Elsevier

Computer Networks

Volume 192, 19 June 2021, 108100
Computer Networks

Base Station switching and edge caching optimisation in high energy-efficiency wireless access network

https://doi.org/10.1016/j.comnet.2021.108100Get rights and content

Abstract

The improvement of the energy efficiency and the reduction of the latency are two of the main goals of the next generation of Radio Access Networks (RANs). In order to achieve the latter, Multi-access Edge Computing (MEC) is emerging as a promising solution: it consists of the placement of computing and storage servers, directly at each Base Station (BS) of these networks. For the RAN energy efficiency, the dynamic activation of the BSs is considered an effective approach. In this paper, the caching feature of the MEC paradigm is considered in a portion of an heterogeneous RAN, powered by a renewable energy generator system, energy batteries and the power grid, where micro cell BSs are deactivated in case of renewable energy shortage. The performance of the caching in the RAN is analysed through simulations for different traffic characteristics, as well as for different capacity of the caches and different spread of it. New user association policies are proposed, in order to totally exploit the MEC technology and reduce the network energy consumption. Simulation results reveal that, thanks to this technology and the proposed methodologies, the experienced delay and the energy consumption drop, respectively, up to 60% and 40%.

Introduction

The Multi-access Edge Computing (MEC) technology, also known as Mobile Edge Computing, has been introduced to push computing and storage resources in physical proximity of end users, placing servers on the edges of the network [1]. In this way, the execution of applications, the pre-processing of data and the caching of popular contents are performed in proximity of end users. In Radio Access Networks (RANs), these servers are co-located on Base Stations (BSs). Therefore, these infrastructures are able to provide storage and computation services, in addition to access services [2]. Several benefits are derived from the introduction of this technology [3], [4]. First, the backhaul traffic load is reduced, since the access to the cloud is unneeded. Second, the quality of the multimedia content can be adapted to the user’s channel. Finally, as proved in [5], [6], [7], the latency is reduced, which is one of the objectives of the 5G. Because of these advantages, this technology has been largely investigated in literature. Many works focus on the optimisation file placement in caches located on BSs of RANs, with the goal of minimising the content delivery delay [8], [9], [10].

Another key goal of the 5th generation of networks, is the improvement of the energy efficiency [11]. Indeed, 5G systems aim at consuming a fraction of the energy consumption of 4G mobile networks, even if the amount of traffic which 5G networks are supposed to manage is much larger than in 4G ones, as highlighted in [12]. According to [12], the mobile IP traffic will reach 77.5 exabyte (EB) per month by 2022, which is an enormous increase compared to 11.5 EB per month in 2017. For this reason, the capacity of the 5G networks, is expected to increase by a factor 1000 more than 4G networks, supporting up to 9 billion of mobile devices, and an heterogeneous range of applications, services and devices [13]. The network energy efficiency has been recognised as a fundamental and urgent aspect of the communication community, since 80% of the total mobile network is consumed by mobile access equipment. As reported in [13], from this 80%, 90% is consumed by the BSs of these networks, whose energy consumption is an important actor of the Operational Expenditure (OPEX), which would grow because of the RAN densification, planned with the 5G RAN deployment [14], [15]. In addition to this, the increase of the RAN energy consumption will contribute to the increase of the carbon emissions, generated during the energy production, which significantly contribute to the climate changes. For this reason, the design of energy efficient RANs has been receiving a lot of attention for many years. The European Commission, in [16], under the need for actions to improve the energy efficiency in communications, formalises a policy that regulates the energy consumption and carbon emissions for Broadband Communication Equipment. Meanwhile, in literature, many works address this issue through the dynamic allocation of the network resources [17], [18], [19], [20], [21]. Indeed, the typical behaviour of the daily traffic demand presents short peaks and long valleys, during which the capacity of the RAN is under-utilised, since the traffic demand is very low. Therefore, during these periods, the unneeded capacity is deactivated, allowing energy saving [17]. Another trend proposes local Renewable Energy Sources (RES) as power supply of RANs, e.g. a wind turbine and/or a Photovoltaic (PV) panel system. This makes these networks self-sufficient and more sustainable, since the amount of energy that is produced by burning fossil fuels reduces [22]. Recently, these two approaches have been combined, so that the BSs of the RAN, supplied by RES, are dynamically switched to sleep mode, when the traffic demand is low, as in [18], or when the amount of renewable energy that is generated by RES is not enough to power the RAN [19].

