A request dispatching method for efficient use of renewable energy in fog computing environments

https://doi.org/10.1016/j.future.2020.08.035Get rights and content

Highlights

  • A request dispatching controller is designed to shape the fog node electricity demands to match the available green energy supply.

  • The controller jointly minimizes service time and maintains system stability at a satisfactory level.

  • Lyapunov Optimization Technique is leveraged to develop effective online algorithms with guaranteed performance.

  • The simulation results, based on real solar irradiation data, show the superiority of our proposed method over the state-of-the-art method.

Abstract

In the emerging era of Internet of Things (IoT), fog computing plays a critical role in serving delay-sensitive and location-aware applications. As a result, fog nodes are envisioned to be heavily deployed and form future distributed data centers. Powering fog nodes with green energy sources (such as solar and wind), not only helps in environmental and CO2 emission control but also paves the way towards a “sustainable IoT technology”. However, the downside of green energy is its variation and unpredictability, which needs to be engineered. In this paper, we use the Lyapunov optimization technique to derive algorithms for dynamic dispatching of the users’ requests among the nearby fog nodes and remote data centers. The proposed algorithms take into account the time constraints of the requests and maintain the system stability while efficiently utilize the available green energy sources. Exhaustive simulation results, based on solar radiation data supplied by the Australian Bureau of Meteorology, confirm the efficiency of the proposed algorithms. In particular, in terms of service time, the number of deadline misses and green energy utilization, the proposed algorithms outperform the state-of-the-art alternative up to 6%, 17% and 12%, respectively.

Introduction

Fog computing is a new paradigm which brings processing, storage and control to the edge of the network, close to the end devices. As a result, fog computing promotes bandwidth conservation, fast response time, and context-aware applications [1]. Potential benefits of fog computing can be fully harvested through proper resource management which addresses service placement and request dispatching. However, the heterogeneity of the computing nodes and variety in the requests and their requirements in terms of computation, communication and energy make the request dispatching challenging [2], [3].

The fog paradigm introduces a highly distributed platform with a large number of computing nodes. The energy consumption of fog nodes becomes challenging if the nodes are all powered by the centralized power grid [4]. Furthermore, due to the wide distribution of the fog nodes, it is not possible to implement the central, intelligent power management techniques deployed in cloud data centers. In addition, It is impossible to reduce the energy cost by placing the nodes where the cheaper energy is available, because they should be placed somewhere in the user premises. Although the fog paradigm implies a different condition compared with cloud from the energy management point of view, it also brings other new opportunities. Therefore, it requires some new methods of its own based on the fog properties [5], [6].

With respect to cloud, fog nodes consume less power and have smaller footprints [7], which lead them to have a better position in effective utilization of on-site cheap renewable energy (green energy) sources. Recently, renewable energy technology is growing very fast (increasing in performance and efficiency, and decreasing in cost). Therefore, it seems to be feasible and reasonable to utilize on-site renewable sources in fog computing environment [8]. However, the renewable energy is unpredictable and intermittent. For example, solar energy is available during the day and depends on weather conditions. Therefore, to mitigate the amount of produced energy, fog nodes can bank green energy in batteries or on the grid network (called net metering). However, both batteries and net metering accompany some problems [9]. In batteries, self-discharge and internal resistance cause energy losses, chemicals that are used are harmful to the environment, and the purchasing and maintaining (P&M) is costly (the cost of P&M can dominate in a solar system). On the other hand, net metering may not be available in every part of the world, and the voltage transformation while feeding the green energy into the power grid also leads to energy losses. Therefore, we focus on fog nodes powered with grid-tied solar systems having no energy storage. Fog nodes use the grid as a ‘backup’ source of energy when renewable energy is not sufficient.

In this paper, we investigate the problem of managing workload to shape the electricity demands to match the available green energy supply. We state the problem as designing a request dispatching controller. The controller dispatches requests among available computing nodes in fog or cloud tier to jointly minimize service time and maintain green energy utilization and system stability at a satisfactory level.

The controller’s optimum decision making process is first formulated as a stochastic optimization problem. The green energy utilization and selection between fog and cloud are presented as the constraints beside the main objective (i.e., service time minimization). We use the idea of virtual queues [10] and turn the satisfaction of these constraints into a pure stability problem. Then Lyapunov Optimization Technique (LOT) is leveraged to develop effective online algorithms with guaranteed performance.

Our main contributions in this paper are summarized as follow:

  • Utilizing the on-site renewable energy sources, we introduce a dynamic request dispatching strategy to lessen the burden on the power grid while minimizing the service time and stabilizing the queues.

