus
  33. J. Miao, O. Hasan, S. B. Mokhtar, L. Brunie, and G. Gianini, “A delay and cost balancing protocol for message routing in mobile delay tolerant networks,” Ad Hoc Networks, vol. 25, pp. 430–443, 2015. View at Publisher · View at Google Scholar · View at Scopus
  34. H. Chen and W. Lou, “Contact expectation based routing for delay tolerant networks,” Ad Hoc Networks, vol. 36, pp. 244–257, 2016. View at Publisher · View at Google Scholar · View at Scopus
  35. F. Xia, L. Liu, J. Li, J. Ma, and A. V. Vasilakos, “Socially aware networking: a survey,” IEEE Systems Journal, vol. 9, pp. 904–921, 2015. View at Publisher · View at Google Scholar · View at Scopus
  36. W. Moreira, P. Mendes, and S. Sargento, “Opportunistic routing based on daily routines,” in Proceedings of the World of wireless, mobile and multimedia networks (WoWMoM), San Francisco, CA, USA, June 2012.
  37. W. Moreira, P. Mendes, and S. Sargento, “Social-aware opportunistic routing protocol based on user’s interactions and interests,” in Proceedings of the International Conference on Ad Hoc Networks, vol. 129, pp. 100–115, Springer International Publishing, Barcelona, Spain, August 2014. View at Publisher · View at Google Scholar · View at Scopus
  38. Website of CRAWDAD project, http://crawdad.org/cambridge/haggle/20090529/.
  39. A. Keränen, J. Ott, and T. Kärkkäinen, “The ONE simulator for DTN protocol evaluation,” in Proceedings of the 2nd International Conference on Simulation Tools and Techniques, Brussels, Belgium, March 2009.