Opportunities for Applications Using 5G Networks:Requirements, Challenges, and Outlook

Aaron Yi DingTU Del�

aaron.ding@tudel�.nl

Marijn JanssenTU Del�

m.f.w.h.a.janssen@tudel�.nl

ABSTRACT�e increasing demand for mobile network capacity drivenby Internet of �ings (IoT) applications results in the needfor understanding be�er the potential and limitations of5G networks. Vertical application areas like smart mobility,energy networks, industrial IoT applications, and AR/VRenhanced services all pose di�erent requirements on theuse of 5G networks. Some applications need low latency,whereas others need high bandwidth or security support.�e goal of this paper is to identify the requirements and tounderstand the limitations for 5G driven applications. Wereview application areas and list the typical challenges andrequirements posed on 5G networks. A main challenge willbe to develop a network architecture being able to dynami-cally adapt to �uctuating tra�c pa�erns and accommodatingvarious technologies such as edge computing, blockchainbased distributed ledger, so�ware de�ned networking, andvirtualization. To inspire future research, we reveal openproblems and highlight the need for piloting with 5G appli-cations, with tangible steps, to understand the con�gurationof 5G networks and the use of applications across multiplevertical industries.

KEYWORDS5G Systems; Pilot; IoT; Smart City; Edge Computing

1 INTRODUCTION�e success of mobile communication stems from its per-vasive coverage and substantial ecosystem that boost rapidpace of innovation in terms of new applications and venturecreations. To withstand its long-term prosperity, the upcom-ing ��h generation (5G) of mobile networks are expected togenerate new opportunities in the era of Internet of �ings(IoT), autonomous driving, augmented and virtual reality(AR/VR) services. �is vision is supported by the on-goingdevelopment of 5G cellular architecture and its air interfaceenhancement [1, 2] to cater for massive deployment of con-nected devices, which are projected to reach more than 75billion by 2025 [3].

Comparing with existing 4G, 5G networks encompassnew wireless interfaces to support higher frequencies andspectrum e�ciency. �ere is signi�cant improvement in

terms of signaling, management and accounting proceduresat the 5G core networks in order to accommodate the needsfrom diverse range of new applications that are outside tra-ditional mobile broadband category [4]. By its design, 5Gdeployment will provide extensive connectivity through itsheterogeneous wireless access, ranging from macrocell (longrange) to femtocell (short range). As shown in Figure 1, thecoverage will span across metropolitan area, municipal areaand down to campuses and buildings. �is pervasive connec-tivity is the key to seamless mobility and service availabilitythat has been centered in the cellular system since its debut.

Given the new demands from IoT, autonomous driving,AR/VR and smart city services, one important pursuit for 5Gis to match its capacity to the scale and growth of various 5Gdriven applications in an economical and sustainable manner.�is mission covers network architecture, communicationtechniques, ecosystem design and actual deployment. Re-cent e�orts have sought the utilization of network functionvirtualization (NFV), so�ware-de�ned networking (SDN),edge computing and o�oading, as well as distributed dataanalytics (e.g., Apache Spark [5]). �ose technical innova-tions have shown promising results [6–16]. Meanwhile, asnew application domains are inspired by 5G on almost dailybasis, we still lack a comprehensive understanding on re-quirements originating from various domains which includeboth technical and governance perspectives, especially whatopportunities and challenges 5G will endure in the currenttransitional phase.

Motivated by the latest advance in 5G and trend of urban-ization [17], this paper tackles the challenges of 5G from theapplication perspective. In particular, we focus on verticalapplication domains, which are built for target enterpriseand entities with speci�c requirements. As a solid step todemystify the requirements in the context of IoT and smartcities, we aim to answer the major question: ”How do weconsolidate 5G driven applications across multiple ver-tical industries to unveil the full potential of 5G?” Be-sides identifying challenges, our work also highlights theopportunities for 5G applications in various vertical domains,which can shed light on future development for researchers,engineers and policy makers from academia, industry andgovernment.

Our key contributions are hence twofold:

Figure 1: Envision of 5G Network Connectivity

• First, we classify the application domains inspiredby 5G, and quantify their key requirements. Ourrequirement analysis covers major aspects includingcommunication range, bandwidth capacity, latency,reliability, energy, security and privacy.

