Proactive caching for edge computing-enabled industrial mobile wireless networks

https://doi.org/10.1016/j.future.2018.06.017Get rights and content

Highlights

  • Considering mobility, a three-layer caching architecture in IMWN is proposed.

  • The sojourn time and effective link time of mobile node are determined.

  • For delivering large amount of data, the optimal caching strategies are proposed.

  • The performances of the proposal are analyzed via a simulation and comparisons.

Abstract

As manufacturing systems shift from automated patterns to smart frameworks such as smart factories in Industry 4.0, industrial wireless networks (IWNs) are serving as promising communication systems that can be applied to the manufacturing field. When the mobile elements and static nodes are introduced into the system, large amounts of data downloaded from mobile networks or tele-servers can be one of the greatest challenges for industrial mobile wireless networks (IMWNs). Mobility and industrial properties have rarely been considered by previous research on download strategies and caching methods. In this paper, we present a three-layer cache architecture based on edge computing and other heritage traditional networks. Then, useful spatial and temporal mobility properties are mapped using different groups and edge computing servers that contain mobile nodes. Then, according to the sojourn time, the capacity of edge computing servers and other neighbouring nodes, we propose a proactive caching strategy for large amounts of data downloaded by mobile networks that considers location and mobile trajectories. Moreover, the superiority of our proposed scheme is demonstrated by comparison case studies of widely used classical schemes. The numerical results show that our proposed strategy achieves higher goodput and real-time and other performance.

Introduction

With the increasing progress of information and communication technologies (ICTs) [[1], [2], [3], [4]], growing ICTs are constantly introduced into multiple industry domains. This trend is more prominent in the manufacturing industry. Manufacturing systems are currently shifting from automated patterns to smart frameworks such as smart factories in Industry 4.0 and are further propelled by popular trends [[5], [6], [7], [8]]. In addition, mobile nodes (MNs) are undergoing rapid growth in smart factories, arising from the flexibility, mobility and extendibility of the technology [[9], [10]]. Industrial mobile wireless networks (IMWNs) bridge smart heterogeneous equipment and create a communication channel among users and manufacturers. Because ubiquitous communications exist in the novel system, mobile data traffic is an active research topic and the main challenge associated with IMWNs. Big data or large amounts of data transmission are gradually magnifying this challenge [[11], [12]]. One main effort to meet such a strong demand is to deploy additional access points or base stations. Although this method can increase bandwidth and access choices, it results in heavy traffic to the backbone of the work, particularly between edge nodes and tele-servers. Therefore, traditional methods are not effective to meet manufacturing industry demands. Edge computing and pre-caching techniques [13] offer a potential solution.

Network computing is migrating from cloud computing to fog or edge computing [14], particularly for the Internet of things, which includes numerous edge nodes. Since edge computing allows more accurate structures, this framework can be merged into the mobile network and form mobile edge computing (MEC) systems. These systems have drawn relatively strong attention from academia and industry. Certain studies [[15], [16]] have focused on network computing resource deployment, edge computing or providing the computing service for edge nodes. However, only rarely have studies addressed caching or data storage based on MEC for accommodating big data transmission. MEC provides an alternative option to accommodate large amounts of data communication. The key research issue in caching storage is finding fast wireless networks. Popular content and quality experiences have been developed; the associated methods are provided in [[17], [18], [19]]. Although the IMWN can benefit from these previous studies, the studies can also take on certain new features. Most previous research has focused on static networks and has neglected mobility and industry properties, which are not suitable for the industry scenario.

There are several reasons for this study [[20], [21], [22]]. Firstly, there are great requirements for distributing data from cloud to node. Secondly, moving nodes increase the complexity of the data distribution for dynamic changes of network topology. Thirdly, in industrial wireless networks, real-time data distribution is a key assessment criterion for industrial devices. Most previous works focused on data transmission, but these studies have to the constraints such as communication rate, energy consumption. Proactive data caching technology provided a new option. Distinct from the previous research, we consider mobility and industrial properties to propose a three-layer caching architecture in IMWN. Then, based on MEC, we present a model that can determine the optimal caching strategies of a large amount of data downloaded from tele-servers. Our strategies can be divided into three stages. First, we map the MEC server and neighbouring nodes according to their mobile trajectories. Second, the sojourn time and effective link time are determined. Third, the finishing layer caching is completed. Then, we analyse the performance of the proposal via a simulation and compare the results to the results of other methods.

