Elsevier

Computer Communications

Volume 152, 15 February 2020, Pages 187-199
Computer Communications

Fog assisted task allocation and secure deduplication using 2FBO2 and MoWo in cluster-based industrial IoT (IIoT)

https://doi.org/10.1016/j.comcom.2020.01.042Get rights and content

Abstract

Fog-assisted Internet-of-Things have dynamically received interest in the research community. There are more duplicate data transmitted over the Internet due to the rise in IoT devices. In this paper, task allocation and secure deduplication are implemented over four layers of Fog assisted Cluster-based Industrial IoT. IoT device layer used to sense data and mitigate security threats. Devices presented in this layer are registered to the cloud server using Elliptic Curve Cryptography based Hybrid Multiplier. Multi-Objective based Whale Optimization algorithm is accessible for cluster section. SHA-3 is presented in the fog layer for secure data deduplication. ECC HM private key is used for data encryption before transmitted the data. In cloud, layer indexing is constructed using Merkle Hash Tree. This is endeavored to provide query search results for IoT users at the service layer. The simulation results prove enhancement in average latency, user satisfaction, network lifetime, energy consumption, and security strength.

Introduction

Internet of Things have emerged recently and Industrial Sector is one of its best application areas. In an industrial area, deployment of wireless sensor–actuator networks (WSAN), and wireless sensor networks (WSN) help to sense information such as Energy Efficiency, Air Quality Management, Fault Prediction, Resource Prediction, and Product Planning. When data are gathered from industrial sensors, machines, devices, and actuators, the access control of data and sensors generation is assisted via the Internet, such practice is known as industrial IoT [1]. Energy consumption and makespan are important factors considered in Smart Factory based IoT [2], [3]. Fog computing is an emerging paradigm that supports a wide range of applications in real-time. In this model, enormous fog nodes are deployed in various locations. The goal of fog computing model is to reduce the delay for users and the remaining traffic (require a large amount of processing time) is sent to the cloud data centers. Several optimization methods such as weighted sum, hierarchical and trade-off methods have been proposed [4]. Some of the characteristics of fog computing are the following: (1) Heterogeneity, (2) Online analytics, (3) Large Scale IoT applications support, and (4) Easy to interplay with cloud [5]. Healthcare, air pollution monitoring, smart grids, smart homes and smart vehicles are the best example applications that provide advantages via fog computing. Air pollution is the major key factor that affects human life, animals and plants. It happens due to harmful amounts of gases, dust/fumes or odor. Nitrogen dioxide (NO2), Carbon dioxide (CO2), Carbon monoxide (CO), Methane (CH4), Hydrogen sulfide (H2S), Hydrocarbons (Ethanol, Propane, Butane, Isobutane, and Toluene) and Ozone (O3) [6], [7]. In air pollution monitoring systems, there will be a huge amount of duplicate data forwarded to the data processing center. It increases the storage capacity and efficiency of the monitoring system. Hence data deduplication scheme is required in which the redundant or similar data is found out and eliminated. This will improve the storage requirement of the data. IoT devices submit the replicate sensing data that must be eliminated to store in it to the cloud server [8], [9]. On the other hand, task allocation is performed via fog computing for a number of IoT devices. It reduces latency, communication overhead and communication cost for IoT devices [10], [11], [12]. We jointly consider these two issues in this research. Clustering is one of the important processes that supported in fog enabled WSN. Cluster Head (CH) is selected on each cluster, and other nodes in the cluster are Cluster Members (CMs). In [13] prolong stable election protocol (P-SEP) is proposed, which functions are mainly are follows: (i) It controls the random selection of CH for each round, (ii) It exploits the heterogeneity energy threshold to avoid less residual energy nodes, which are nominated as CH in the next rounds, (iii) CH optimizes the minimum distance between the CHs and fog nodes, and (4). It increases energy efficiency for fog nodes and decreases the overhead. To choose the optimum CH. Fog nodes aggregate the sensing data and forward the computing results and then data upload to the cloud servers. Importance of fog computing for task allocation is followed [14], [15]:

Latency: In many real world environments, low latency is a major constraint. To reduce the amount of time for task processing, fog nodes are deployed on the edge of the network.