• Second, we pinpoint open challenges and opportuni-ties based on reviewing the state-of-the-art researchand project initiatives. Our discussions cover bothtechnical aspect and also regulation and governance.We further stress the need to pilot 5G experimentaltestbed through tight cooperation across universities,network operators, equipment vendors and govern-mental institutes.

We note that although this work provides an extensivesampling of existing and emerging 5G applications, our studydoes not a�empt to cover every nuance. �e rest of this paperis organized as follows. Section 2 provides an overview ofapplication domains in 5G. Section 3 illustrates the applica-tion requirements. Section 4 highlights open challenges andpotential opportunities in 5G. We discuss related work andproject initiatives and conclude with our outlook in Section5.

2 5G ENABLED APPLICATIONS�e advances in mobile networking have created a myriadof diverse applications to improve the life quality of endusers, including smart mobility, digital commerce, socialnetworking and health care. From a broader perspective,mobile applications are part of the Internet services, whichhave witness a rapid evolution over the past decades.

As illustrated in Figure 2, the Internet services have evolvedfrom conventional point-to-point data exchange, world wideweb (WWW), mobile and social applications, to the recentIoT services and forthcoming tactile Internet [18–20]. Inspeci�c, the tactile Internet applications will facilitate theintegration between digital sphere and our physical environ-ments, covering advanced use cases of machine-to-machine(M2M) communication. �ose new applications are charac-terized by the need for a network having ultra low latency,high availability, reliability and security. Many of these ap-plications are also context-aware where the context is sensedfor triggering actions, e.g., smartphones are nowadays awareif the owner is driving and can avoid interrupting the driver[21].

2

Figure 2: Advancement of Internet Services

Among several application domains, the IoT-empoweredsmart city has become a focal concern of 5G. In this context,smart city integrates traditional and modern Information andCommunication Technology (ICT) for a uni�ed and simpleaccess to services for the city administration and the resi-dents. �e aim is an enhanced use of resources, improvingquality of services for citizens while reducing operationalcosts of public administration [22] and reducing the admin-istrative burden for citizens and businesses. For example,smart transportation should reduce congestion and pollutionand at the same time result in higher utilization of transport.

On the one hand, IoT has quickly advanced from an ex-perimental technology to the driving force of 5G systems.To fully exploit the opportunities behind IoT, 5G has placedIoT at a vital position in its ecosystem. On the other hand,realizing the IoT vision of smart city depends on a carefulintegration with 5G telecommunication technologies to pro-vide scalable and robust connectivity. Comprehensive andscalable supports from 5G are hence required to overcomethe economic and technical constraints of state-of-the-artconceptualizations and implementations, while maintainingboth practical and commercial appeals.

For 5G driven applications, we highlight �ve domainsthat can bene�t from a tight integration with 5G and nextgeneration cyber-physical infrastructure.

• SmartMobility: Mobility applications in 5G rangesfrom traditional road/route planning to the emerg-ing autonomous driving services (connected vehi-cles) and extended sharing economics of smart trans-portation. �e bene�ts of smart mobility includetra�c balancing, e�cient routing, accident preven-tion, energy saving, cost and emission reduction [23].From this group of applications, there is a strong de-mand for 5G to support pervasive connectivity, low

latency, high speed and link reliability, security andlow power consumption.

• Smart Energy: �is category of applications in-cludes power plant monitoring and management,smart grid networking, power failure detection andresponse, new consumption saving services for homesand o�ce buildings, energy marketplace and smartcharging stations for electronic vehicles. Smart en-ergy is expected to enhance e�ciency and reliabilityof power systems with renewable energy and achieveintelligent distribution. �e major demands for 5Gare on link reliability, security and privacy [24, 25].

• Smart Health: Health applications are becomingpopular among mobile users owing to the grow-ing awareness of �tness and well-being. Togetherwith the advance of smart wearables, applications ofthis category have covered mobile based conditionmonitoring and diagnosis, environmental quality in-spection. With more data collected from sensorsdeployed on wearable devices, smart health will pos-itively in�uence the medical and healthcare systems[2]. Another emerging application in this domainis the AR/VR enabled surgery, which will demandlow latency and high bandwidth, on top of the gen-eral requirements of low power, security and dataprivacy from 5G.

