Elsevier

Computer Communications

Volume 144, 15 August 2019, Pages 66-75
Computer Communications

GreenLoading: Using the citizens band radio for energy-efficient offloading of shared interests

https://doi.org/10.1016/j.comcom.2019.05.002Get rights and content

Abstract

Cellular networks are susceptible to being severely capacity-constrained during peak traffic hours or at special events such as sports and concerts. Many other applications are emerging for LTE and 5G networks that inject machine-to-machine (M2M) communications for Internet of Things (IoT) devices that sense the environment and react to diurnal patterns observed. Both for users and devices, the high congestion levels frequently lead to numerous retransmissions and severe battery depletion. However, there are frequently social cues that could be gleaned from interactions from websites and social networks of shared interest to a particular region at a particular time. Cellular network operators have sought to address these high levels of fluctuations and traffic burstiness via the use of offloading to unlicensed bands, which may be instructed by these social cues. In this paper, we leverage shared interest information in a given area to conserve power via the use of offloading to the emerging Citizens Broadband Radio Service (CBRS). Our GreenLoading framework enables hierarchical data delivery to significantly reduce power consumption and user fairness variation and includes a Broker Priority Assignment (BPA) algorithm to select data brokers for users. With the use of in-field measurements and web-based Google data across four diverse U.S. cities, we show an order of magnitude power savings via GreenLoading over a 24-hour period, on average, and power savings up to 97% at peak traffic times. Finally, we consider the role that a relaxation of wait times can play in the power efficiency of a GreenLoading network.

Introduction

While the focus of cellular congestion is frequently placed on the frustration that users experience when they lack the ability to call, text, or receive web-based information, there is a byproduct of excessive re-transmissions in a congested state: severe battery expenditure. The problem could be even more extreme for battery-limited devices associated with the emergence and potential explosive growth expected for Internet of Things (IoT), Vehicle-to-Vehicle (V2V) and Machine-to-Machine (M2M) based communication in LTE and 5G networks [1], [2]. Such congestion is hard to alleviate since the data-intensive services and scale of mobile devices has grown the global data traffic 18-fold over the last 5 years and expected to reach 49 exabytes per month by 2021 [3].

One promising solution is to deploy traffic offloading, where cellular traffic is moved to less-congested, often unlicensed spectrum [4]. The primary objective of traffic offloading is to support more capacity-hungry services while simultaneously preserving satisfactory Quality-of-Service (QoS). Small cells, WiFi networks, and opportunistic communications have recently emerged as the main cellular offloading technologies [5]. While smaller cells have helped, simply offloading traffic from macro cells to small cells may not increase the transmission rate, improve the user experience, or reduce power expenditure. Some work has investigated this relationship between energy savings and traffic offloading to small cells [4]. Other work has focused on the switching times and performance improvements for cellular offloading to WiFi [6]. The use of WiFi or even white space bands has also been advocated for energy-efficient cellular offloading [7].

However, there remains a yet untapped spectrum resource for cellular offloading: the Citizens Broadband Radio Service (CBRS) for shared wireless access using the carrier frequencies of 3550–3700 MHz (3.5 GHz Band). Unlike WiFi and white spaces, CBRS is expected to be fully operational in the context of 5G. CBRS access will be managed by a dynamic spectrum access system, conceptually similar to the databases used to manage TV white space devices but at faster time scales. The three tiers of access are: Incumbent Access (existing users of 3650–3700 MHz), Priority Access (network operators may purchase up to seven 10 MHz Priority Access Licenses (PALs) in a census track from 3550–3650 MHz), and General Authorized Access (unallocated bandwidth from the first two tiers). Hence, up to 150 MHz may be available in a given area for opportunistic use [8].

The time and place of network congestion can often stem from mutually-shared environmental factors, causing a surge in data (e.g., roadway conditions, audio/video from live events, or emergency situations). These shared interests and the data redundancy thereof have largely been overlooked when optimizing offloading strategies in terms of capacity and power consumption [9]. In this paper, we leverage shared interest information in a given area to conserve power via the use of offloading to the emerging CBRS spectrum. To do so, we use a data broker where mutual information can be broadcast to the interested parties with the following hierarchical structure: consider one extreme where all devices connect directly to the macro cell and no data broker is needed. In this situation, the channel will be divided for all users and the interference generated could cause poor data rates over the network. Now, consider the other extreme where all devices work through a data broker to receive their information. If the amount of overlap in the shared interests is extremely high, there are tremendous savings of the spectrum. However, if the amount of overlap in the shared interests is extremely low, there could be severe delays and greater power consumption in working through a data broker to deliver unique content to individual users.