Data confidentiality: Large amount of data generated and forwarded in a network that cause major attacks. To avoid malicious activity, data encrypted and forwarded to end users or other processing centers.

Indexing: In order to perform data retrieval among end users, efficient indexing is required.

However, mining the data stored by enormous edge servers is difficult in a distributed computing environment. Cloud server or proxy server extracts this information and provide support to on-demand services to end-users. Cloud server outsourcing non-sensitive information to end-users, but it does not guarantee data security and integrity [16]. In the following, we described about the fog assisted cloud environment for IIoT applications.

Today IoT has adopted various applications of the industry. Some examples of IIoT are healthcare service, air pollution monitoring, smart manufacturing, and so on. In healthcare applications, medical sensors are deployed in patients and continuous sensing and transmission of following data such as body temperature, blood pressure, heartbeats rate and breathing activity, which are used to enhance the service quality in the health sector. Fog computing is the new era in IIoT. Fog enabled IIoT would be meaningless without the participation of WSAN and WSN. Fog nodes carried out to reduce the service latency by remote users. Fog nodes and IoT devices are involved with the transmission and deletion of data in the cloud. Integrating IoT, Fog and Cloud have given better-off in current industrial systems. Fig. 1 shows fog assisted cloud of IIoT.

Remote monitoring and control system design for the application is motivated by several types of research. In IIoT each IoT device has a function for both sensing and actuation. In a cloud system, data storage occurs and is comprised of two functioning: visualization of historical information and remote monitoring of sensing environment. In Industry 4.0, fog nodes act as Gateways, which respond faster than cloud servers. This paper is provided with the new innovation of task allocation and secure deduplication in the combination of Industrial IoT, Fog and Cloud environments. Hereby we addressed the issues in this integration [17], [18], [19], [20]. Table 1 shows key features of Fog, IoT and Cloud computing.

The main motivation of this paper is based on research challenges underlying in fog enabled IIoT. IoT devices suffer from large latency from the cloud server. Despite mobile crowdsensing, the cloud server does not provide services to users on time. So our endeavor in this paper, is to design a task allocation and secure deduplication for fog enabled IIoT, using the concept of novel secure and optimization algorithms. Our proposed system architecture of air pollution monitoring in FaCIIoT estimate several air pollutants such as NO2, CO2, SO2, O2, PM2,5, PM10, etc. In this paper, we presented task allocation and secure data deduplication in cluster-based Industrial IoT. The main contributions of this paper are summarized as follows:

  • We firstly register all IoT devices to the cloud server via trusted authority using Elliptic Curve Cryptography (ECC) with Hybrid Multiplier

  • Then we cluster similar IoT devices using the Multi-Objective Whale Optimization (MoWo) algorithm based on node residual energy, node degree (no. of neighbor connections), and distance between nodes.

  • For secure data deduplication, SHA-3 is presented in which hash values are generated for the verification of data deduplication. After hash generation, CH sends Duplicate Service Check (DSC) and Duplicate Service Response (DSR). Then we select the optimal fog node from the CH using 2FBO2 (two fitness based One-to-One Matching) algorithm.

  • Before transmitted to the cloud server, data packets are encrypted using ECC HM private key. Here packets are encrypted using ECC HM algorithm and decrypted at the proxy server.

  • The experimental results show that the proposed scheme is outperforms than the previous works based on the QoS metrics: average latency, energy consumption, user satisfaction, network lifetime and security strength.

