FOG COMPUTING ENABLED INTERNET OF EVERYTHING

Zhaolong Ning

Lei Guo

Joel Rodrigues

Mohammad S.Obaidat

With the rapid development of ubiquitous networks and smart cities,the connection and communication of Internet of Everything (IoE) have drawn great attention from both academia and industry.The main challenge for constructing IoE is to enable real-time communication and high-efficiency computing among mobile devices.Mobile fog computing is promising to lower communication delay and offload network traffic.However,how to realize fog-enabled communication and computing in IoE with high-dynamic and heterogeneous network characters has not been fully investigated.Furthermore,deployment and reliable communications among fog nodes are also challenging.

The purpose of this special issue is to take the opportunity to introduce the current developments,progress,and advancements of the technical elements of IoE combined with fog computing,from both theoretical and practical perspectives.We focus on architecture and framework establishment,real-time communications and protocols,resource allocation and management,network virtualization technologies in fog computing enabled IoE and so on.

Eight papers are selected from the submissions in this special issue,covering both theoretical analyses and practical systems.Details are illustrated as follows:

Shifting from traditional cloud computing to fog computing models has shown some significant benefits in 5G and Internet of Things (IoT) based networks.These benefits include better network control,storage for latency critical applications,and security improvement for the processing of data at edge devices.However,there are some challenges such as users' privacy,resource allocation,unavailability of testing software and programming models,which need to be addressed before the largescale implementation of fog computing models.Motivated by these observations,Aliet al.discuss a three-tier architecture of fog networks in detail along with its standardization problems.A swap matching based algorithm is proposed for efficient resource allocation in end IoT devices of fog networks.

With the increase of mobile communication equipment and service demands,the high-efficiency and low-latency service response is critical to the development of smart cities.Fog radio access networks (F-RAN) have drawn universal attention from academia and industry because of their advantages,such as low communication delay,high frequency spectrum utilization,flexible networking and wide coverages.Meanwhile,due to the scarcity of wireless resources and the serious waste phenomenon of resources,it is especially important to develop an effective wireless resource management mechanism.By the considerations mentioned above,Yuet al.design an F-RAN based framework,focusing on dynamic coverages of base stations and user requirements.Furthermore,the authors extend the application of game theory models,and propose a spectrum pricing and allocation scheme based on a Stackelberg game model.The designed method can realize the allocation of resources on demand,avoid waste of resources,and maximize the revenue of operators.At the same time,the authors use spectrum reuse technologies to improve the efficiency of network resource usage.By analyzing the simulation results,the effectiveness of the proposed wireless resource management scheme can be verified.

With the rise of IoT and artificial intelligence,fog computing has been proposed to fulfill the low-latency requirements of latency-sensitive applications and reduce the computing pressure of the cloud.Intuitively,the resource allocation of fog computing is a key issue and deserves to be explored.However,online allocation can enhance the utilization of resources and save the cost used to maintain fog nodes with respect to the problem of fog nodes' states,when the extra cost can be generated by changing the state from off to on.The paper,entitled “Game-Theoretic Online Resource Allocation Scheme on Fog Computing for Mobile Multimedia Users”,designs a resource allocation scheme for fog computing based on the Stackelberg Game.As the resources' owner - the scheduler's goal is to minimize the cost of operations under the premise of satisfying multimedia media users' needs.In order to obtain the required quality of experience,each multimedia media user's goal is to minimize the response time of the task.We study the allocation of fog resources under the multimedia background.Specifically,three cases including the single-round of static requests,multiple rounds of future predictable requests,and multiple rounds of future unpredictable requests are discussed in detail and related experiments are designed.Experimental results show that with the increasing number of calculation rounds,the overall operating costs of future predictable requests in multiple rounds are obviously less than that of static requests in the single-round.

More and more applications are being developed in the IoE era.However,the device resource is generally constrained to fulfill all the network requirements.The paper entitled “Energy-efficient computation offloading and resource allocation in fog computing for Internet of Everything” works on the computation offloading problem.Authors focus on the resource synergy problem between fog computing and cloud computing,and jointly consider offloading decisions,computation resource allocation and transmission power.Specifically,an energy-efficient computation offloading and resource allocation scheme is presented for the sake of minimizing the system cost.Performance evaluation verifies the effectiveness of the presented method.

With the growing trend of physical objects connected to the Internet (i.e.,IoE),our daily life is more involving with service which utilizes real time generated sensor data collected from physical objects.Therefore,a search engine for IoE which dynamically integrates the sensor information plays an important role for IoE data service.However,compared to the traditional search engine which conducts data statistics,a search engine for IoE needs to process a huge amount of streaming data generated by the live sensors.Such processing turns out to be the crucial point to meet the requirement of real time smart device data collection and searching.In order to improve the performance of IoE data integration,Suet al.propose a search framework for IoE based on a multi-storage of sensor data and data statistics results.A prediction model is presented for dynamic sensor data behavior prediction so that the designed engine does not need to wait for the collected data to generate the final search results,making it suitable for the online usage.The effectiveness of the designed dynamic prediction model is demonstrated by a public sensor dataset from Intel Lab.

The computing issue for contact plan designs in the spatial-node-based IoE is studied by Daiet al.to realize high-efficiency data transmission and computing,wide-area coverages and connectivity.First,a three-layer space communication system model is constructed,consisting satellites,high altitude platforms,and ground stations.After that,two heuristic computing methods are proposed by comprehensively considering the time-varying topology,the intermittent connectivity,and the adaptive transmission feature.For the first algorithm,the contact plan is designed under the limited-resource condition,and the search scope can be adjusted dynamically to ensure the continuous searching capability through the population-based iterative computing.For the second algorithm,the contact plan decisions are made based on the current state,and the strategies of optimal matching and load balancing are designed with the greedy algorithm and the state-based computing method.Simulation results show that the proposed computing methods outperform the existing solutions in terms of fitness and delivery time and the overall delay.

With the objective of accommodating the asymmetric characteristic of Internet traffic,Penget al.present a cloud-fog architecture based on multi-flow optical transponders.Its advantages are:1) flexible adjustment for downstream or upstream traffic transmission; 2) multiple sub-channels can use flexible modulation formats to serve traffic demands.The optimization objectives are both minimizing the required number of sub-channels and maximizing the volume of traffic transmission.Numerical results demonstrate the effectiveness of the constructed architecture.

Advanced vehicular applications challenge the computational and communication demands of vehicles.A fog-cloud computational offloading scheme is designed by Wanget al.in Internet of Vehicles (IoV) to minimize the power consumption and computational facilities of vehicles.The formulated offloading problem is NP-hard.After that,a predictive combination transmission mode is designed for vehicles,and a deep learning based model is established to obtain the optimal workload allocation.Simulation results show that the presented method can increase energy efficiency and lower network latency.

We would like to thank the anonymous reviewers for their great efforts in reviewing the submitted manuscripts.We would like to thank the editors of this special issue.