Elsevier

Information Sciences

Volume 369, 10 November 2016, Pages 34-50
Information Sciences

Trustworthiness-hypercube-based reliable communication in mobile social networks

https://doi.org/10.1016/j.ins.2016.05.048Get rights and content

Abstract

In mobile social networks (MSNs), the routing packet is forwarded from any user of in a group to any user of the other group until it reaches the destination group - the group where the destination is located. However, it is inevitable that malicious groups could compromise the quality and reliability of data. To alleviate such effect, analyzing the trustworthiness of a group has a positive influence on the confidence with which a group conducts transactions with that group. In our previous work, the feature-based first-priority relation graph (FPRG) of MSNs is proposed, in which two vertices (groups) are connected iff they have a first-priority relationship. In this paper, the trustworthiness computation of a group is firstly presented in the algorithm TC (Trustworthiness Computing) based on the FPRG. The trustworthiness of a group is evaluated based on the trustworthiness of neighbors and the number of malicious users in the group. We then establish the Trustworthiness-Hypercube-based Reliable Communication (THRC) algorithm in MSNs. The algorithm THRC can provide an effective and reliable data delivery routing. Finally, we also give two scenario simulations to elaborate the processes of the trustworthiness computation and reliable communication.

Introduction

The mobile social network (MSN) paradigm is proposed as a mean of ferrying data through mobile devices using human social contacts [34] (see Fig. 1). MSNs are considered as a case of socially-aware Delay/Disruption Tolerant Networks (DTNs) [17] that are characterized by intermittent connectivity and limited network capacity. MSNs take advantage of social contacts to opportunistically create data paths over time [34]. As storage capacities of mobile devices increase, and support for short-range data transfer protocols (e.g., WiFi and Bluetooth) becomes more prevalent, we can use these devices to forward data along an efficient and reliable path in a store-carry-forward fashion.

In MSNs, users move around based on their common interest and come in contact each other more frequently if they have more social features in common. A group consists of some users (individuals) with same key social features. The routing packet is forwarded from any user of in a group to any user of the other group until it reaches the destination group - the group where the destination is located [41], [42], [43]. However, some users in MSNs may be malicious. A malicious group can compromise the quality and reliability of data. To counter such adverse effect, evaluating the trustworthiness for each group will facilitate the arrangement of the most trustworthy route for inter-groups data delivery (see Fig. 1). Moreover, trustworthiness system are also proved useful in assessing the quality of received information to provide network security services such as access control, authentication and secure resource sharing [2], [5], [25], [26], [33], [35].

Due to the independent movement of component users and relationships, the trustworthiness computation is a highly challenging issue in MSNs. In this paper, we design simple, fast and accurate strategies to compute the trustworthiness for each group.

A. Motivations and goals

Our work is motivated by two specific situations: Efficient trust computation for large-scale mobile social networks using a fuzzy implicit social graph [11] and Cluster-group based trusted computing for mobile social networks using implicit social behavioral graph [12]. In our previous work, we [46] detect malicious users in MSNs by the algorithm FPRG-MUD-PMC (First-Priority Relation Graph-based Malicious Users Detection under the PMC Model). This paper firstly proposes a novel trustworthiness computation scheme by considering more comprehensive aspects, and then aims to design a reliable communication in malicious mobile social networks.

In malicious mobile social networks, a data delivery routing without malices are necessary. If each unmalicious group on the reliable data delivery routing has less malicious neighbors, then the reliable data delivery routing is more reliable.

(1) Trustworthiness computation: high-level accuracy

The trustworthiness computations in [11], [12] don’t consider the influence of both internal social features and neighbors’ trustworthiness status. In order to overcome the above shortcoming, we [46] propose the feature-based first-priority relation graph, say FPRG, of MSNs, in which two vertices (groups) have a link iff they have a first-priority relationship. Moreover, the trustworthiness of a group is evaluated based on the trustworthiness of neighbor groups and the number of malicious users in the group along the FPRG, both of which are predictable. Because the trustworthiness status of a group’s neighbors can tremendously affect the trustworthiness of that group. Current many studies ignore the influence. The trustworthiness computation of a group based on static topology, e.g., hypercube, is relatively simple and efficient. Trustworthiness computations of groups based on the FPRG are accurate because first-priority relationships represent the intensity of connection between groups.

