Industrial Internet of Things: Challenges, Opportunities...

1551-3203 (c) 2018 IEEE Personal use is permitted but republicationredistribution requires IEEE permission See httpwwwieeeorgpublications_standardspublicationsrightsindexhtml for more information

This article has been accepted for publication in a future issue of this journal but has not been fully edited Content may change prior to final publication Citation information DOI 101109TII20182852491 IEEETransactions on Industrial Informatics

IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS VOL X NO X APRIL 2018 1

Industrial Internet of Things ChallengesOpportunities and Directions

Emiliano Sisinni Member IEEE Abusayeed Saifullah Member IEEE Song Han Member IEEE UlfJennehag Member IEEE and Mikael Gidlund Senior Member IEEE

AbstractmdashInternet of Things (IoT) is an emerging domainthat promises ubiquitous connection to the Internet turningcommon objects into connected devices The IoT paradigm ischanging the way people interact with things around them Itpaves the way to creating pervasively connected infrastructuresto support innovative services and promises better flexibility andefficiency Such advantages are attractive not only for consumerapplications but also for the industrial domain Over the last fewyears we have been witnessing the IoT paradigm making its wayinto the industry marketplace with purposely designed solutionsIn this paper we clarify the concepts of IoT Industrial IoTand Industry 40 We highlight the opportunities brought in bythis paradigm shift as well as the challenges for its realizationIn particular we focus on the challenges associated with theneed of energy efficiency real-time performance coexistenceinteroperability and security and privacy We also provide asystematic overview of the state-of-the-art research efforts andpotential research directions to solve Industrial IoT challenges

Index TermsmdashIndustrial Internet of Things (IIoT) WirelessSensor Network (WSN) Real-time communication ReliabilitySecurity

I INTRODUCTION

Internet of Things (IoT) is a computing concept describingubiquitous connection to the Internet turning common objectsinto connected devices The key idea behind the IoT conceptis to deploy billions or even trillions of smart objects capableto sense the surrounding environment transmit and processacquired data and then feedback to the environment It isexpected that by the year 2021 there will be around 28 billionconnected devices [1] Connecting unconventional objects tothe Internet will improve the sustainability and safety ofindustries and society and enable efficient interaction betweenthe physical world and its digital counterpart ie what isusually addressed as a Cyber-physical System (CPS) IoT isusually depicted as the disruptive technology for solving mostof present-day society issues such as smart cities intelligenttransportation pollution monitoring connected healthcare toname a few As a subset of IoT (see Fig 1) IndustrialIoT (IIoT) covers the domains of machine-to-machine (M2M)and industrial communication technologies with automation

Emiliano Sisinni is with the Department of Information EngineeringUniversity of Brescia 25123 Brescia Italy e-mail emilianosisinniunibsit

Abusayeed Saifullah is with the Department of Computer Science WayneState University Detroit USA e-mail saifullahwayneedu

Song Han is with the Department of Computer Science and EngineeringUniversity of Connecticut Storrs USA e-mail songhanengruconnedu

Ulf Jennehag and Mikael Gidlund are with Department of InformationSystems and Technology Mid Sweden University SE-851 70 SundsvallSweden e-mail firstnamelastnamemiunse

applications IIoT paves the way to better understanding ofthe manufacturing process thereby enabling efficient andsustainable production

Flexibility and scalability required by IoT communicationsare typically addressed using wireless links In the pastwireless technologies in industrial applications were mostlybased on ad hoc solutions eg individually developed forconnecting moving parts or hard-to-reach devices Only re-cently standards purposely designed for the industry (egWirelessHART [2] and ISA10011a [3]) were released How-ever they face limitations in terms of scalability and coveragewhen very large areas need to be covered While cellulartechnologies such as 345G technologies promise to connectmassive devices over long distances they require infrastructuresupport and licensed band [4] IIoT applications typicallyrequire relatively small throughput per node and the capacityis not a main concern Instead the need of connecting a verylarge number of devices to the Internet at low cost withlimited hardware capabilities and energy resources (eg smallbatteries) make latency energy efficiency cost reliability andsecurityprivacy more desired features [5]