• Industrial Applications: Applications such as In-dustry IoT 4.0 [26] represent the next generationof cyber-physical services in terms of manufactur-ing, machine-to-machine (M2M) communication, 3Dprinting and AI supported construction. �e impactof those industrial applications will extend beyondfactories and plants, directly bene�ting the entiresociety. �e major demands for 5G include critically

3

Figure 3: Application groups empowered by 5G

high reliability, ultra low latency, support of massivedeployment, security and privacy.

• Consumer Applications: �e vast amount of con-sumer applications (Apps) re�ect the potential of 5Gmobile business and technology innovations. As weare familiar with typical mobile applications runningon smartphones and tablets, the emerging applica-tions include ultra HD (4K/8K) mobile streaming,blockchain based �nancial technology (FinTech), per-vasive gaming (like Pokemon GO 1), mobile AR/VRmixed reality services supported by unmanned aerialvehicles, and holographic technology such as HoloLens2. All those advanced services are demanding 5G tosupport extensive connectivity, high bandwidth, lowlatency, low energy footprint, link reliability andsecurity.

3 REQUIREMENT ANALYSISTo carry out a �ne-grained analysis that re�ects technicalinterdependency, we break down the aforementioned appli-cation domains (detailed Section 2) into four distinct types.As shown in Figure 3, 5G driven applications are divided intofour categories: 1) domestic type with short communication1h�ps://www.pokemongo.com/en-us/2h�ps://www.microso�.com/en-us/hololens

range, 2) remote type with long range, 3) latency critical,and 4) massive scale. Our goal is to quantitatively manifesttheir requirements for facilitating future development anddeployment of 5G systems. We also note that although thisgrouping includes a wide range of existing and emergingapplications, our discussions do not a�empt to cover everynuance.

3.1 General RequirementsFor each type of applications, we highlight the general re-quirements in Table 1, covering communication range, band-width capacity, latency, link reliability, energy consumption,security and privacy.

3.1.1 Domestic - short range. As shown in Figure 3, thisgroup of applications include consumer applications in thecontext of smart homes and o�ce buildings [22]. Owing totheir communication pa�ern, 5G needs to support low powernetworking, which is crucial for wearable devices. Giventhe exposed security issues at smart homes [32], there is astrong demand to regulate unwanted tra�c on the wirelessinterfaces.

3.1.2 Remote - long range. Applications in smart farm-ing and urban monitoring demand 5G support especially interms of communication coverage. Since devices deployed

4

Table 1: Vertical Application Requirements in 5G

Applications CommunicationRange

BandwidthCapacity

Latency LinkReliability

Energy SecurityPrivacy

Smart buildings short range 10 - 1000 Mbps Median Median Low HighSmart devices short range 10 - 1000 Mbps Median Median Low HighSmart farming long range 1 - 100 Mbps Tolerant Median Low MedianUrban monitoring long range 1 - 100 Mbps Tolerant Median Low MedianAutonomous driving long range 10 - 5000 Mbps Critical High High CriticalAR/VR services short range 100 - 5000 Mbps Critical Median High MedianSmart energy median range 10 - 1000 Mbps Median High Median HighSmart mobility long range 10 - 1000 Mbps Median High Median Median

for farming and urban monitoring need to operate over longtime period, energy saving is another key requirement.

3.1.3 Latency critical. Industrial applications are typicallytied to safety in manufacturing and hence demanding highlevel of security. For consumer domain such as autonomousdriving and AR/VR services, low latency (critical level) andhigh bandwidth must be supported in 5G communication.Due to the safety concern, autonomous driving also demandshigh link reliability.

3.1.4 Massive scale. For scenarios of massive deploymentsuch as in smart grid and transportation systems, 5G needs toelastically scale, to cater for increased tra�c demand, numberof end devices, and applications, and with acceptable cost. Inparticular to smart energy, high link reliability and securityare also required.

3.2 Requirements from Emerging ServicesFor emerging applications in both smart city and vertical in-dustries, 5G architecture needs to consider the requirementsfrom several new angles. �e �rst one rises from the swi� oftra�c pa�ern from downlink driven to uplink driven. �isis mainly due to the introduction of high volume of datagenerated from smart vehicles, drones, and industrial IoTdeployment. �e tra�c pa�erns might change and demandfrom di�erent vertical industries and can shi� over time. Be-ing able to adapt to the �uctuation will be a key requirementfor the 5G networks.

