1 Edge-Computing-Enabled Smart Cities: A Comprehensive Survey Latif U. Khan, Ibrar Yaqoob, Senior Member, IEEE, Nguyen H. Tran, Senior Member, IEEE, S. M. Ah- san Kazmi, Tri Nguyen Dang, Choong Seon Hong, Senior Member, IEEE Abstract—Recent years have disclosed a remarkable prolifer- not limited and can refer to many different processes, resulting ation of compute-intensive applications in smart cities. Such in more reliable, secure, sustainable, and innovative cities with applications continuously generate enormous amounts of data unique entrepreneurial opportunities. The key contributors to which demand strict latency-aware computational processing capabilities. Although edge computing is an appealing technology smart cities are platform developers, research communities, to compensate for stringent latency related issues, its deployment governments, citizens, service providers, and application de- engenders new challenges. In this survey, we highlight the velopers [7]. role of edge computing in realizing the vision of smart cities. A remarkable increase in city populations is expected in First, we analyze the evolution of edge computing paradigms. the foreseeable future, which will impose substantial computa- Subsequently, we critically review the state-of-the-art literature focusing on edge computing applications in smart cities. Later, we tional complexity regarding advancements in the smart cities. categorize and classify the literature by devising a comprehensive Moreover, this rapid rise in populations will impose further and meticulous taxonomy. Furthermore, we identify and discuss challenges on the city infrastructure due to an exponential key requirements, and enumerate recently reported synergies of increase in data generated by different devices, such as smart- edge computing enabled smart cities. Finally, several indispens- phones, global positioning systems, smart sensors, and smart able open challenges along with their causes and guidelines are discussed, serving as future research directions. cameras. One way to cope with the challenges of massive scale computing and storage is the use of expensive dedicated com- Index Terms—Smart cities, Internet of Things, Mobile cloud puting hardware with sufficient storage capacity. To eliminate computing, Cloudlet, Fog computing, Mobile edge computing, Micro data centers. the need for dedicated expensive computing hardware, cloud computing was introduced. This is a powerful technology that is intended to enhance the Quality of Experience (QoE) by I. INTRODUCTION providing on-demand storage and processing capabilities in a The unprecedented plethora of miniaturized sensing tech- cost-effective and elastic manner [8]. The striking features of nologies has shown a remarkable breadth of smart cities vision. cloud computing, such as scalability, elasticity, multitenancy, Smart cities enable citizens to experience sustainable, secure, sufficient storage capacity, and resource pooling, have made its and reliable developments [1], [2]. Smart cities are defined for- adoption possible in different areas [9]–[11]. Recently, cloud mally in numerous ways [3]–[6] and thus, there is no absolute computing with virtually unlimited resources has emerged as definition of the smart cities. Generally, these definitions refer a promising paradigm to tackle the high computational com- to advancements related to information and communication plexity associated with smart cities [12]. However, its inherent technologies (ICT) infrastructure, that enable citizens to realize limitations of high latency, non-context-aware behavior, and high quality of life via smart services. Smart city services are no support for mobility pose serious limitations on its use in real-time smart environments. Apart from these downsides, arXiv:1909.08747v2 [cs.NI] 12 Oct 2020 cloud computing suffers from processing time inefficiency due This work was supported by Institute of Information & communications to the large overhead of smart city device data. On the other Technology Planning & Evaluation (IITP) grant funded by the Korea govern- ment(MSIT) (No.2019-0-01287, Evolvable Deep Learning Model Generation hand, edge computing extends the cloud computing resources Platform for Edge Computing) and partially supported by the MSIT(Ministry to the network edge and offers context-awareness, low latency, of Science and ICT), Korea, under the Grand Information Technology Re- mobility support, scalability, to name a few. Hence, to address search Center support program(IITP-2020-2015-0-00742) supervised by the IITP(Institute for Information & communications Technology Planning & the limitations of cloud computing for enabling real-time smart Evaluation). *Dr. C S Hong is the corresponding author. city environments, edge