Meeting the above mentioned requirements poses a numberof key challenges on the evolution of IIoT Addressing thesechallenges is critical in order to ensure a massive roll-outof IIoT technologies In this paper we clarify the conceptsof IoT IIoT and the current trend of automation and dataexchange in manufacturing technologies called Industry 40We highlight the opportunities brought in by IIoT as well asthe challenges for its realization In particular we focus onthe challenges associated with the need of energy efficiencyreal-time performance coexistence interoperability and withthe security and privacy issues We also provide a systematicoverview of the state-of-the-art research efforts and potentialfuture research directions to address Industrial IoT challenges

The rest of this paper is organized as follows Section IIcompares IoT IIoT and Industry 40 Section III providesan overview of the recent activities on the definition of theIIoT architecture protocol stack as well as the standardizationefforts Section IV describes opportunities that IIoT will offerand challenges that have to be solved Finally we give someconcluding remarks in Section V

II IOT IIOT AND INDUSTRY 40

IoT IIoT and Industry 40 are closely related concepts butcannot be interchangeably used In this section we provide arough classification of these terms Regarding the IoT several




definitions exist each one trying to capture one of its funda-mental characteristics It is often considered as a sort of webfor the machines highlighting the aim of allowing things toexchange data However application fields are so diverse thatsome requirements (especially those related to communicationaspects) can be very different depending on the intendedgoals and end-users the underlying business models and theadopted technological solutions What is usually addressed asIoT could be better named as consumer IoT as opposed toindustrial IoT [6] [7]

Consumer IoT is human-centered the ldquothingsrdquo are smartconsumer electronic devices interconnected with each otherin order to improve human awareness of the surroundingenvironment saving time and money In general consumerIoT communications can be classified as machine-to-user andin the form of client-server interactions

On the other hand in the industrial world we are assisting tothe advent of the digital and smart manufacturing which aimat integrating Operational Technology (OT) with InformationTechnology (IT) domains [8] In very few words the IIoT (thebasic pillar of digital manufacturing) is about connecting allthe industrial assets including machines and control systemswith the information systems and the business processes Asa consequence the large amount of data collected can feedanalytics solutions and lead to optimal industrial operationsOn the other hand smart manufacturing obviously focuses onthe manufacturing stage of (smart) products life-cycle with thegoal of quickly and dynamically respond to demand changesTherefore the IIoT affects all the industrial value chain andis a requirement for smart manufacturing

As underlined in the following communication in IIoT ismachine oriented and can range across a large variety ofdifferent market sectors and activities The IIoT scenariosinclude legacy monitoring applications (eg process moni-toring in production plants) and innovative approaches forself-organizing systems (eg autonomic industrial plant thatrequires little if any human intervention) [9]

While the most general communication requirements of IoTand IIoT are similar eg support for the Internet ecosystemusing low-cost resource-constrained devices and network scal-ability many communication requirements are specific to eachdomain and can be very different eg Quality of Service(QoS) (in terms of determinism latency throughput etc) theavailability and reliability and the security and privacy IoTfocuses more on the design of new communication standardswhich can connect novel devices into the Internet ecosystemin a flexible and user-friendly way By contrast the currentdesign of IIoT emphasizes on possible integration and inter-connection of once isolated plants and working islands or evenmachineries thus offering a more efficient production and newservices [9] For this reason compared with IoT IIoT can beconsidered more an evolution rather than a revolution TableI gives a qualitative comparison of these technologies

Regarding the connectivity and criticality IoT is more flex-ible allowing ad hoc and mobile network structures and hav-ing less stringent timing and reliability requirements (exceptfor medical applications) On the other hand IIoT typicallyemploys fixed and infrastructure-based network solutions that

TABLE ICOMPARISON BETWEEN CONSUMER IOT AND INDUSTRIAL IOT

Consumer IoT Industrial IoT

Impact Revolution Evolution

Service Model Human-centered Machine-oriented

Current Status New devices and stan-dards

Existing devicesand standards

Connectivity Ad-Hoc (infrastructureis not tolerated nodescan be mobile)