Secondly, due to the enforcement of General Data Protec-tion Regulation (GDPR) 3, data privacy is becoming an avidissue. Data referring to an identi�able or identi�ed personfalls under this regulation. Especially with more embeddeddevices and autonomously �ying drones/robots to collectdata for surveillance purposes, 5G needs to guarantee secu-rity in communication and ensure privacy-by-design. �ela�er refers to ensuring data protection by having a properarchitecture.3h�ps://www.eugdpr.org/

Besides technical requirements, 5G must take into accountthe requirements from governmental and economical angles.In this context, connectivity of 5G in the future will be re-garded as one of the mandatory common-pool resources(CPR) similar to water and electricity. �is has strong impli-cation on the regulation and management of 5G networksin terms of interoperability across operators, cost of mainte-nance, public-private sector ownership, wireless spectrumbidding and allocation (especially above 3 Ghz). Being a pub-lic resource, safety of large-scale operations will also becomea key requirement.

4 OPPORTUNITIES AND OPENCHALLENGES

4.1 Technology OpportunitiesWe identify four technology advancement that can bene�t5G, including blockchain inspired distributed ledger, light-weight virtualization, so�ware-de�ned Networking, andedge computing.

Blockchain based Distributed Ledger

Besides upholding user privacy on the Internet, GDPRis also accelerating the development of distributed ledgertechnologies (i.e., blockchain based protocol design). Speci�-cally, we are witnessing a strong demand nowadays to unifythe data management across end users, companies and gov-ernment. �is includes providing the appropriate means toreceive, track, and ful�ll user requests, and to update thedata as requested.

Given the high risk for enterprise in the face of steep �nes,Blockchain Technology (BCT) which can store a secure his-torical record of transactions in a tamper-proof format willplay a more visible role for many data driven applications in5G. In this context, BCT stores information at di�erent nodes.�e past information cannot be removed and informationcan only be added when the nodes possess it [27]. BCT was

5

Figure 4: Edge-enabled Platform for 5G

introduced for Bitcoin and is based on Distributed LedgerTechnology (DLT), in which each participant has access toa shared ledger which is stored in many nodes [28]. Alltransactions are stored in a ledger and all nodes have a copyof this. In turn, blocks are time stamped batches of validtransactions in which each block includes the hash of theprior block. Creating new blocks is known as mining [29].By linking the blocks a chain is formed which has resultedin the use of the name ‘blockchain’. Blockchain uses can befound in many sectors [30]. Especially for 5G empoweredIoT and mobile �nancial services, we will see a merge be-tween IoT security solutions [31–35] and novel blockchainand cryptocurrency designs to achieve be�er accountabilityand privacy in 5G.

Virtualization and So�ware-De�ned Networking

Network Function Virtualization (NFV) is a solid technol-ogy to organize network related computation in 5G. By itsdesign, NFV utilizes virtualization technologies to decou-ple physical network equipment from the functions runningon them [6]. �is way, various virtual network functionscan be implemented and deployed on one or more physical

servers. In particular, the new lightweight virtualization tech-nologies such as Docker and Unikernels [13] will facilitate5G to support new services of IoT domain which demandsmulti-tenancy, low cost, e�cient resource utilization, andlow power consumption.

Meanwhile, for managing 5G network tra�c, So�ware-De�ned Networking (SDN) is another powerful tool whichhas been successfully applied to data centers and commercialnetworks. In its essence, SDN decouples the data and controlplane so that all the control functions can be implemented ina centralized network controller. Its design transfers the con-trol functionality to so�ware based entities, which eliminatesthe use of vendor speci�c back-box hardware and promotesthe use commodity switches in data plane over proprietaryappliances [36]. On top of its security bene�ts [7], SDN canbe�er support multi-tenancy for large scale deployment of5G services, such as in smart city and industrial operations,e.g., by using frameworks such as So�O�oad [9, 10, 37].

Mobile Edge Computing

�e convergence of mobile Internet and wireless systemsin 5G can trigger an explosive growth in resource-hungryand computation-intensive applications, which cover a broad

6

paradigms of IoT. �ese IoT systems include real-time video/ audio surveillance, smart e-health, smart transportation,and Internet of Vehicles (IoV). Mobile edge computing, bycomplementing various cloud resources and bringing com-putation closer to smart devices/objects, has been envisionedas an enabling and highly promising technology to reap thepotential of IoT in 5G.