Paper Outline: The remaining part of the paper is structured as follows: Section 1 describes the state-of-the-art in the field of fog assisted IIoT. In Section 2, we present the major problem statements. In Section 2.1, we briefly explain our proposed system design and architecture. In addition, describe each new idea is a well-organized way. In Section 3, experimental settings for the proposed system design is presented and also evaluate the comparison between the proposed as well as previous approaches. Finally, the conclusion and future work of this paper are presented and summarized in Section 4. State-of-the-Art, in this section we describe latest literatures that have been done in the fog computing, cloud computing and IoT.

In [21] authors have proposed multi criteria based decision making approach for task allocation in several nodes, which is implemented at the edge nodes. In this paper, tasks implemented at nodes in local manner or presented in peer topology. The proposed decision making scheme were follows two decisions for optimal task allocation. Energy consumption is high due to high latency. In [22] spatial crowdsourcing assisted task owners based task allocation and data aggregation have been proposed through fog computing. Server is enabled to collect sensed information from mobile users. Data aggregation is a specific task, which has drawn much attention in mining massive spatial crowdsensing data. Fog nodes are deployed in several regions, which can assist the server to distribute and aggregate data in a privacy aware manner. Privacy for mobile users are lacking due to inefficient cryptography scheme. In [23] IIoT-based fog computing technology was presented, which was implemented for the smart factory application. In this paper, a hierarchical fog servers based deployment was invoked, which categorizes sensed data into two forms: high priority and low priority. High priority requests are scheduled firstly since it is based on emergency/urgent demands. Furthermore, a workload assignment algorithm was used to offload high traffic load of fog nodes to higher fog tiers. End-to-end delay is large due to large amount of workloads at fog tier. Adaptive configuration of fog nodes were implemented in [24] over IIoT environment. It offers several IoT services such as imminent failure detection, and automatic monitoring control at fog nodes. It improves the performance of industrial systems. Lyapunov optimization and parallel Gibbs sampling methods were proposed in this paper for adaptive fog nodes configuration. However, it is not adopted for any real-time application. In [25] smart resources partitioning was proposed in fog assisted IIoT environment. Firstly authors have exploited Zipf’s law to compute the relationship between popularity ranks of computing control layer from data processing layer of IIoT. Experimentation was conducted to show the successful achievement in terms of response time, response rate and delay time. It causes large complexity due to large number of fog nodes that are geographically distributed in the fog layer. In [26] hybrid approach such as reinforcement learning and fuzzy logic algorithms were proposed to minimize latency for healthcare applications over IoT environment. The reason for this hybrid approach is to assign packets to different processors of virtual machines available for gateway. The limitation of this work is to high service latency in the application layer. Indexing is very important for providing right services based on the given request. In [27] authors have proposed energy efficient indexing scheme over fog environment. This scheme was designed for IIoT environment and provides most appropriate services for users. The index was constructed as a basis of relationship between users. It was effectively considered service retrieval and maintenance for industry based applications. In this work, multi-level indexing was implemented and also performed various indexing operations. This approach may not suitable for dynamic IoT environment and latency is high. Similar to indexing, security is not primarily concentrated on recent works. It is essential for fog enabled cloud environment. In [28] matrix based key agreement and lightweight authentication model was proposed, which focused to make communication through fog computing and also verifies identify of multiple parties. In this work, healthcare data was encrypted and uploaded into cloud server. Encryption and decryption time is high. An intelligent traffic control system application was implemented in fog based security framework. The proposed framework is referred as Intelligent Transportation Control System (FSF-ITLCS) that comprised of department of motor vehicles (DMV), roadside units (RSUs), and vehicles. The proposed addresses various security attacks such as Sybil, DoS, Replay and Impersonation. Overall computation time is high.