The energy efficiency in RAN and the employment of the MEC technology in these networks have been largely investigated in the literature, but the impact of the MEC technology employment on the network energy efficiency is usually neglected. Meanwhile, the effect of the BSs switching on the MEC technology performance is ignored, as well. Indeed, these two topics are typically considered separately and their coexistence has not been investigated yet. For this reason, in our previous work, presented in [23], the simultaneous employment of the MEC technology and BSs switching is considered, providing an overview of their mutual effects. The growth of the energy consumption due to the supply of the MEC servers installed on each BS is analysed, as well as the effects of the BSs switching, and consequently of the MEC server deactivation, on the experienced delay. In this paper, we deepen the analysis of the energy saving strategy and we use association procedures, which aim at further improving the network energy efficiency, as well as the latency reduction. The MEC switching is also introduced to guarantee the load balancing among BSs. To do this, we considered a portion of an heterogeneous RAN in the city centre of Ghent, in Belgium, which dynamically adapts its capacity to the traffic demand. It is composed of a set of macro cell BSs, each supported by 4 micro cell BSs and powered by a PV panel system, energy batteries and the power grid. Each BS of the considered RAN is equipped with a caching server, where the most popular contents are stored. In case the renewable energy generation is not sufficient for the network supply, the micro BSs of the network are deactivated. The contributions of this work are:

  • Using a simulation-based approach, we quantify the gain as well as the cost, which derive from the usage of caching servers, placed at each BS. The gain is measured in experienced delay drop, and the cost is expressed in growth of the energy consumption of the network. These quantities are also evaluated when an energy reducing strategy is used, which deactivates each micro cell BS in case of local renewable energy shortage.

  • By simulations, we derive the impact of the different traffic characteristics and of the different capacities of each cache.

  • Different spread of the cache capacity among the BSs of the network is investigated and we observe that caching on the macro BSs is always needed to significantly reduce delays, while caching also on the micro cells relieves the effort on the macro cell. This allows the micro cells to often respond without any involvement of the macro cell, providing a slight delay reduction.

  • Different association policies are proposed, which aim at minimising the RAN energy consumption and/or the experienced delay, in order to maximise the benefits provided by the MEC technology usage, ensuring also the achievement of energy efficiency.

The paper is organised as follows. Related works are revised in Section 2. In Section 3, the scenario and the methodology of our work are presented, while the used Key Performance Indicators (KPIs) are described in Section 4. Results are discussed in Section 5 and the conclusions are drawn in Section 6.

Section snippets

Energy efficiency in wireless access networks

The employment of the RES for the power supply of the BSs of RANs has been receiving much attention because it reduces the carbon emissions and of the electricity bill  [24], [25], [26]. Various papers address the critical issue of properly dimensioning RE generation systems to power mobile networks [24], [25], [26]. The sizing process brings to a trade off among self-sustainability, cost and feasibility constraints due to the installation of a RE generation system. The RE system sizing problem

Methodology

In this work, the heterogeneous RAN portion considered in [22], covering an area of 0.3 km2 of the city centre of Ghent, in Belgium, is considered (orange rectangle in Fig. 1).

The RAN that covers this area is composed of 8 macro cell BSs, marked by the blue points in Fig. 1. In order to provide additional capacity during high traffic demand periods, each macro cell BS is supported by 4 micro cell BSs. These are indicated with the brown points in the figure and their radio coverage overlaps with

Energy consumption

The energy consumption of the network during the simulation, in Wh, is given by E=t=1TEtot(t)where Etot(t) is the energy consumption of the network at time t and it is computed as reported in (2) and T is the duration of the simulation.

Green energy

The green energy, in Wh, accounts for the amount of used energy which has been produced by the PV-panel system, which is locally installed.