Recently, multi-access mobile edge computing (MA-MEC),which actively exploits a systematic and adaptive integrationof wireless access technologies in 5G, will further enhancethe access capacity between smart devices and mobile edgeplatforms. �e design of MA-MEC is in line with the evolu-tion towards ultra-dense deployment of small-cells (micro/ pico / femto cells) in future 5G. In speci�c, the denselydeployed 5G small cells can enhance the capacity and qual-ity of the connections. As an example, the emerging dual-connectivity in 5G networks can enable smart objects tocommunicate with conventional macro-cells and o�oad datatra�c to small cells simultaneously [38–42]. �is enhancesthe access capacity of mobile edge cloud at small cells. Inaddition, the existing computational o�oading techniques[43–45], including edge o�oading [14], will further com-plement the needs for speeding up computation and lowcommunication latency.

Edge-enabled 5G Service Framework

To exemplify how to combine new technologies into 5G,we propose an edge-enabled platform for 5G to consolidatedata management in large scale cyber-physical system de-ployment. As shown in Figure 4, through this platform, the5G edge layer is expected to bring a variety of bene�ts, suchas i) ultra-low latency between smart devices and edge cloudfor real-time, interactive, and mission-critical applications,e.g., industrial operations; ii) privacy and security in localcommunications; and iii) fast data processing at the pointof capture for IoT applications. For instance, the proposedplatform will provide robust and ultra-low latency connec-tions for smart vehicles to e�ciently access the edge layerdeployed on road-side units for real-time information pro-cessing. To build the edge layer, we can utilize the SDNframework [37], IoT management tool [46, 47], and Ka�aframework. 4

4.2 Open ChallengesA combination of promising technologies like NFV and edgecomputing is needed to meet the demands of new applica-tions. Nevertheless, the success of 5G still requires tacklingmany other challenges.

4h�ps://ka�a.apache.org/

Technical Challenges

For 5G network operation, security is a major concern. �erole of encryption, especially the operator driven pervasiveencryption [48], has raised lots of discussions across serviceproviders (e.g., Google, Amazon), ISPs e.g., KPN, T-Mobile,equipment vendors (e.g., Nokia, Ericsson) and standardiza-tion units such as IETF 5 and ETSI 6.

In the context of cellular systems, the conventional net-work management, security operations, and performanceoptimization have been conducted over a large majority ofdata tra�c �ows without encryption. While unencryptedtra�c could facilitate troubleshooting and management op-erations at all network layers, it has also made pervasivemonitoring by unseen parties possible. With support fromservice providers (e.g., Google) and increased awareness ofprivacy on the Internet [49], more and more tra�c are en-crypted in an end-to-end manner. �is trend has created achallenge for 5G since existing management, operational,and security practices have depended on the availability ofclear text to function. For 5G operators, it is important toinvestigate if critical operational practices can be met by lessinvasive means.

Besides conventional tra�c balancing between real-timeand typical web tra�c [50], 5G needs also to prioritize tra�ctypes with �ne granularity. In some vertical applicationsthe quick response is needed to avoid failure, whereas otherapplications response-time is less an issue. However, thistra�c di�erentiation is correlated with the net neutralitydebate whether the freedom and fairness of Internet will bea�ected.

To e�ciently exploit computation and storage resources atmobile edge nodes, a joint optimization of placement of com-putation/storage resource and cell-association with radio re-source allocation are required. Such joint optimization mustbe self-adaptive and with minimum manual e�orts. �e adap-tation needs to take into account time-varying environments,such as the varying wireless channel states when users moveacross the cells and computation/storage resource utiliza-tions.

Challenges from Regulation and Governance

As connectivity becomes a common-pool resource (CPR),there is a need for governance to manage fair usage, ensuresu�cient bandwidth and scalability, enforce interoperabil-ity and give priority to certain vertical applications. Forthis change, regulations might interfere with the role of 5Gproviders in the future. Latency critical application like con-nected vehicle might be given priority to avoid car collisions5h�ps://www.ietf.org/6h�ps://www.etsi.org/

7

Table 2: Comparison of 5G Pilot Initiatives

Pilots Experiment Scale Technology Operation Model FocusEnvisioned 5G Pilot City Scale 5G driven Public-Industry hybrid Consumer and

public servicesSingtel 5G Regional (Buona

Vista area,Singapore)

5G driven Company driven - Ericsson(vendor)

Network services

Toronto Waterfront Regional (Port areain Toronto)

Fixed network Company - Google (cloudservice provider), andpartially public sector

Infrastructureoriented

over other applications. Also the distributed nature mightdemand redundant coverage of areas to avoid problems incase of mall-function. Back-up and recovery plans might berequired by regulations.