Structured (nodesare fixed central-ized network man-agement)

Criticality Not stringent (exclud-ing medical applica-tions)

Mission critical(timing reliabilitysecurity privacy)

Data Volume Medium to High High to Very High

are well designed to match communication and coexistenceneeds In IIoT communications are in the form of machine-to-machine links that have to satisfy stringent requirements interms of timeliness and reliability Taking process automationas an example domain where process monitoring and controlapplications can be grouped into three sub-categories moni-toringsupervision closed loop control and interlocking andcontrol While monitoring and supervision applications are lesssensitive to packet loss and jitter and can tolerate transmissiondelay at second level closed loop control and interlocking andcontrol applications require bounded delay at millisecond level(10-100ms) and a transmission reliability of 9999 [5]

Comparing the data volume the generated data from IoT isheavily application dependent while IIoT currently targets atanalytics eg for predictive maintenance and improved logis-tics This implies that very large amount of data are exchangedin IIoT For example it is reported that the Rio Tinto minegenerates up to 24TB of data per minute according to CiscoGlobal Cloud Index

The concept of Industry 40 (where 40 represents the fourthindustrial revolution) arises when the IoT paradigm is mergedwith the Cyber-Physical Systems (CPSs) idea [10] Originallydefined in Germany the Industry 40 concept has gained aglobal visibility and it is nowadays universally adopted foraddressing the use of Internet technologies to improve produc-tion efficiency by means of smart services in smart factoriesCPSs extend real-world physical objects by interconnectingthem altogether and providing their digital descriptions Suchinformation stored in models and data objects that can beupdated in real time represents a second identity of the objectitself and constitutes a sort of ldquodigital twinrdquo Thanks to thedynamic nature of these digital twins innovative services thatwere not possible in the past can be implemented acrossthe whole product lifecycle from inception to disposal ofmanufactured products In summary IIoT is a subset of IoTwhich is specific to industrial applications The manufacturingphase of the product lifecycle is where the IoT and Industry40 meet originating to the IIoT Figure 1 shows intersectionsof IoT CPS IIoT and Industry 40

As a concluding remark it has to be highlighted thatthe IIoT paradigm is not intended for substituting traditionalautomation applications but aims at increasing the knowledge




IIoTIoT CPS

Ind

ustr

y 4

0

Fig 1 IoT CPS IIoT and Industry 40 in Venn Diagram

about the physical system of interest As a consequence theIIoT (at least today) is not related to control applications atthe field level where bounded reaction time (ie determinism)must be ensured On the contrary as previously stated IIoT ap-plications including supervision optimization and predictionactivities are typically grouped into the so called Digital orCloud Manufacturing (CM) The growing interest toward thistopic is confirmed by the wide range of literature A surveyabout CM is reported in [11] In the past the supervisionactivities were dominated by the man but efficient machine tomachine communications make human intervention superflu-ous and extend the operating range to geographical scale Forinstance the availability of reliable short latency connectionson such a large scale may increase the revenue [12] Thework in [13] highlights the importance of real-time large-scaleapproach for equipment maintenance applications An IIoT-based dynamic production logistics architecture is presentedin [14] for real-time synchronization of internal and publicproduction logistics resources In [15] the optimization ofproduction scheduling is based on IIoT decentralized energyprediction algorithms fed by the current state of the machinesAs a concluding remark the progressive reduction of latencyand jitter of Internet-based connectivity will increase the rangeof possible applications as reported in [16]

III STATE OF THE ART

As IIoT interconnects a large number of components lever-aging sensing communication and data processing technolo-gies it is not possible to have a comprehensive descriptionof all the recent advancements in such a diverse field How-ever some foundational aspects can be highlighted ie thearchitecture the connectivity and the standardization