The remainder of the paper is structured as follows. In Section 2, we briefly review and discuss the related works for edge computing and caching. Then, the three-layer storage architecture and the outline of our strategies are provided in Section 3. Section 4 describes the system model and the solution procedure of the proposed schemes. The performance evaluation and conclusions are presented in Sections 5 Based on mobile position and data download strategy, 6 Experiments and results, respectively.

Section snippets

Related works

In the section, we briefly review the related works about edge computing and caching methods.

Industrial wireless network (IWN): Critical time or real-time data delivery for IWNs has been an active research direction in recent years. A brief survey of previous work can be found in [[23], [24]]. Specifically, in [25], a novel routing protocol that achieves high-quality real-time performance was proposed for uploading data from nodes to a base station; this approach estimates the communication

Architecture of industrial MEC and strategy

In this section, in accordance with manufacturing system properties and previous work on the structure of IWNs, we first construct a single-hop clustered industrial wireless mobile network. Then, we present a three-layer storage system based on cloud and edge computing in the industrial domain for the delivery of large amounts of data. Moreover, we briefly review our strategies for the different stages of big data traffic for working MNs.

Fig. 1 demonstrates the scene of a smart factory, where

Proactive cache optimization problem

In this section, we present the system model of IMWNs for large amounts of data and any related assumptions considered in this study. We can compute the transmission time of the big data in an MN and in the corresponding constraints via the proposed model.

Network coverage model : We consider an IMWN with multiple equipment. The physical network devices include multiple cluster heads with CESs, static nodes (SNs) (e.g., sensors), fixed machines, MNs and a TCS. Let L denote the path of the MNs.

Based on mobile position and data download strategy

Considering the previous architecture of the models and formulation, our proactive cache approach for MNs in IWN edge computing includes two stages: a proactive cache is selected to store large amounts of data, and the steps to fetch the data for the MN are provided. Specifically, the sub-problem in Eq. (17) must be efficiently solved. We develop an efficient coverage segmentation algorithm to solve the second problem in Eq. (22), and a distributed improved Hungarian algorithm is presented.

Experiments and results

In this section, the numerical results of the proposed algorithm are presented. First, we describe the evaluation methodology that includes the simulation configuration, performance metrics and emulation parameters. Then, we discuss and present the simulation results in detail.

Conclusions

In this paper, we have studied big data distribution from cloud server to wireless moving nodes, based on industrial MN properties and MN distributed caches for edge computing in IMWNs. Firstly, a three-layered storage system in an edge computing framework is proposed. Then, we analysed the MNs and edge cluster network properties. The employed spatial and temporal mobility properties are mapped for different groups and for the CES of the MNs. We divided big data distribution into data cache and

Acknowledgements

This work was supported in part by the National Key Research and Development Program of China (No. 2017YFE0101000), the Fundamental Research Funds for the Central Universities (No. x2jqD2170480), the Science and Technology Program of Guangzhou, China (No. 201802030005), and the Key Program of Natural Science Foundation of Guangdong Province (No. 2017B030311008).

Xiaomin Li is working towards the Ph.D. degree of mechanical and electrical engineering in South China University of Technology. His research interests are in wireless sensor networks, industrial wireless network, data mining.

References (40)

  • HassanM.M. et al.

    Resource provisioning for cloud-assisted software defined wireless sensor network

    IEEE Sens. J.

    (2016)
  • KimB. et al.

    AWNIS: Energy-efficient adaptive wireless network interface selection for industrial mobile devices

    IEEE Trans. Ind. Inform.