Brown energy

The brown energy, in Wh, indicates the energy bought from the power grid, for the network supply. As already mentioned, in

Performance evaluation

In this section, we discuss the results obtained, when the considered RAN is simulated, assuming that it operates for 1 week. The values of latency, as well as the value of the parameter ωMEC of (3) are reported in Table 3. They are taken from [43] and [49], respectively.

Conclusion

In this paper, mechanisms to move towards two of the objectives of 5G networks, the delay reduction and the network energy efficiency, are revised and proposed. A portion of a RAN, composed by 8 macro cell BSs, each supported by 4 micro cell BSs is considered, where the MEC technology is employed, to push the most popular contents closer to users so as to reduce latency. The considered RAN is powered by a PV panel system and an energy battery and is connected to the power grid. Different users

CRediT authorship contribution statement

Greta Vallero: Conceptualization, Methodology, Software, Writing - original draft. Margot Deruyck: Data curation, Conceptualization, Supervision. Michela Meo: Data curation, Conceptualization, Supervision. Wout Joseph: Data curation, Conceptualization, Supervision.

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.

Acknowledgement

M. Deruyck is a Post-Doctoral Fellow of the FWO-V (Research Foundation - Flanders, Belgium , no. 12Z5621N).

Greta Vallero was born in Torino (Italy) on September 21st, 1993. She got her bachelor degree in Computer Engineering, in Politecnico di Torino, in 2015; she then obtained the master degree with summa cum laude, in ICT for Smart Societies (Telecommunication Engineering), in October 2017, in Politecnico di Torino. From March 2017 to August 2017, she was hosted by Ghent University, in Belgium, to work on her master thesis, supervised by Prof. Michela Meo, in collaboration with Prof. Wout Joseph

References (59)

  • LiuD. et al.

    Caching at the wireless edge: design aspects, challenges, and future directions

    IEEE Commun. Mag.

    (2016)
  • DengR.

    DASH based video caching in MEC-assisted heterogeneous networks

    Multimedia Tools Appl.

    (2020)
  • SunY. et al.

    Delay analysis and optimization in cache-enabled multi-cell cooperative networks

  • UgwuanyiE.E. et al.

    Co-operative and hybrid replacement caching for multi-access mobile edge computing

  • OsseiranA. et al.

    Scenarios for 5G mobile and wireless communications: the vision of the METIS project

    IEEE Commun. Mag.

    (2014)
  • ForecastC.V.

    Cisco Visual Networking Index: Forecast and Trends, 2017–2022, White paper

    (2019)
  • GatiA. et al.

    Key technologies to accelerate the ICT Green evolution–An operator’s point of view

    (2019)
  • PompiliD. et al.

    Elastic resource utilization framework for high capacity and energy efficiency in cloud RAN

    IEEE Commun. Mag.

    (2016)
  • SabellaD. et al.

    Energy efficiency benefits of RAN-as-a-service concept for a cloud-based 5G mobile network infrastructure

    IEEE Access

    (2014)
  • BertoldiP.

    EU Code of Conduct on Energy Consumption of Broadband Equipment

    (2017)
  • BudziszŁ. et al.

    Dynamic resource provisioning for energy efficiency in wireless access networks: A survey and an outlook

    IEEE Commun. Surv. Tutor.

    (2014)
  • DalmassoM. et al.

    Radio resource management for improving energy self-sufficiency of green mobile networks

    ACM SIGMETRICS Perform. Eval. Rev.

    (2016)
  • DeruyckM. et al.

    Reducing the power consumption in LTE-advanced wireless access networks by a capacity based deployment tool

    Radio Sci.

    (2014)
  • ShankarT.

    A survey on techniques related to base station sleeping in green communication and CoMP analysis

  • BuzziS. et al.

    A survey of energy-efficient techniques for 5G networks and challenges ahead

    IEEE J. Sel. Areas Commun.

    (2016)
  • DeruyckM. et al.

    Accounting for the varying supply of solar energy when designing wireless access networks

    IEEE Trans. Green Commun. Netw.

    (2017)
  • ValleroG. et al.

    Caching at the edge in high energy-efficient wireless access networks

  • PamukluT. et al.

    Reducing the total cost of ownership in radio access networks by using renewable energy resources

    Wirel. Netw.