  • Based on the Lyapunov optimization technique and the concept of virtual queues, we introduce easy-to-implement online request dispatching algorithms.

  • The proposed methods are evaluated by extensive simulations based on real solar irradiation data.

  • Analysis is performed to assess the sensitivity of the proposed method to different conditions.

In the remainder of the paper, the related works are summarized in Section 2, the system model is described in Section 3, and Section 4 is devoted to the problem statement. The basics of our methods is explained in Section 5. The proposed methods are introduced in Section 6. Section 7 presents the simulation results and Section 8 provides further discussion. Finally, the paper is concluded in Section 9.

Section snippets

Related works

In this section, some of the most relevant works are presented and discussed. We classify the prior research works into four subsections and separately discuss them.

System model

We consider a fog environment consisting of N heterogeneous fog nodes, a cloud at a remote data center, IoT nodes, and a request dispatching controller. We assume that the fog nodes can be powered with on-site renewable energy sources, the power grid network, or both. In contrast to fog nodes, the only source of energy for the cloud is the grid.

The fog nodes are highly distributed and heterogeneous. Therefore, they have different amounts of computing capability, and the IoT devices may

Problem statements

In this section, we first deal with expressing the green energy utilization constraint satisfaction by defining two lateral constraints. Then, we go through the formulation of the main problem by defining the objective function and considering related constraints.

Queue-stable optimized solution

In this section, we first describe the Lyapunov optimization technique. Then we explain the idea of virtual queues which are used to satisfy lateral constraints of an optimization problem.

Proposed request dispatching algorithms

In this section we employ the LOT to solve the optimization problem P1. We define the virtual queues Zi and G related to the constraints (10a), (10b), respectively.

Corresponding to the constraint (10a) which imposes Ūi(t)ϒ, we define yi(t)=yˆ(ci(t))=Uˆ(ci(t))ϒ to rewrite the constraint in the form of ȳi(t)0. Now, we can define the virtual queue Zi(t) with the updating equation Zi(t+1)=max[Zi(t)+yi(t),0].

The constant value ϒ in the constraint (10a) and in the definition yi(t) is a

Evaluation and simulation results

In this section, we first describe our simulation setup. Then, we present the results of the fixed-parameters simulations, analyzing the effect of control parameter V, and evaluating the scalability of the algorithms. Finally, we go through different scenarios (by changing a specific parameter at each scenario; such as arrival rate, computing and communication demands) to show how the proposed algorithms behave under different conditions.

Discussion

Remark 2

In this work, by leveraging the LOT, we proposed two online algorithms, JoA and GrA, as solutions for request dispatching problem. The methods are independent of the system dynamics and work based on the current system conditions and the queues’ backlog information. The JoA exhaustively searches the requests’ assignment space at each time slot and dispatches all the requests at once (near optimum solution), while GrA greedily dispatches the requests one by one. Request dispatching is a

Conclusions and future work

There are a large number of computing nodes in the fog paradigm, which are geographically distributed over a wide area. Providing the required energy for this large number of nodes becomes challenging if they are all powered by the main grid. Utilizing renewable energy sources emerges as a feasible solution, which not only reduces the burden on the main power grid but also leads to lower CO2 emission and environmental harms.

Considering the on-site renewable energy sources, we investigated the

CRediT authorship contribution statement

Aref Karimiafshar: Conceptualization, Methodology, Software, Validation, Formal analysis, Investigation, Resources, Writing - original draft, Writing - review & editing, Visualization. Massoud Reza Hashemi: Conceptualization, Methodology, Validation, Formal analysis, Resources, Writing - review & editing, Supervision, Project administration. Mohammad Reza Heidarpour: Conceptualization, Methodology, Validation, Formal analysis, Resources, , Writing - review & editing, Supervision. Adel N. Toosi:

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.

Aref Karimiafshar received the M.Sc. degree in computer engineering from Isfahan University of Technology, Isfahan, Iran, in 2013, where he is currently pursuing the Ph.D. degree. His primary research interests include operating systems, cloud/fog/edge computing, and Internet of Things. He is currently working on resource management in cloud/edge computing and integrating renewable energy.

References (36)

  • MarkakisE.K. et al.

    EXEGESIS: Extreme edge resource harvesting for a virtualized fog environment

    IEEE Commun. Mag.

    (2017)
  • GoiriÍ. et al.

    Greenslot: scheduling energy consumption in green datacenters

  • NeelyM.J.

    Stochastic Network Optimization with Application to Communication and Queueing Systems

    (2010)
  • NiL. et al.

    Resource allocation strategy in fog computing based on priced timed petri nets

    IEEE Internet Things J.