    (2013)
  • RahmanS.M.M. et al.

    Privacy preserving secure data exchange in mobile P2P cloud healthcare environment

    Peer-to-Peer Netw. Appl.

    (2016)
  • BiS. et al.

    Wireless communications in the era of big data

    IEEE Commun. Mag.

    (2015)
  • ZhangY. et al.

    CADRE: Cloud-assisted drug recommendation service for online pharmacies

    Mobile Netw. Appl.

    (2015)
  • LiQ. et al.

    Cooperative edge caching in software-defined hyper-cellular networks

    IEEE J. Sel. Areas Commun.

    (2017)
  • LiX. et al.

    Adaptive transmission optimization in SDN-based industrial internet of things with edge computing

    IEEE Internet Things J.

    (2018)
  • GirolamiM. et al.

    Human-enabled edge computing: Exploiting the crowd as a dynamic extension of mobile edge computing

    IEEE Commun. Mag.

    (2017)
  • TandonR. et al.

    Harnessing cloud and edge synergies: toward an information theory of fog radio access networks

    IEEE Commun. Mag.

    (2016)
  • LiC. et al.

    QoE-driven mobile edge caching placement for adaptive video streaming

    IEEE Trans. Multimed.

    (2017)
  • Cited by (26)

    • Geometric-thermal error control system for gear profile grinding machine

      2022, Advanced Engineering Informatics
      Citation Excerpt :

      The cloud computing provides powerful computing power and storage capability [42–44]. The edge computing [45–49] and fog computing [50–54] have emerged in recent years. The edge computing and fog computing can share part of the computing tasks, and then relieve the computing pressure of the cloud computing layer, and then the decisions are made more agilely and efficiently.

    • Improving distributed anti-flocking algorithm for dynamic coverage of mobile wireless networks with obstacle avoidance

      2021, Knowledge-Based Systems
      Citation Excerpt :

      Nowadays, mobile wireless networks [1] are playing a vital role in our world and are widely used in many kinds of cases and applications, such as environment monitoring [2], communication [3], and anti-intrusion detecting [4].

    View all citing articles on Scopus

    Xiaomin Li is working towards the Ph.D. degree of mechanical and electrical engineering in South China University of Technology. His research interests are in wireless sensor networks, industrial wireless network, data mining.

    Jiafu Wan is a Professor in School of Mechanical and Automotive Engineering at South China University of Technology, China. His research interests include Industry 4.0, industrial wireless networks, cyber–physical systems, internet of things, cloud computing, embedded systems, and industrial robotics. He has directed 17 research projects, including the National Key Research and Development Program of China, the National Natural Science Foundation of China, the High-level Talent Project of Guangdong Province, and the Key Program of Natural Science Foundation of Guangdong Province. Thus far, he has published more than 150 scientific papers, including 90+ SCI-indexed papers, 30+ IEEE Trans./Journal papers, 14 ESI Highly Cited Papers and 4 ESI Hot Papers. His research results have been published in some famous IEEE Journals & Magazines, such as IEEE Transactions on Industrial Informatics, IEEE/ASME Transactions on Mechatronics, IEEE Communications Surveys and Tutorials, IEEE Communications Magazine, IEEE Transactions on Intelligent Transportation Systems, IEEE Internet of Things Journal, IEEE Network, IEEE Wireless Communications, IEEE Systems Journal, and IEEE Sensors Journal. According to Google Scholar Citations, his published work has been cited more than 4850 times (H-index = 35). He is an Associate Editor for IEEE Access (SCI) and Editorial Board of PLOS ONE (SCI) and Computers & Electrical Engineering (SCI), and he is a Managing Editor for IJAACS (Ei Compendex) and IJART (Ei Compendex). He is a Leading Guest Editor for several SCI-indexed journals, such as IEEE Systems Journal, IEEE Access, Elsevier Computer Networks, Mobile Networks & Applications, Computers and Electrical Engineering, Wireless Communications and Mobile Computing, and Microprocessors and Microsystems.

    View full text