    (2020)
  • da SilvaA.P.C. et al.

    Small solar panels can drastically reduce the carbon footprint of radio access networks

  • Cited by (18)

    • Energy-efficient proactive caching with multipath routing

      2022, Computer Networks
      Citation Excerpt :

      A hybrid caching policy for reducing latency and energy consumption is designed in [15] which consists of device local caching, device-to-device (D2D) caching, base station (BS) caching. Except for saving the traditional energy in the above works, energy harvesting technology together with cooperative caching has also received scholarly attention in recent years, such as BS caching in [16,17] and D2D caching in [18]. In particular, the authors in [16] formulate a BS caching problem as a mixed integer linear programming model (MILP) and then decompose it as two sub-problems, and the work [18] investigates the caching performance with stochastic geometry.

    View all citing articles on Scopus

    Greta Vallero was born in Torino (Italy) on September 21st, 1993. She got her bachelor degree in Computer Engineering, in Politecnico di Torino, in 2015; she then obtained the master degree with summa cum laude, in ICT for Smart Societies (Telecommunication Engineering), in October 2017, in Politecnico di Torino. From March 2017 to August 2017, she was hosted by Ghent University, in Belgium, to work on her master thesis, supervised by Prof. Michela Meo, in collaboration with Prof. Wout Joseph (Ghent University) and Dr. Margot Deruyck (Ghent University). In 2018, she starts officially her Ph.D., under the supervision of Professor Michela Meo, at Politecnico di Torino. Her main research interests are Multi-Access Edge Computing, as well as the energy efficiency in Radio Access Networks, using the support of Machine Learning algorithms, through radio resource management and network renewable energy supply.

    Margot Deruyck received the M.Sc. degree in Computer Science Engineering and the Ph.D. degree from Ghent University, Ghent, Belgium, in 2009 and 2015, respectively. From September 2009 to January 2015, she was a Research Assistant with Ghent University — IMEC – WAVES (Wireless, Acoustics, Environment & Expert Systems – Department of Information Technology). Her scientific work is focused on green wireless access networks with minimal power consumption and minimal exposure from human beings. This work led to the Ph.D. degree. Since January 2015, she has been a Postdoctoral Researcher at the same institution where she continues her work in green wireless access network.

    Michela Meo received the Laurea degree in electronic engineering and the Ph.D. degree in electronic and telecommunications engineering from the Politecnico di Torino, Italy, in 1993 and 1997, respectively, where she has been a Professor since 2006. She has co-authored about 200 papers and edited a book with Wiley and special issues of international journals, including ACM Monet, Performance Evaluation, and Computer Networks. Her research interests include performance evaluation and modelling, green networking, and traffic classification and characterisation. She is an Associate Editor of the IEEE Communications Surveys & Tutorials and an Area Editor of the IEEE Transactions on Green Communications and Networking. He was an Associate Editor of the IEEE Transactions of Networking. She chairs the Steering Committee of IEEE OnlineGreenComm and the International Advisory Council of ITC. She was the Program Co-Chair of several conferences, including ACM MSWiM, IEEE Online GreenComm, IEEE ISCC, IEEE Infocom Miniconference, and ITC.

    Wout Joseph received the M.Sc. degree in electrical engineering from Ghent University (Belgium) in July 2000. From September 2000 to March 2005, he was a research assistant at the Department of Information Technology (INTEC) of the same university. During this period, his scientific work was focused on electromagnetic exposure assessment. His research work dealt with measuring and modelling of electromagnetic fields around base stations for mobile communications related to the health effects of the exposure to electromagnetic radiation. This work led to a Ph.D. degree in March 2005. Since April 2005, he has been a postdoctoral researcher for IBBT-Ugent/INTEC (Interdisciplinary institute for Broadband Technology). Since October 2007, he has been a Post-Doctoral Fellow of the FWO-V (Research Foundation – Flanders). Since October 2009 he has been a professor in the domain of “Experimental Characterisation of wireless communication systems.” His professional interests are electromagnetic field exposure assessment, propagation for wireless communication systems, antennas and calibration. Furthermore, he specialises in wireless performance analysis and Quality of Experience.

    View full text