Air pollution is a key factor that affects the health life of the Human, Animals and Plants. There are several application areas related to air pollution monitoring include roadside pollution monitoring, industrial perimeter monitoring, site selection for reference monitoring stations, and indoor air quality monitoring [29], [30]. In [31] authors proposed post-process de-duplication method, which was used to determine redundant air pollutant data and removed it efficiently for manage cloud storage systems. This scheme was flexible for data access control and revocation process. A detailed content of all gases present in the air is important for accurate deduplication. With the rapid growth of Urbanization and Industrialization, air pollution monitoring system was presented in WSN. All sensor information is forwarded to Cloud Server (ThingSpeak an Open Source API) [32] for er analysis and computes air quality index (AQI), which is used to visualize location of air quality efficiently. Timely sensing and transmitting framework is required without processing delay in cloud storage. Authors in [33] have proposed air pollution system for monitoring global airs. The main contributions of this paper are three-fold: (1) Determine air pollutants in the given area from Gas Sensors, (2) Design portable interface and user friendly interface i.e. Android Application, in which remote users can access pollution range in the given area. In this work, delay is high when dealing with huge amounts of sensors information and thus fog based routers/edge devices are required to reduce delay and also minimizes computation complexity in cloud computing. In [34] IoT based 3D air quality sensing system was presented, which was designed as a real-time, power efficient and fine grained architecture. It was designed with four layers: sensing layer (data collection), transmission layer (bidirectional communications support), processing layer (data processing and analysis), and finally presentation layer (provide graphical interface for users). However, data security was not considered on such air pollution monitoring applications. In order to protect air quality monitoring system framework, authors in [35] were focused on data integrity and security for low cost air quality sensor, which are used to collect sensors information and manage such pollutants under three cases: Sensor in Physical Possession, Sensor MAC address knowing (geographical information) environment, and Automatic air pollution monitoring in large scale environment. In [36] mobile crowdsensing challenge (increases number of participating mobile users) was tackled through fog computing. User’s task allocation is invoked by the selection of nearest fog nodes and also task assignment is done based on user mobility. Here fog assisted secure data deduplication scheme was proposed which ensures data confidentiality. Fog nodes detect and remove replicate data using BLS-obvious pseudorandom number function and chameleon hash function. These were used to hide user’s information to anonymous mobile users. Currently, air pollution is an emerging topic in IoT, such system sense huge volume of data. For such kind of application, this application may not suitable. Task allocation/assignment does not effective since it find fog nodes based on users local information (preferences and mobility patterns), Boneh–Lynn–Shacham (BLS) -oblivious pseudorandom number function is used for deduplication verification and unauthorized users can easily generate pseudorandom number of deduplication and also it is time consuming process.

In Fog based IoT (sensors, devices) environment, energy efficient cluster-based routing is an essential thing that presented in [37], here authors have proposed a new P-SEP based fog computing model. It follows two procedures such as Fog based Energy Efficient Routing using ant colony optimization (FEAR) and Fog based Energy aware Cloud Routing (FECR). This new P-SEP reduces 9% and 8% of energy usage in FECR and FEAR, respectively. Similarly, 74% of network lifetime is increased by 74% and 83% in FECR and FEAR, respectively. The clustering process is not effective since a node with more residual energy is selected as CHs, and optimal fog node allocation does not investigate and implemented properly because it randomly selects the adjacent fog nodes (the major for invoking fog computing is that reduces processing delay and meets users while request application or services). Adaptive block compressed sensing was proposed in [38], which is based on sensor-cloud data acquisition method over fog environment. Adaptive block compressed sensing is an image compression technique, which was proposed in the lower WSN layer. The drawback of this scheme is large complex and it does not lightweight. In addition, it causes high energy utilization in fog nodes due to virtual clusters formation in lower WSN layer. In [39], fog enabled cloud environment was considered in IIoT. A large amount of data was generated in different sources. However IoT devices are vulnerable and insecure to several threats. AVL tree was constructed in cloud server for indexing different data sources, which causes high query processing costs and consumes more time to rebalance the tree. Secure KNN was proposed to ensure data confidentiality, which is very expensive and thus data searching time is high. Furthermore, KNN is not suitable for dense area and also it does not suitable for processing large amounts of data, particularly in real-world dataset processing.