Without regulation, it is challenging to ensure properfunctioning when some of the components are restricted orfail (e.g., due to market failure). For instance, the spectrumallocation is open to discussion as some spectra are alreadyoccupied by Department of Defense applications in certaincountries. Failures can be disastrous for critical vertical ap-plications areas. Edge computing architectures might beneeded to be able to operate independently of the networkto avoid failure of the larger system. Also security shouldbe enforced in such a way that the whole system cannot bebreached by a hack. Another aspect will be interoperabilitybetween di�erent providers and platforms. roaming betweenproviders should be possible to ensure proper functioning ofthe vertical applications which are likely to be operated bymultiple 5G providers.

�e General Data Protection Regulation (GDPR) repre-sents the largest change to European Union (EU) data protec-tion laws in decades. For 5G applications, one major criteriais on the private data collected from both end users and phys-ical infrastructure. Privacy-by-design should be guaranteedwhen using the 5G applications. In addition for be�er inte-gration of 5G, we also need to draw lessons from studies onstandardization [51], ambidexterity [52], and applying opendata to smart cities [53].

5 INITIATIVE AND OUTLOOKA main challenge for 5G will be to create a network archi-tecture that adapts to �uctuating tra�c pa�erns, consistsof promising technologies like edge computing, so�warede�ned networking, virtualization, and combines wired andwireless elements to deal with the requirements of variousvertical industries. �e vertical industries yield various re-quirements on 5G and the actual usage might �uctuate. �isrequires that the architecture is dynamic, able to prioritize

tra�c, and can ensure that edge computing power (as envi-sioned in Figure 4) is available for fast and e�cient process-ing and response.

Re�ecting on our main pursuit of this paper, ”How dowe consolidate 5G driven applications across multiplevertical industries to unveil the full potential of 5G?”we believe that this answer is non-trival and the answershall be sought from developing a comprehensive pilotingtestbed integrating the various technologies and in whichvertical industries are involved.

�is envisioned network testbed pilot needs to integratevarious technologies and be compliant to regulation and gov-ernance. To shed light on the 5G pilot, which will combinethe e�orts with Del� Green Village 7, we compare it againstthe Singtel 5G initiative 8 and Toronto Waterfront 9. We sum-marize our observations in Table 2 in terms of experimentscale, driven technology, operation model, and project focus.

Given the challenges we outlined, this pilot project mustbridge the gap between research community, industrial stake-holders, and governmental institutes. In particular from tech-nical perspectives, the envisioned 5G pilot should allow usto: 1) experiment novel radio access technologies and theirfeasibility for di�erent 5G applications; 2) incubate novelapplications by creating a trail infrastructure before enteringmass market; 3) expose unforeseen limitations of networkcon�gurations; and 4) illustrate how to minimize unneces-sary replacement costs through a feasible migration path,which can lead to signi�cant deployment scale.

We must note that although our work provides an exten-sive sampling of existing and emerging vertical applications,this study does not a�empt to cover every nuance. Furtherpiloting can reveal new challenges and be used to understandthe nature of the challenges. Besides that, the requirementanalysis and technologies discussed can be applied to a broadspectrum of scenarios on top of 5G context. In addition to

7h�ps://www.thegreenvillage.org/8h�ps://www.singtel.com/about-Us/news-releases/journey-to-5g-singtel-and-ericson-to-launch-singapores-�rst-5g-pilot-network9h�ps://sidewalktoronto.ca/

8

open challenges, our work highlights the opportunities for5G-enabled applications from both technical and governanceperspectives, which can shed light on future developmentfor researchers, engineers and policy makers from academia,industry and government.

ACKNOWLEDGMENTSWe thank Martin Kienzle (IBM), Inge van de Water (GemeenteDel�), and Dennis Meerburg (TU Del�) for their contributoryfeedback.