A The IIoT architecture

A reference architecture is a higher level of abstractiondescription that helps identify issues and challenges for dif-ferent application scenarios The design of a IIoT architectureneeds to highlight extensibility scalability modularity andinteroperability among heterogeneous devices using differenttechnologies Several reference architecture frameworks orig-inated in the past in different application contexts for both

IoT and IIoT [17] The typically adopted approach is a multi-layer description organized around the services offered at eachlevel depending on the selected technologies business needsand technical requirements For instance the InternationalTelecommunication Union (ITU) supports an IoT architecturemade of five layers sensing accessing networking middle-ware and application layers Jia et al [18] Domingo [19] andAtzori et al [20] suggested the identification of three majorlayers for IoT perception layer (or sensing layer) networklayer and service layer (or application layer) Liu et al [21]designed an IoT application infrastructure that contains thephysical layer transport layer middleware layer and applica-tions layer In [22] a four-layered architecture is derived fromthe perspective of offered functionalities that includes thesensing layer the networking layer the service layer and theinterface layer The Reference Architectural Model Industrie40 (RAMI 40) [23] focuses on next-generation industrialmanufacturing systems it identifies a 3-D model whose axesare the Life Cycle amp Value Stream related to productslife cycle and the Hierarchy Levels related to the differentcomponent functionalities The Hierarchy axis describes theIT representative and includes a communication layer

Recently the Industrial Internet Consortium released theldquoReference Architecturerdquo document [24] In particular itfocuses on different viewpoints (formally business usagefunctional and implementation views) and provides modelsper each one The implementation viewpoint is focused onthe technologies and the system components that are requiredfor implementing the functionalities prescribed by the usageand functional viewpoints Thus it provides not only thedescription of the IIoT system general architecture (ie itsstructure and the distribution of components and the topologyby which they are interconnected) but includes a descriptionof interfaces and protocols as well Roughly speaking twodifferent kinds of information are transferred in IIoT systemsdepending on if the data have to be processed yet (data flow)or they are the results of some elaborations (control flow)

Some architectural patterns are also emerging and provid-ing coherent system implementations and paving the way toinnovative business models and services usually in a multiple-tier arrangement dictated by the very heterogeneous devicesand networks In the widely accepted three-tier pattern [25]edge platform and enterprise tiers are connected by proximityaccess and service networks The edge defines the domain inwhich IIoT components interact one with each other Thusit consists of sensors controllers actuators interconnectedby independent local area networks (the proximity networksusually in the form of fieldbuses) to an edge gateway whichin turn connects to larger networks (access network) of theplatform tier providing global coverage Finally the platformtier leverages on the service network to establish links with theenterprise tier that implements domain-specific applicationsand provides end user interfaces The Fig 2) tries to graphi-cally depict the complexity of the IIoT hybrid architecture inparticular the increased latency and data aggregation of thedifferent tiers is highlighted




ENTERPRISE busines

and application domains

(analytics management

archive UIs)

EDGE monitoring

control safety

domains

PLATFORM provides a

secure shared message

bus (performs data

collection and

transformation)

IIoT Edge

Gateways

Proximity

Networks

TIERS

Actuators Sensors Actuators

Controllers

Networks

Service

Fig 2 The three-tier IIoT architecture

B The IIoT Connectivity

The connectivity of todayrsquos IIoT varies depending on whichcombination of backbone and edge architecture is usefulin a given situation and a combination of wireless andorwired technologies is adopted A key goal is to avoid iso-lated systems based on proprietary solutions and enable datasharing and interoperability among these closed subsystems(brownfield) and the yet-to-come applications (greenfield)within and across industries Neither the seven-layer OpenSystems Interconnect (OSI) nor the five-layer Internet modelis adequate to take into account the distributed nature ofsensors controllers gateways and other components involvedin IIoT and different layering is required The IIoT initiativesfeasibility requires communication protocols able to supportefficient timely and ubiquitous information aggregation andavailability Lower levels of the stack must adequately respondto scalability and flexibility requirements Upper levels mustallow so called ldquosmart devicesrdquo (ie offering both computationand communication capabilities) to transport ldquosmart datardquonot limited to the information of interest but also providingawareness of the users they are intended to and all the semanticrules to be correctly understood at destinations as well Threemacro layers can be identified ie networking (dealing withframes and packets) connectivity (dealing with messages)and information (dealing with end-user data structures) Theprotocol heterogeneity of the IIoT is mirrored in a hourglass-shaped stack (see Fig 3) The neck is represented by thenetwork layer ie the Internet (and its different flavors asIPv4 IPv6 6LowPAN RPL etc) but above and belowsublayers are not yet clearly defined despite they are of criticalimportance for ensuring interoperability at different levels