    (2017)
  • YousefpourA. et al.

    On reducing IoT service delay via fog offloading

    IEEE Internet Things J.

    (2018)
  • ZhangH. et al.

    Computing resource allocation in three-tier IoT fog networks: A joint optimization approach combining Stackelberg game and matching

    IEEE Internet Things J.

    (2017)
  • DengR. et al.

    Optimal workload allocation in fog-cloud computing toward balanced delay and power consumption

    IEEE Internet Things J.

    (2016)
  • NaranjoP.G.V. et al.

    Design and energy-efficient resource management of virtualized networked Fog architectures for the real-time support of IoT applications

    J. Supercomput.

    (2018)
  • Cited by (5)

    • Energy-efficient and secure mobile fog-based cloud for the Internet of Things

      2022, Future Generation Computer Systems
      Citation Excerpt :

      An increasing energy cost is a potential risk to MFBCs because such a change increases the total cost of ownership and decreases the return on investment of MFBC setups [14]. Decreasing the energy consumption of the MFBC is a challenging and difficult task because emerging computing applications are being developed rapidly; thus, larger servers and disks are required to operate within a specific time period [15]. The reduction in energy consumption is of paramount significance to ensure that the ongoing development of MFBC computing will be workable [16].

    • Elastic Power Utilization in Sustainable Micro Cloud Data Centers

      2023, IEEE Transactions on Sustainable Computing
    • GreenFog: A Framework for Sustainable Fog Computing

      2022, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
    • Intent-based Decentralized Orchestration for Green Energy-Aware Provisioning of Fog-native Workflows

      2022, Proceedings of the 2022 18th International Conference of Network and Service Management: Intelligent Management of Disruptive Network Technologies and Services, CNSM 2022
    • Sustainable Green Fog Computing for Smart Agriculture

      2021, Wireless Personal Communications

    Aref Karimiafshar received the M.Sc. degree in computer engineering from Isfahan University of Technology, Isfahan, Iran, in 2013, where he is currently pursuing the Ph.D. degree. His primary research interests include operating systems, cloud/fog/edge computing, and Internet of Things. He is currently working on resource management in cloud/edge computing and integrating renewable energy.

    Massoud Reza Hashemi received the Ph.D. degree in electrical and computer engineering from the University of Toronto, Canada, in 1998. From 1998 to 1999, he was a Postdoctoral Fellow with the University of Toronto. He was a Founding Member and the Lead Systems Architect with AcceLight Networks in 1999, where he developed some of the key system elements of a multi-terabit multiservice core switch. Since 2003, he has been with the Isfahan University of Technology where he is currently an Associate Professor. As the head of the university IT center from 2005 to 2013 he restructured and consolidated the foundations of IT in the campus, including the IT center, the university campus network and the university data center. From 2016 to 2017, he was a Visiting Scholar with the University of Toronto on sabbatical leave. His current research interests include Software Defined Networks, IoT and Fog Computing.

    Mohammad Reza Heidarpour received B.Sc. And M.Sc. degree in electronics and communication engineering from Isfahan University of Technology, Iran, in 2006 and 2008, respectively, and the Ph.D. degree in electrical engineering from University of Waterloo, Ontario, Canada, in 2013. He has done post-doctoral research at the Coding and Signal Transmission (CST) laboratory in University of Waterloo from May 2013 to November 2015, and at the Electrical and Computer Engineering Department of Isfahan University of Technology from December 2016 to October 2017. He is currently an Assistant Professor in the Department of Electrical and Computer Engineering at Isfahan University of Technology. Dr Heidarpour’s research interests include broad areas of wireless communication, data networks, and network algorithms.

    Adel N. Toosi is a lecturer (a.k.a. Assistant Professor) in the Department of Software Systems and Cybersecurity at Faculty of Information Technology, Monash University, Australia. Before joining Monash, Dr Toosi was a Postdoctoral Research Fellow at the University of Melbourne from 2015 to 2018. He received his Ph.D. degree in 2015 from the School of Computing and Information Systems at the University of Melbourne. His Ph.D. thesis was nominated for CORE John Makepeace Bennett Award for the Australasian Distinguished Doctoral Dissertation and John Melvin Memorial Scholarship for the Best Ph.D. thesis in Engineering. His research interests include scheduling and resource provisioning mechanisms for distributed systems. Dr Toosi’s research interests include Cloud/Fog/Edge Computing, Software Defined Networking, Green Computing and Energy Efficiency. Currently, he is working on green energy harvesting for Edge/Fog computing environments. For further information, please visit his homepage: http://adelnadjarantoosi.info.

    View full text