REFERENCES[1] Wang et al. Cellular Architecture and Key Technologies for 5G Wire-

less Communication Networks. IEEE Communications Magazine 52, 2(2014), 122-130.

[2] Akpakwu et al. A Survey on 5G Networks for the Internet of �ings:Communication Technologies and Challenges. IEEE Access 6 (2018),3619-3647.

[3] h�ps://www.statista.com/statistics/471264/iot-number-of-connected-devices-worldwide/ (visited on 07/31/2018)

[4] NGMN Alliance. NGMN 5G White Paper. 2015.[5] h�ps://spark.apache.org/ (visited on 07/31/2018)[6] Han et al. Network function virtualization: Challenges and opportu-

nities for innovations. IEEE Communications Magazine 53, 2 (2015),90-97.

[7] Aaron Yi Ding, Jon Crowcro�, Sasu Tarkoma, Hannu Flinck. So�wareDe�ned Networking for Security Enhancement in Wireless MobileNetworks. Computer Networks 66 (2014), 94-101.

[8] Costa-Requena et al. So�ware de�ned 5G mobile backhaul. In Proceed-ings of 1st International Conference on 5G for Ubiquitous Connectivity2014.

[9] Aaron Yi Ding, Yanhe Liu, Sasu Tarkoma, Hannu Flinck, Jon Crowcro�.Demo: An Open-source So�ware De�ned Platform for Collaborativeand Energy-aware WiFi O�oading. In Proceedings of ACM MobiCom2015.

[10] Aaron Yi Ding, Yanhe Liu, Sasu Tarkoma, Hannu Flinck, HenningSchulzrinne, Jon Crowcro�. Vision: Augmenting WiFi O�oadingwith An Open-source Collaborative Platform. In Proceedings of ACMMobiCom MCS 2015.

[11] Yi Ding. Collaborative Tra�c O�oading for Mobile Systems. PhDDissertation, University of Helsinki, 2015.

[12] Flores et al. Large-scale O�oading in the Internet of �ings. In Pro-ceedings of IEEE PerCom PerIoT 2017.

[13] Morabito et al. Consolidate IoT Edge Computing with LightweightVirtualization. IEEE Network 32, 1 (2018), 102-111.

[14] Cozzolino et al. FADES: Fine-Grained Edge O�oading with Unikernels.In Proceedings of ACM SIGCOMM HotConNet 2017.

[15] Cozzolino et al. Empowering Cyber-Physical Systems with FADEX.In Proceedings of ACM MobiSys 2018.

[16] Pu et al. Low Latency Geo-distributed Data Analytics. In ACM SIG-COMM 2015.

[17] h�ps://esa.un.org/unpd/wup/publications/�les/wup2014-highlights.Pdf (visited on 07/31/2018)

[18] Zanella et al. Internet of �ings for Smart Cities. IEEE Internet of �ingsJournal 1, 1 (2014), 22–32.

[19] Fe�weis et al. �e Tactile Internet: Applications and Challenges. IEEEVehicular Technology Magazine 9, 1 (2014), 64-70.

[20] Simsek et al. 5G-Enabled Tactile Internet. IEEE Journal on SelectedAreas in Communications 34, 3 (2016), 460-473.

[21] Shishkov et al. �ree Categories of Context-Aware Systems. In Pro-ceedings of International Symposium on BusinessModeling and So�wareDesign 2018.

[22] L. G. Anthopoulos. Understanding the smart city domain: A literaturereview. In Transforming City Governments for Successful Smart Cities.Springer. 2015. 9–21.

[23] Benevolo et al. Smart Mobility in Smart City. In Empowering Organi-zations 2016. 13–28.

[24] Fang et al. Smart Grid - �e New and Improved Power Grid: A Survey.IEEE Communications Surveys and Tutorials 14, 4 (2012), 944–980.

[25] Wang et al. Cyber security in the Smart Grid: Survey and challenges.Computer Networks 57, 5 (2013), 1344–1371.

[26] Wang et al. Towards smart factory for industry 4.0: a self-organizedmulti-agent system with big data based feedback and coordination.Computer Networks 101, 4 (2016), 158–168.

[27] Ølnes et al. Blockchain in government: Bene�ts and implications ofdistributed ledger technology for information sharing. GovernmentInformation �arterly 34, 3 (2017), 355–364.