Additionally it is worth mentioning that most of currentindustrial applications exploit fieldbuses each having its ownecosystem thus providing poor interoperability Fieldbuses arevertical solutions covering most of the functionalities of thecommunication stack Fortunately latest technologies (eg themany different flavors of the real-time Ethernet solutions)natively adopt Ethernet and IP protocols thus making it easierto provide technical interoperability ie the ability to sharepackets in a common format [26] Due to its full IP compatibil-

NETWORKING

CONNECTIVITY

INFORMATION

Number of

Protocols

Protocol age

and mutability

Fig 3 The hourglass-shaped IIoT protocol stack

ity and incorporation of the Common Industrial Protocol (CIP)and reliance on standard Internet and Ethernet technology(IEEE 8023 combined with the TCPIP Suite) EtherNetIPmakes itself a particularly suitable for IIoT As an examplemyriads of motion applications in industries feature a bevy ofconnected components ndash from IO blocks and vision sensorsto servo and variable frequency drives EtherNetIP can uniteall of these moving parts via CIP communications running onEthernet [27] Since it is built on the IP suite EtherNetIP isgaining momentum from the development and refinement ofassociated protocols In addition to TCPUDP at the transportlayer it can access higher-level functionality through HTTPConnectivity between industrial equipment Ethernet networksand the Internet can enable time-sensitive communicationsto streamline plant operations thereby enabling real-timemanufacturing for enterprises with global supply chains

1) Stack Lower Layers In IIoT stack the lowest layer isthe physical one which refers to the exchange of physicalsignals on media linking the participants Above it lies thelink layer which connects adjacent participants allowing toexchange frames by means of signaling protocols It hasto be noticed that the already available solutions explicitlydesigned for the industrial market have some limits Well-accepted standards defined in the IEC62591 and IEC62743(commercially known as WirelessHART and ISA10011a) arebased on IEEE802154 compliant radio and are not designedto connect a large number of devices as in typical IIoTapplications Consequently several independent networks mustbe deployed each one with its own IIoT gateway On thecontrary Low-Power Wide-Area Network (LPWAN) solutionsare gaining momentum in recent years for occupying the lowertwo levels of the protocol stack with multiple competingtechnologies being offered or under development [28]

LPWANs allow to communicate over long distances (severalkilometers) at very low transmission power SigFox [29] and




LoRaWAN [30] are two of the most interesting proposals [31]ndash[33] However SigFox based on ultra narrowband technology(ie communication channels with a bandwidth on the orderof 100Hz) is mainly intended for smart city applicationseg smart metering since a device can send at most 140messages per day each one typically having 3s air timeThus it is not suitable for many industrial applications that re-quire real-time performance or frequent sampling LoRaWAN(maintained by the LoRa alliance) leverages on proprietaryLoRa radios and offers 125kHz or 250kHz-wide channelsand low data rate (from about 10kbps down to less than400bps) It has been demonstrated that by mimicking thetime-slotted channel hopping of typical wireless industrialcommunications thousand of communication opportunitiesper second are affordable [34] As a final remark it has tobe highlighted that LPWANs generally operate in the sub-GHz region that ensures good coverage but is often limitedby duty cycled transmission of 1 or 01 or ldquoListen BeforeTalkrdquo (LBT) medium access strategy Also both SigFoxand LoRaWAN are primarily uplink-only LoRaWAN canenable bidirectional communication but it has to rely on timesynchronized beacons and schedules which is an overheadThe recently developed SNOW [35]ndash[38] is an LPWAN thatenables concurrent bidirectional communications thus makingit suitable for control applications However SNOW operatesover the TV white spaces and thus its performance dependson the availability of white spaces