(2) Trustworthiness-based reliable communication: high-level security

A reliable communication should take multiple factors into account, e.g., trustworthiness and path-length. The rationale is that a short route with less malicious uses can provide higher level of security and a better delivery efficiency. In fact, the trustworthiness-hypercube-based routing algorithm as we proposed is reliable, efficient and relatively security.

MSNs have many advantages, one of which is that the deployment cost of MSNs is low, the other is that the user’s physical location is linked with the social interaction. Hence, MSNs without doubt are well-known for their broad development prospects such as social network services, medical services, healthcare services, positioning services and wearable services, and so on. For example, healthcare services can provide information platform for the patient communication at any time, share the symptoms and treatment information [6]. GPS or cellular base station positioning can facilitate users to find friends nearby or a specific infrastructure such as a bank, a restaurant and so on [18]. Routing is the basis of any network technology and information transmission. Hence, the trustworthiness computation and reliable communication are very important to these realistic scenarios.

B. Contributions

Our aim is to design a trustworthiness computation scheme and propose a reliable routing algorithm satisfying aforementioned goals.

We first present a simple and efficient trustworthiness computation scheme to analyze the trustworthiness for each group in malicious mobile social networks. The trustworthiness of a group is calculated by considering the influence with the trustworthiness of neighbor groups and the number of malicious users within group. We believe the proposed scheme provides a comprehensive balance between the trustworthiness of neighbor’s groups and the number of malicious users in the group, and thus it is a promising scheme in trusted system in malicious mobile social networks.

We then focus on reliable communication so that packets can be delivered safely even some users in MSNs are malicious.

We give a novel design of reliable routing based on the trustworthiness and the first-priority relation graph:

  • (1)

    The design can provide a feature-based local optimal reliable path;

  • (2)

    The proposed feature-based local optimal reliable path is of high-level average trustworthiness and relatively small path-length without malicious users.

  • (3)

    The trustworthiness computation and reliable routing algorithm can be extended to the generalized hypercube.

C. Related works

MSNs have received a lot of attentions in the routing field [3], [24], [29], [32], [34], [37], [41], [42], [43]. Song et al. [37] parallelize the influence propagation based on communities and consider the influence propagation crossing communities to improve the performance. Bhosale and Kulkarni [3] propose that location based community greedy algorithm is used to find the influence node based on location and consider the influence propagation within particular area. Sagduyu and Shi [32] firstly obtain analytical formulations for both average delay and success probability as a function of separation between source and destination pairs, and then quantify how mobility and link failure effects decrease routing performance. Trustworthiness computations have also received a lot of attentions in different networks, including delay tolerant networks [8], [10], web spam [22], P2P networks [15], [40], [44], mobile ad hoc networks [13], [19], social networks [1], [4], [7], [21], [38], [39], [47] and mobile social networks [11], [12], [14]. In the following, we only review some results which are most relevant to this paper.