[28] S. Nakamoto. Bitcoin: A Peer-to-Peer Electronic Cash System. h�ps://bitcoin.org/bitcoin.pdf (visited on 09/03/2018)

[29] Narayanan et al. Bitcoin and Cryptocurrency Technologies: A Com-prehensive Introduction. Princeton University Press.

[30] Tapsco� et al. �e Impact of Blockchain Goes Beyond Finan-cial Services. Harvard Business Review. h�ps://hbr.org/2016/05/the-impact-of-the-blockchain-goes-beyond-�nancial-services (visited on09/03/2018)

[31] Hafeez et al. Demo: Cloud-based Security As a Service for Smart IoTEnvironments. In Proceedings of ACM MobiCom S3 2015.

[32] Hafeez et al. Securebox: Toward Safer and Smarter IoT Networks. InProceedings of ACM CoNEXT CAN 2016.

[33] Hafeez et al. IOTURVA: Securing Device-to-Device (D2D) Communi-cation in IoT Networks. In Proceedings of ACM MobiCom CHANTS2017.

[34] Hafeez et al. Toward Secure Edge Networks: Taming Device-to-Device(D2D) Communication in IoT. CoRR abs/1712.05958. 2017

[35] Hafeez et al. Real-time IoT Device Activity Detection in Edge Networks.In Proceedings of 12th International Conference on Network and SystemSecurity 2018.

[36] Kreutz et al. So�ware-De�ned Networking: A Comprehensive Survey.IEEE Communications Surveys & Tutorials 103, 1 (2015), 14–76.

[37] h�ps://www.cs.helsinki.�/group/eit-sdn/so�o�oad.html (visited on07/31/2018)

[38] Ding et al. NAO: a framework to enable e�cient mobile o�oading. InProceedings of ACM Middleware PDT 2011.

[39] Korhonen et al. Controlling Tra�c O�oading Using Neighbor Dis-covery Protocol. IETF Internet Dra� 2012.

[40] Ding et al. Enabling energy-aware mobile data o�oading for smart-phones through vertical collaboration. In Proceedings ofACMCoNEXTstudent workshop 2012.

[41] Korhonen et al. Toward Network Controlled IP Tra�c O�oading.IEEE Communications Magazine 51, 3 (2013), 96–102.

[42] Ding et al. Enabling Energy-Aware Collaborative Mobile Data O�oad-ing for Smartphones. In Proceedings of IEEE SECON 2013.

[43] Cuervo et al. MAUI: making smartphones last longer with code o�oad.In Proceedings of ACM MobiSys 2010.

[44] Chun et al. CloneCloud: elastic execution between mobile device andcloud. In Proceedings of ACM EuroSys 2011.

[45] Kosta et al. �inkAir: Dynamic resource allocation and parallel execu-tion in the cloud for mobile code o�oading. In Proceedings of IEEEINFOCOM 2012.

[46] Haus et al. Managing IoT at the Edge: �e Case for BLE Beacons. InProceedings of ACM MobiCom SMARTOBJECTS 2017.

9

[47] Haus et al. Demo: iCon�g: What I See Is What I Con�gure. In Pro-ceedings of ACM CHANTS 2017.

[48] Moriarty et al. E�ects of Pervasive Encryption on Operators. RFC 84042018.

[49] Haus et al. Security and Privacy in Device-to-Device (D2D) Commu-nication: A Review. IEEE Communications Surveys & Tutorials 19, 2(2017), 1054–1079.

[50] Jarvinen et al. E�ect of Competing TCP Tra�c on Interactive Real-time Communication. In Proceedings of 14th International Conferenceon Passive and Active Measurement 2013.

[51] Ding et al. Bridging the Gap Between Internet Standardization and Net-working Research. ACM SIGCOMM Computer Communication Review44, 1 (2014), 56–62.

[52] Matheus et al. How to Become a Smart City?: Balancing Ambidexterityin Smart Cities. In Proceedings of 10th International Conference on�eory and Practice of Electronic Governance 2017.

[53] Janssen et al. Big and Open Linked Data BOLD to Create Smart Citiesand Citizens: Insights from Smart Energy and Mobility Cases. In Pro-ceedings of 14th IFIP International Conference on Electronic Government2015.

10