computing is a viable solution [13]– [16]. L. U. Khan, I. Yaqoob, T. N. Dang and C. S. Hong are with the Department of Computer Science & Engineering, Kyung Hee University, Yongin-si 17104, Figure 1 elucidates a high-level overview of Internet South Korea. of Things (IoT) based smart city environments enabled by N. H. Tran is with School of Computer Science, The University of Sydney, edge computing. These IoT based smart environments leverage Sydney, NSW 2006, Australia. S. M. A. Kazmi is with the Networks & Blockchain Lab, Secure System connected smart IoT devices and sensors to improved life and Network Engineering, Innopolis University, 420500 Tatarstan, Russia, and standards for citizens. Apart from that, IoT can be defined the Department of Computer Science and Engineering, Kyung Hee University, in different ways because it is associated with a variety of Yongin-si 17104, South Korea. technologies and concepts in literature [17]–[19]. Mostly, the definitions of IoT are inspired from research projects [20]. In 2 Smart Buildings Smart Airport Waste Recycling plant Smart People Smart Healthcare Smart Waste Smart Police bin Cloudlet` Augmented Reality Cloud glasses Blockchain enabled FOG Node Banking Mobile Edge Computing Server hospital Cellular BS hospital Road Smart Parking Side Unit Smart Smart Grid Agriculture 1 2 3 4 5 Fig. 1: An overview of edge computing enabled smart city. [17], the authors found the common and frequently occurring context-awareness, mobility support, and low latency. It mi- features among different IoT definitions, which include: grates computing resources, such as computing power, data, • Existence of a global network infrastructure that enables and applications, from the remote cloud to the network edge, a unique addressing scheme, seamless integration, and and thus enables numerous real-time smart city services. connectivity among heterogeneous IoT nodes. Recently, three different emerging technologies, cloudlet, fog • IoT nodes must be addressable, readable, locatable, rec- computing, and mobile edge computing, have been used in ognizable, and autonomous. the literature to bring the striking features of cloud computing • Existence of heterogeneous technologies that must be to the edge [24]. For example, edge computing reduces the given assiduous attention. network bandwidth usage by alleviating the transmission of • Association of services to objects to emphasize its object data from end users to the remote cloud [13]. dependent nature. • Intelligent interfaces between things and humans. A. Research Trend and Statistics We can consider the above-mentioned features (as emphasized Figure 2 illustrates edge computing and smart cities re- in [17]) in a platform to constitute IoT. However, it may not search trends showing the exponential increases in the numbers always be possible for IoT system to possess all of these of publications in both domains. Apart from the research features. IoT is broadly classified into Industrial Internet of publications trend, according to statistics, the percentages of Things (IIoT) and Consumer Internet of Things (CIoT) in people living in cities were 29% and 50% in the years 1950 [18]. The IoT assisted applications for smart cities include and 2008, respectively, which is expected to reach 65% in smart grids, smart lighting systems, smart transportation, smart 2040 [25]. Approximately 1.3 million people are moving health-care, smart homes, augmented reality, and smart agri- to cities every week. In 1975, there were three megacities, culture [1], [21]–[23]. Therefore, it is evident from literature but currently there are 21 mega cities having a population that IoT is an integral part of smart cities. The next step is more than 10 million. Moreover, the number of megacities in enabling the resource intensive and strict latency IoT based 2025 is expected to increase by one city in Africa, two in smart city applications. Edge computing provides a promising Latin America, and five in Asia. Furthermore, cities utilize way of enabling these applications by offering computation approximately 60%-80% of the world energy and generate and storage resources with low latency. 50% of greenhouse gas emissions. Edge computing is a cutting-edge computing paradigm The considerable rise in the number of smart cities in the characterized primarily by its geo-distributed operation, foreseeable future will likely to increase smart cities market 3 autonomous cars ranging from level 0 to level 5 [28]. The level 0 represents no automation and level 5 represents full automa- tion, such as operating without driver assistance. Moreover, autonomous cars can handle lane changing, warning signals, collision avoidance, and in-car infotainment, to name a few. To enable real-time analytics