Chen et al. [10] and Cho et al. [13] seek to combine a class of integrated social trust and quality of service (QoS) trust to obtain a composite trust metric in delay tolerant networks and mobile ad hoc networks, respectively. Walter et al. [39] propose a novel trust metric for social networks which is suitable for application to recommender systems. Caverlee et al. [7] point out a user’s trustworthiness should be determined by: (i) the number and trustworthiness of the users who are in a relationship with her; (ii) the relationship quality of each of these users; and (iii) the feedback rating of each user. Varlamis et al. [38] generate personalized content recommendations based on the analysis of implicit of explicit link information between users and user provided content. Guo et al. [21] propose a trust-based privacy-preserving friend recommendation scheme for online social networks (OSNs), where OSN users apply their attributes to find matched friends, and establish social relationships with strangers via a multi-hop trust chain. Chen et al. [11] propose an efficient trust inference mechanism based on fuzzy communities, which we call κ-FuzzyTrust. They propose an algorithm for detection of community structure in complex networks under fuzzy degree κ and construct a fuzzy implicit social graph. They then constructed a mobile social context including static attributes and dynamic behavioural characteristics based on the fuzzy implicit social graph. They infer the trust value between two mobile users using this mobile social context. Chen et al. [12] also describe the implicit social behavioral graph, i.e., ego-i graph which is formed by users’ contacts, and present an algorithm for initiating ego-i graph. They rate these relationships to form a dynamic contact rank, which enables users to evaluate the trust values between users within the context of MSNs. They, then, calculate group-based trust values according to the level of contacts, interaction evolution, and users’ attributes. Based on group-based trust, they obtain a cluster trust by the aggregation of inter group-based trust values. Chatfield et al. [9] propose cover image government surveillance disclosures, bilateral trust and Indonesia–Australia cross-border security cooperation: Social network analysis of Twitter data. Lyu et al. [27] establish efficiently predicting trustworthiness of mobile services based on trust propagation in social networks.

These above trustworthiness computations are dynamic and unstructured, in which the information is hard to collect. Moreover, these trustworthiness computations don’t consider the influence produced by first-priority neighbor groups. Therefore, we propose a simple and effective trustworthiness computation scheme of a group based on the trustworthiness of neighbor groups and the number of malicious users in the group along the FPRG. Moreover, we propose a reliable and efficient trustworthiness-based routing algorithm by applying the trustworthiness of groups and the FPRG in MSNs.

Organization. The remainder of this paper is organized as follows. Section 2 shows the preparatory works. Section 3 proposes the trustworthiness computation method in MSNs. Section 4 describes reliable communication algorithm in MSNs. Section 5 gives scenario simulations for reliable routing in MSNs. Section 6 discusses an extension with the generalized hypercube. We conclude our work in Section 7.

Section snippets

Preliminaries

This section first reviews general definitions of combinatorial network theory and first-priority relation graph. Sub section 2.2 briefly gives the model assumption.

Trustworthiness computations in MSNs

In MSNs, a malicious group can wreak considerable damage and adversely affect the quality and reliability of data (see Fig. 6). Therefore, analyzing the trustworthiness of a group has a positive influence on the confidence with which a group conducts transactions with that group.

Trustworthiness of a user is the level of trust that the trusting entity has in that entity [16]. Trustworthiness means simply that something is worthy of being trusted to satisfy its specified requirements, typically

Trustworthiness-hypercube-based reliable communication

This section describes an efficient design of reliable communication based on trustworthiness and FPRG. The details of our design, together with the performance, are given in the following subsections.

Scenario simulations

In this section, we give two scenario simulations for the reliable routing algorithm in MSNs with malicious users. In this paper, let α=0.5. That is to say, the trustworthiness level and the length of paths are the same important for the feature-based local optimal reliable routing.

Scenario 1. There are 100 mobile users (individuals) in a special MSN (see Fig. 13(a)). Assume there has four key features in the MSNs and the 1st, 2nd feature has three distinct values, the 3rd, 4th feature has two

Extensions

In previous sections, we discussed the trustworthiness and the reliable routing in MSNs based on the hypercube. We extend these trustworthiness and the multi-path reliable routing scheme to the generalized hypercube with multiple distinct values in each dimension without compression.

The trustworthiness of a group in the MSN, in which the first-priority relation graph is the generalized hypercube (GQn). The trustworthiness computation is similar in MSNs based on Qn.

The routing in GQn is exactly

Conclusion

Trustworthiness computations and management are highly challenging issues and exciting fields of research in MSNs due to computational complexity constraints and the independent movement of component users and relationships. The rich literatures growing around trustworthiness give us a strong indication that this is an important area of research. There is no single solution suitable in all contexts and applications. While designing a new trustworthiness system, it is necessary to consider the

Acknowledgment

This work was partly supported by the National Natural Science Foundation of China (Nos. 61072080, U1405255, 61572010), Natural Science Foundation of Fujian Province (Nos. 2013J01221, 2013J01222, 2016J01289), Fujian Normal University Innovative Research Team (No. IRTL1207).