The use of unlicensed spectrum has raised certain reliabilityissues since there is no guarantee of service availability inaddition to the aforementioned duty-cycle and LBT regula-tions For this reason fifth generation cellular access (5G)is often envisioned as a viable IIoT solution in additionto regular telecommunication applications using the cellularinfrastructure Currently there is no finalized standard for5G (which actually is an umbrella for many specifications)However the cost of technical solutions to be applied at thephysical layer to satisfy industrial needs can be an importantissue Only a sound business model and a strong argument forusing licensed frequency bands (both missing today) couldbring market acceptance within industrial automation for 5G[4] Narrowband LPWAN technology standard to operate oncellular infrastructure and bands as NB-IoT received attentionrecently but despite its potential there are some issues re-garding scalability and network resource slicing between IoTapplications and other broadband services that need furtherstudies [39] In licensed cellular spectrum EC-GSM-IoT [40]and LTE Cat M1 (LTE-Advanced Pro) [41] are also underdevelopment A key requirement of all these technologicalsolutions is that they need cellular infrastructure

Bluetooth low energy (BLE) [42] is another interestingalternative for IIoT since it offers ultra-low power consump-tion but the initial doubts for BLE was due to its rangelimitations since it only supports star network and limitednumber of devices [43] To overcome those limitations BLEmesh networking standard was recently released and initiallyconsidered for home automation The main challenge withBLE mesh networking targeting real-time communication isthat the connection establishment procedure introduces a long

delay (eg several hundred ms) To overcome this problemmany upper layer protocols such as mesh and beacon try toleverage on the connection-less scheme since there is no needto establish connections before sending data However thisdoes not ensure reliable communication due to lack of a goodmedium access control Besides the throughput is much lowerthan 1 Mbps since there is a limitation of sending packet in thisbearer ie at least 20 ms interval is required in order to reduceintra-interference and avoid collisions Recently there has beensome interesting work about using BLE mesh networking forreal-time communication targeting low latency applications inindustrial automation In [44] the authors presented a real-time protocol aimed to overcome the problem with rangelimitations of mesh technology and support bounded real-timetraffic Their protocol exploits time division multiple access(TDMA) with an optimized transmission allocation to providedata packets with real-time support It works on standard BLEdevices In [45] the authors presented a bandwidth reservationmechanism for partitioning the radio transceiver between twoprotocols namely the BLE and a real-time custom protocol

2) Stack Upper Layers The aim of upper layers of the IIoTstack is to facilitateensure so called syntactic interoperabilityie the capability to use a common data structure and set ofrules for information exchanges [46] [47] It is the actualapplication that finally provides the semantic interoperabilityie the capability to interpret exchanged data unambiguously[26] In light of this requirement the Industrial Internet Con-sortium proposed to separate upper layer protocols into justtwo levels the lower is occupied by the transport layer thatis in charge of exchanging variable length messages amongthe involved applications the upper constitutes the frameworklayer which manages the transfer of structured data havinghigher abstraction (eg state events streams etc) Accordingto this classification the transport layer is loosely relatedto the transport layer of OSI (and Internet) model indeedUDP and TCP are foundations for other transport protocolsHowever some functionalities of the session presentation andapplication layers are included as well

A well-accepted and widely used solution for implementinghorizontal integration relies on messaging protocols (oftenimplemented by message oriented middleware) These pro-tocols support the publishersubscriber paradigm where bothsides of the actual data exchange are in general not directlyconnected The application that wants to publish a messageconnects to a so-called message queue broker for placing it ina queue subsequently subscribers automatically receive themessage as a push notification The delivering modality issaid to be persistent if it survives a broker failure Messagingsolutions ensure scalability since the applications do not haveto know each other Today a prevailing messaging protocol isMQTT (Message Queue Telemetry Transport) standardizedby the OASIS alliance A different approach relies on re-questresponse data delivery and synchronous or asynchronousdata exchanges are permitted In the synchronous data ex-changes the requestor waits for replies before issuing thenext request In an asynchronous case the reply is returnedat some unknown later time to the requestor A well-knownexample of requestresponse protocol is CoAP (Constrained