Our proposed scheme overwhelms the issues that are mentioned in the previous works.

Section snippets

System model

In this section, we present the system model for the proposed task allocation and secure deduplication via FaCIIoT. The proposed system architecture for the operations (task allocation and secure deduplication) is depicted in Fig. 2.

Experimental results and discussion

In this section, we well described the simulation part and also discussed the performances of the proposed scheme using several metrics. Table 2 shows the simulation parameters.

Conclusion and future work

Air pollution is one of the major problems in the industrial sector in which deduplication is the critical factor to minimize storage capacity and latency of cloud server and fog nodes, respectively. The main purpose of this application consideration is that today air pollutants range exceeds its threshold range. Therefore it causes severe health issues for people. To overwhelm this issue, we designed this paper over the Fog assisted Cloud environment for IIoT, which is carried out two

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

References (39)

  • AdhikariM. et al.

    Energy efficient offloading strategy in fog-cloud environment for IoT applications

    Internet Things

    (2019)
  • BinsyM.S. et al.

    User configurable and portable air pollution monitoring system for smart cities using IoT

    Lect. Notes Data Eng. Commun. Technol.

    (2018)
  • BommiR.M. et al.

    A surveilance smart system for air pollution monitoring and management

    Lect. Notes Data Eng. Commun. Technol.

    (2018)
  • ClementsA.L. et al.

    Low-cost air quality monitoring tools: From research to practice (A workshop summary)

    Sensors

    (2017)
  • YanZ. et al.

    Encrypted data management with deduplication in cloud computing

    IEEE Cloud Comput.

    (2016)
  • KolomvatsosK. et al.

    Multi-criteria optimal task allocation at the edge

    Future Gener. Comput. Syst.

    (2018)
  • LiJ. et al.

    Virtual fog: A virtualization enabled fog computing framework for internet of things

    IEEE Internet Things J.

    (2018)
  • NaranjoP.G.V. et al.

    P-SEP: A prolong stable election routing algorithm for energy-limited heterogeneous fog-supported wireless sensor networks

    J. Supercomput.

    (2016)
  • MoysiadisV. et al.

    Towards distributed data management in fog computing

    Wirel. Commun. Mobile Comput.

    (2018)
  • Cited by (27)

    • S-FoS: A secure workflow scheduling approach for performance optimization in SDN-based IoT-Fog networks

      2023, Journal of Information Security and Applications
      Citation Excerpt :

      As a result, their protection is based on different levels of IDSs that are chosen based on probability. Shivi et al. [9] explained the usage of encryption to ensure the safety of the data in IoT. In another paper, Baniata et al. [36] employed blockchain to enable identity verification and privacy for IoT applications.

    • Fog-GMFA-DRL: Enhanced deep reinforcement learning with hybrid grey wolf and modified moth flame optimization to enhance the load balancing in the fog-IoT environment

      2022, Advances in Engineering Software
      Citation Excerpt :

      These characteristics can interact and inhibit one another [10]. The fog computing paradigm aims to decrease user delays, with the remaining traffic (which needs a lot of processing time) being transmitted to cloud data centers [4]. Also, fog computing is used in industrial IoT, the Internet of vehicles, and 5G heterogeneous networks [7].

    • Towards edge computing in intelligent manufacturing: Past, present and future

      2022, Journal of Manufacturing Systems
      Citation Excerpt :

      The ciphertext of the data key is re-encrypted at data deletion, which is not decryptable by the user’s private key, making the data irrecoverable. Fog layer can perform secure deduplication as Sharma et al. [125] implemented in a four-layer cluster-based IIoT. All the edge devices are registered to the cloud server using a Cryptography-based Hybrid Multiplier, and data is encrypted before transmitting.

    View all citing articles on Scopus
    View full text