Hence, the crux of our work is establishing when it would be beneficial to use a data broker based on the number of users in an area, their mutual overlap of shared interests, the QoS response time required for a given application, and the availability of spectrum for offloading. These five factors are considered in our Broker Priority Assignment (BPA) algorithm. With the use of crowdsourced Google Maps measurements, we build a data transformation model that allows analysis across four diverse U.S. cities. We show that GreenLoading with shared interest data in a given area and the use of CBRS channels can reduce the power consumed by an order of magnitude over a 24-h period, on average. At peak traffic times, we find that our framework can reduce power expended by 97%.

The main contributions of our work are as follows:

  • We leverage Google Maps data to create a relationship between the travel time and number vehicles over a 24-h period in four major U.S. cities so that the commuting pattern of users on the road can be characterized.

  • We consider the data demand characteristics of these users in these four cities and use it to motivate and analyze a GreenLoading data sharing framework, which uses our BPA algorithm to quantify the power savings of our scheme.

  • We perform measurement-driven numerical evaluations of various QoS scenarios and user distributions to show that CBRS offloading can reduce the power consumption by up to 97%. We further show that the power savings can be reduced by 95% from a cellular-only configuration with a CBRS channel and the GreenLoading framework. In dense urban areas, we show the average power consumption over a 24-h period can be reduced by over 10 times versus a cellular-only network.

  • We perform measurement-based evaluation of user fairness to show that CBRS offloading can improve the max–min ratio by 81% and Jain fairness index by 64%.

  • We show the role that the relaxation in user wait times plays on the energy savings that one may experience using the GreenLoading framework with a wide range of realistic scenarios in our analysis.

The remainder of the paper proceeds as follows. In Section 2, we motivate the use of shared interest demand profiles to construct the GreenLoading framework, introduce our BPA algorithm, and analytically model the key aspects of their performance. We then consider four major U.S. cities and quantify various QoS scenarios and the energy savings that our GreenLoading framework offers in Section 3. We discuss related work in Section 4. Finally, we draw conclusions in Section 5.

Section snippets

GreenLoading framework

Cellular offloading refers to the mixed use of cellular data traffic with various available unlicensed bands such as Bluetooth, WiFi, white space, and CBRS networks. Cellular network operators are motivated to leverage these unlicensed bands for greater capacity and higher QoS. If offloading additionally provided power savings, there would be a reduction in ongoing operating costs of the network and potentially a solution in select rural infrastructures that have begun to depend on solar power 

Evaluation of GreenLoading framework

In this section, we introduce the experimental setup and evaluation. In particular, we analyze the Broker Priority Assignment (BPA) Algorithm in our GreenLoading framework and compare the power consumption of BPA across various levels of cellular offloading in a variety scenarios in four major U.S. cities.

Related work

Various forms of multiple radio and multiple channel opportunistic networking exist where devices can communicate with various radios, frequency bands, and/or channels [31]. While this simultaneous use of multiple radios for communication offers the opportunity for data sharing in small groups, previous works have focused on specific applications such as IoT or self-driving cars [32]. An opportunistic communication model for cellular traffic was proposed in [5]. More to the point, energy

Conclusion

In this paper, we created the GreenLoading framework to efficiently offload cellular network traffic to the emerging CBRS band via the use of shared interest information and data brokers. To achieve this goal, we developed a Broker Priority Assignment (BPA) algorithm to select the shared-interest user groups for the data brokers to broadcast traffic. With the use of in-field measurements and web-based Google Maps data across four diverse U.S. cities in both dense and sparse areas, we showed

Acknowledgment

This work was in part supported by the following National Science Foundation (NSF), USA grant: CNS-1526269.

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.

References (39)

  • HanB. et al.

    Cellular traffic offloading through opportunistic communications: a case study

  • SinghS. et al.

    Offloading in heterogeneous networks: Modeling, analysis, and design insights

    IEEE Trans. Wirel. Commun.

    (2013)
  • CuiP. et al.

    Whitecell: Energy-efficient use of unlicensed frequency bands for cellular offloading

  • WilesW.W.

    Federal reserve system

    System

    (2015)
  • PietiläinenA.K. et al.

    Mobiclique: middleware for mobile social networking

  • ShiW. et al.

    Edge computing: Vision and challenges

    IEEE Internet Things J.

    (2016)
  • NahF.F.-H.

    A study on tolerable waiting time: how long are web users willing to wait?

    Behav. Inf. Technol.

    (2004)
  • WangY. et al.

    Informing the use of vibrotactile feedback for information communication: an analysis of user performance across different vibrotactile designs

  • JainR.K. et al.

    A Quantitative Measure of Fairness and Discrimination

    (1984)
  • View full text