References (47)

  • CaringBridge. [Online]. Available:...
  • ChangM.J. et al.

    Trust-threshold based routing in delay tolerant networks

    IFIP Adv. Inf. Commun. Technol.

    (2011)
  • I.-R. Chen et al.

    Integrated social and QoS trust-based routing in delay tolerant networks

    Wirel. Pers. Commun.

    (2012)
  • J.-H. Cho et al.

    A survey on trust management for mobile ad hoc networks

    IEEE Commun. Surv. Tut.

    (2011)
  • K.S. Cook

    Trust in society

    Russell Sage Foundation Series on Trust, New York

    (2003)
  • H. Elshaafi et al.

    Trustworthiness monitoring of dynamic service compositions

    Proceedings of the 6th Workshop on Emerging Web Services Technology, WEWST 2011, Lugano, Switzerland, September 14

    (2011)
  • K. Fall

    A delay-tolerant network architecture for challenged internets

    Proceedings of the ACM SIGCOMM’03 Aug

    (2003)
  • Google Latitude. [Online]. Available:...
  • K. Govindan et al.

    Trust computations and trust dynamics in mobile adhoc networks: a survey

    IEEE Commun. Surv. Tut.

    (2012)
  • I. Guyon et al.

    An introduction to variable and feature selection

    J. Mach. Learn. Res.

    (2003)
  • GuoL. et al.

    A trust-based privacy-preserving friend recommendation scheme for online social networks

    IEEE Trans. Dependable Secure Comput.

    (2015)
  • Z. Gyöngyi et al.

    Combating web spam with trustrank

    VLDB

    (2004)
  • S. Jain et al.

    Routing in a delay tolerant network

    Proceedings of the ACM SIGCOMM

    (2004)
  • Cited by (20)

    • Generalized fault-tolerance for enhanced hypercubes

      2024, Discrete Applied Mathematics
    • How many triangles and quadrilaterals are there in an n-dimensional augmented cube?

      2019, Theoretical Computer Science
      Citation Excerpt :

      Therefore, selecting an appropriate and adequate topological structure of interconnection networks becomes a critical issue in the field of parallel computing [4]. The hypercube is the most famous interconnection model for parallel computing [3,6–8,14]. In order to reduce the diameter and keep most of the desirable properties of the hypercube, many variants of the hypercube have been proposed as alternative interconnection networks, such as the crossed cube [9,13], the twisted cube [16,18], the locally twisted cube [10,15,17] and the Möbius cube [19,20].

    • RVCCC: A new variational network of cube-connected cycles and its topological properties

      2019, Theoretical Computer Science
      Citation Excerpt :

      The hypercube and its related networks, such as the Twisted cube [4,9,15], Folded hypercube [8], Generalized hypercube [1,16], Hierarchical hypercube [25] and the k-ary n-cube [5,6,24,37], are popular and are excellent choices for the architecture of multipurpose parallel computer systems. Recently, the hypercube network and its topological properties are still widely concerned [7,23,38]. However, the degree of each processor element in the hypercube topology concomitantly increases with its dimension, which renders the hypercube ill-suited for VLSI/UVSI implementation.

    • On exploiting priority relation graph for reliable multi-path communication in mobile social networks

      2019, Information Sciences
      Citation Excerpt :

      To further the above approaches, a priority relationship [20] between two users was defined if they differ in no more than k feature values. We [21] have proposed a reliable communication in mobile social networks based on trustworthiness-hypercube. In this paper, we observe that the PRG has many structural properties, including the nature order, diameter, regular degree and multi-path technology.

    • Social influence modeling using information theory in mobile social networks

      2017, Information Sciences
      Citation Excerpt :

      Mobile social networks [6,19] are the networks where individuals with similar interests converse and connect with one another through their mobile phones and/or tablets.

    View all citing articles on Scopus
    View full text