Resource management of <scp>IoT</scp> edge devices: Challenges, techniques, and solutions

With the growth in the Internet of things (IoT) paradigm, there has been a tremendous makeshift in how the distributed devices work to achieve a common goal. However, it remains essential that all these devices work in a coherent manner to perform a collective action. This makes the task of resource provisioning extremely important in such a paradigm. The end-user level in IoT mostly comprises of low computation and communication powered devices. Improper utilization of the available resources in such a scenario burdens the complete system and degrades the quality of service. In such a scenario, the use of cloud computing techniques can help to manage the resources effectively. More so, with the emergence of relatively newer cloud-based technologies such as edge and fog computing, resource management in the IoT has become far more effective. These technologies bring the computation and communication capabilities closer to the IoT devices where some of the services can be offloaded to the edge devices. These devices are called IoT edge devices and they provide a unique opportunity to tackle some of the existing and pertinent issues for resource management in IoT paradigms; yet at the same time, they face their own set of challenges. However, the use of IoT edge devices in a traditional IoT paradigm results in better utilization of the available resources as well as improving the overall quality of service. Keeping this in mind, this special issue addressed some of the aspects related to resource management in IoT edge devices with the focus on various challenges faced, and potential techniques and solutions to address such challenges by leveraging IoT edge devices. We received numerous submissions in the issue, and we accepted 13 high-quality submissions for publication as a result after following a rigorous review process. Each of the accepted papers is summarized as follows. In the first paper, Khan et al.1 presented “A cache-based approach toward improved scheduling in fog computing” for efficient resource allocation in the fog computing environment, while maintaining the quality of service. The authors use first-in first-out scheme to place the jobs in queue and cache the job type, fog server, arrival time, time to leave, and internal processing time. The jobs are then moved from the queue by the fog broker which selects fog server having sufficient required power and resources to execute the job. The authors' proposed cache-based scheme showed promising results in terms of reducing the execution time, latency, processing delays and power consumption as compared to the conventional first-come-first-serve and shortest job first policies. The second paper on “Extensive review of cloud resource management techniques in industry 4.0: Issue and challenges” by Dewangan et al.2 sheds light on various types of resource provisioning schemes and classified those into different categories (to help understand them better) on the basis of the underlying technique and their overall objective. This survey helps to understand the optimal schemes for catering to different performance metrics such as time, cost, energy, service level of agreement rate, power consumption, resource utilization, etc. Moreover, the authors also highlighted some of the current research challenges in the domain of resource management. The next paper, “An energy efficient and low overhead fault mitigation technique for internet of thing edge devices reliable on-chip communication” by Ibrahim et al.3 presents a coding scheme to make the network-on-chip fault-tolerant. The network-on-chip provides communication backbone in the underlying network for which the proposed scheme handled both single and multibit adjacent bit errors. The next paper in this issue is on “Design and data analytics of electronic human resource management activities through Internet of Things in an organization” by Nasar et al.4 The authors focus on designing a data analytical human resource management system for IoT devices in an organization for ensuring the policies, strategies, and practices within the organization. The activities covered under this improved system include e-recruitment, e-Selection, e-performance management, e-learning, and e-compensation and the performance of the system was validated on four Kaggle databases. In the fifth paper on “A Mobile Data Offloading Framework based on a Combination of Blockchain and Virtual Voting”, Hassija et al.5 enable mobile users to offload computation tasks to resource-rich mobile-devices in order to reduce energy consumption and enhance performance. The authors used directed acyclic graphs (DAGs) for mobile offloading algorithm where the users can securely submit a transaction (powered by blockchain) request for task offloading a DAG, while a game-theoretic scheme was employed in order to model the interactions between various mobile devices for bargaining cost and time. The sixth paper by Lu is on “Security of Internet of Things edge devices”.6 The paper focuses on securing the edge nodes and edge gateways in IoT to meet its future security needs to eliminate the data leakage risk. The edge nodes were optimized by using a cache replacement algorithm, namely Max-PSN and the results illustrate that the proposed mechanism performed superiorly to the lead frequently used and least recently used algorithms with respect to the hit rate and average response speed of centralized and distributed systems. In the seventh paper, Balasubramanian and Jolfaei present “A scalable framework for healthcare monitoring application using the Internet of Medical Things”.7 The authors made use of IoT for providing real-time alarm and assistance in order to ease the activities of pregnant women by merging the advantages of event-driven and assistive care loop framework architecture. In the next paper, Bodkhe and Tanwar shed some light on “Secure data dissemination techniques for IoT applications: Research challenges and opportunities”.8 As the name suggests, the authors presented a comprehensive summary of secure data dissemination schemes present in the existing literature for IoT applications along with their potential research issues and possible countermeasures. The majority of the researched literature in this survey covers the Internet of Vehicles, Internet of Drones, and Internet of Battlefield things with respective open issues and challenges of each of these. As countermeasures, the authors researched opportunities in the directions of the requirement of secure dissemination protocols, efficient data aggregation methods, and cluster-based data dissemination. The ninth paper on “Comparative study of support vector machines and random forests machine learning algorithms on credit operation” by Teles et al.9 compares the support vector machine (SVM) and random forest (RF) scheme for their application to predict financial risks on credit operation. The outcomes of this paper suggest that while both can be effectively used for the specified task, RF has an advantage of the speed and operational simplicity over SVM; while SVM has the benefit of higher classification accuracy. The tenth paper presented by Zhao et al. titled “Message-Sensing Classified Transmission Scheme Based on Mobile Edge Computing in the Internet of Vehicles”.10 The authors make use of mobile edge computing for secure message transmission by prioritizing secure messages using the analytic hierarchy process to guarantee a higher transmission level for urgent messages. Moreover, using the Lagrangian relaxation method, an optimal task offloading model was devised for delay and energy loss by assigning different weight factors to these parameters. The next paper is “FPFTS: A Joint Fuzzy PSO Mobility-aware Approach to Fog Task Scheduling Algorithm for IoT Devices” by Javanmardi et al.11 The authors build a fog task scheduler leveraging the particle swarm optimization along with fuzzy theory to assign tasks of the users to fog devices. The proposed task schedular was tested on iFogSim simulator and results show that it outperformed first-come-first-serve and delay-priority algorithms with respect to delay and network utilization. Zhang et al.,12 in their paper “Service offloading oriented edge server placement in smart farming” made use of the edge resources to support the real-time intelligent controls in smart farming. The authors presented a service offloading oriented architecture for reducing delay in data transmission from sensors to the edge servers while balancing the load on the servers and optimizing the energy consumption. The final accepted paper in this special issue is on “A metaheuristic optimization approach for energy efficiency in the IoT networks” by Iwendi et al.13 The authors proposed a hybrid metaheuristic algorithm, namely, WOA-SA, for optimizing the energy consumption of the sensors in IoT-based wireless sensor networks. The two metaheuristic approaches, namely, whale optimization algorithm and simulated annealing for choosing the cluster heads in order to optimize the energy consumption in the network. The proposed approach was found to be more effective than its counterparts in terms of load, temperature, residual energy, and cost function. We sincerely hope that after reading the accepted contributions in this special issue would help the readers of the journal and a wider research community to gain knowledge on the presented research challenges, techniques and solutions, and encourage them to further work on different aspects of resource management in IoT devices. We thank the editor-in-chief and editorial board members for providing us with the opportunity to conduct a special issue in Software: Practice and Experience. We also like to thank the administrative staff, reviewers and most importantly, the authors, for their help and contributions in successful organization of this issue.