Toward better data veracity in mobile cloud computing: A context-aware and incentive-based reputation mechanism
Introduction
Mobile Cloud Computing (MCC) combines cloud computing and mobile computing to provide mobile users with data storage and processing services in clouds, such as Amazon, Google AppEngine and Microsoft Azure, that perform resource-intensive computing [31], [32]. MCC is a highly promising technology trend for the future of mobile computing, and for this reason, there has been a phenomenal burst of research activities in MCC. Although MCC has attracted significant research and development efforts, there are salient open issues and challenges in the area of security and trust in MCC, which is an essential factor for the success of the burgeoning MCC paradigm [1], [26], [27], [31], [32].
MCC enables the large-scale collection and processing of personal private data such as individuals’ locations and electronic medical records [5], [6], [26], [32]. The processing of this information using big data analytics has become a hot topic in MCC. In order to avoid making decisions based on the analysis of uncertain and imprecise data, it is crucial to maintain a high level of data veracity, which is often overlooked. But it is just as important as the other three V's of Big Data: Volume, Velocity and Variety. Data veracity includes two aspects: data certainty defined by their statistical reliability; and data trustworthiness defined by a number of factors including data origin, collection and processing methods, such as trusted infrastructure and facility [29]. Thus, apart from data confidentiality and privacy, data provenance must be certified and data must be accurate, complete and up-do-date as well [3], [25]. Since ubiquitous access to the Internet in MCC exposes critical data and privacy information to new security threats, a number of research works have been focused on security and trust to cope with these new threats and enhance the data veracity in MCC.
Data veracity shows how much the data used are trusted, authentic and protected from unauthorized access and modification. There are many security challenges in data veracity such as external denial-of-service, credential stealing, remote code injection, data integrity attacks, internal attacks, and supply chain attacks [10]. Consequently, the availability, confidentiality, and integrity of both the original data and the data analytics results are threatened by these attacks, e.g., the degraded availability of a big data system, the compromised confidentiality of the data and analytics, and the violated integrity of the data and analytic results.
As an effort to tackle the aforementioned challenges, this paper focuses on the aspect of data trustworthiness to enhance data veracity through designing a reputation mechanism to defend against internal attacks in MCC. A new Category-based Context aware and Recommendation incentive reputation Mechanism (CCRM) is proposed, which incorporates innovative approaches in terms of data categories, context sensing, security relevance and recommendation incentive. To the best of our knowledge, our work is one of the first to describe this front of data veracity in MCC. The major contributions of this work include the following:
- (1)
This paper proposes a new Category-based Context aware Reputation Mechanism (CCRM) to defend against the internal threats for enhancing data veracity in MCC.
- (2)
The CCRM incorporates three key innovations: a Vickrey-Clark-Groves (VCG)-based distributed cheat-proof recommendation incentive scheme, a security level-based data category method, and a user context sensing technology.
- (3)
Extensive OPNET simulation experiments demonstrate that the CCRM improves the performance of the reputation mechanism compared to the state-of-the-art including the RP-CRM [14], ARTSense [23] and Harmony [22] mechanisms. The CCRM can effectively defend against internal collusion attacks and bad mouthing attacks to enhance data veracity in MCC.
The remainder of this paper is organized as follows. Section 2 presents a brief review of related work, Section 3 describes network and adversary models, Section 4 introduces the implementation details of the CCRM, Section 5 analyzes the cost and evaluate the performance of the CCRM. Finally, Section 6 presents the paper's conclusions .
Section snippets
Related work
Data veracity is becoming a research hotspot of big data and there have been many related studies in the literature [2], [4], [10], [16], [20], [15]. For example, Kepner et al. [10] introduced a new technique called Computing on Masked Data (CMD) to improve data veracity while allowing a wide range of computations and queries to be performed with low overhead by combining efficient cryptographic encryption methods with an associative array representation of big data. Bodnar et al. [4] proposed
Network model
In this paper, we focus on the network environment of MCC, which is considered to be a viable solution to implement fast, large-scale big data applications [11], [34]. A typical MCC architecture, which consists of a mobile client network, a wireless mesh backbone network and a cloud service platform, is depicted in Fig. 1. The mobile client is connected to the base transceiver station (BTS) and accesses the mesh backbone via the mesh router; these are linked to each other and communicate with
Category-based Context-aware Reputation Mechanism (CCRM)
In this section, we elaborate on the proposed Category-based Context-aware Reputation Mechanism (CCRM), which integrates the reputation evaluation with data category [17], [23], context-awareness technologies [19] and the VCG mechanism [7], [18], [24] to defend against the insider threat and enhance the data veracity in MCC. The CCRM is implemented in both mobile clients and cloud service providers to perform bidirectional reputation evaluation. The CCRM includes three phases: direct reputation
Cost analysis and performance evaluation
In the section, we first elaborate on the communication cost and the computation complexity, and then present the performance evaluation of the proposed reputation mechanism CCRM.
Conclusions
In this paper, we investigated the problem of protecting against internal attacks for enhancing data veracity in Mobile Cloud Computing (MCC). A new category-based context aware and recommendation incentive reputation mechanism named CCRM has been proposed, which incorporates innovative technologies in terms of data categories, context sensing, security relevance and recommendation incentive. The simulation-based experiments and performance analysis have verified that the CCRM is effective and
Acknowledgements
This work is supported by the National Natural Science Foundation of China (61363068, 61472083, 61671360), the Pilot Project of Fujian Province (formal industry key project) (2016Y0031), the Foundation of Science and Technology on Information Assurance Laboratory (KJ-14-109) and the Fujian Provincial Key Laboratory of Network Security and Cryptology Research Fund.
References (34)
- et al.
Security in cloud computing: opportunities and challenges
Inf. Sci.
(2015) - et al.
Addressing data veracity in big data applications
Data security – challenges and research opportunities
Secure Data Management
(2014)- et al.
Increasing the veracity of event detection on social media networks through user trust modeling
- et al.
κ -Fuzzytrust: efficient trust computation for large-scale mobile social networks using a fuzzy implicit social graph
Inf. Sci.
(2015) - et al.
A partnership-based approach to improve QoS on federated computing infrastructures
Inf. Sci.
(2016) - et al.
Truthful mechanisms for secure communication in wireless cooperative system
IEEE Trans. Wireless Commun.
(2013) - et al.
A trust management system for ad-hoc mobile clouds
- et al.
Variable rate variable power MQAM for fading channels
IEEE Trans. Commun.
(1997) - et al.
Computing on masked data: a high performance method for improving big data veracity