Application Protocol) defined by the the IETF ConstrainedRESTful Environments (CORE) working group [47]

The framework layer provides services to the above appli-cation and manages the lifecycle of any piece of data fromthe creation to the deletion Protocols at this level offer theability to discover and identify data objects and can understandthe transported data meaning (ie are not opaque) Thisawareness is exploited for optimally delivering the informationat the destination The open platform communications - unifiedarchitecture (OPC-UA a multi-part document set managedby the OPC foundation formally known as the IEC62541)is an example of such a framework It describes a ServiceOriented Architecture (SOA) based on clientserver architec-ture in which the server models data information processesand systems as objects that are presented to clients togetherwith services that the client can use

C The Standardization of IIoT

Standardization is an important step for a technology tobe widely supported and well-accepted Interesting to notemost of the past standardization activities focused on veryspecific domains thus resulting in disjoint and somewhat re-dundant development The standardization process has to faceseveral challenges currently there is a plethora of competingstandardization bodies and consortia initiatives at every layerof the IIoT stack referring to a variety of fragmented ofteninconsistent and opponent requirements Obviously such anapproach is detrimental to IIoT whose fundamental aim isto bring together and share information coming from veryheterogeneous things The actual fragmentation is effectivelyhighlighted by the ETSI technical report ETSI TR 103375whose aim is to provide the roadmaps of the IoT standardsGenerally speaking the ongoing standardization activities in-clude horizontal standards aiming at ensuring interoperabilityvertical standards aiming at identifying requirements of indi-vidual applications and use cases and promotional activitiessupported by industrial consortia and government groups

Focusing on industrial applications the most significant andimportant efforts are those carried out by the IEC (Inter-national Electrotechnical Commission) which created manydifferent Study Groups and Technical Committees on thesubject and published a couple of white papers about IIoTand the smart factory with the aim of assessing potentialglobal needs benefits concepts and pre-conditions for thefactory of the future It is worth noting that regarding theconnectivity issues the aforementioned IEC62541 is the onlystandard originated in the industrial vertical context

Standardization activities for 5G targeting IIoT and crit-ical communication is ongoing in 3GPP and falls underthe umbrella of Ultra reliable Low Latency Communications(URLLC) with the aim of providing 1 ms latency One wayto reduce the latency in URLLC is to provide a reliabletransmission time interval (TTI) operation

Considering that a relevant part of IIoT communications willprobably be implemented as wireless links coexistence issuesarise as well The IEC62657 provides a sort of glossary ofindustrial automation requirements for harmonizing concepts

and terms of the telecommunication world and defines coex-istence parameters (in the form of templates) and guidelinesfor ensuring wireless coexistence within industrial automationapplications along the whole lifecycle of the plant

IV OPPORTUNITIES AND CHALLENGES

A key reason for adopting IIoT by manufacturers utilitycompanies agriculture producers and healthcare providers is toincrease productivity and efficiency through smart and remotemanagement As an example Thames Water [48] the largestprovider of drinking and waste-water services in the UK isusing sensors and real-time data acquisition and analyticsto anticipate equipment failures and provide fast response tocritical situations such as leaks or adverse weather eventsThe utility firm has already installed more than 100000 smartmeters in London and it aims to cover all customers withsmart meters by 2030 With more than 4200 leaks detectedon customer pipes so far this program has already savedan estimated 930000 liters of water per day across LondonAs another example the deployment of 800 HART devicesfor real-time process management at Mitsubishi chemicalplant in Kashima Japan has been increasing the productionperformance by saving US$20-30000 per day that also averteda $3million shutdown [49]

Precision agriculture powered by IIoT can help farmersbetter measure agricultural variables such as soil nutrientsfertilizer used seeds planted soil water and temperature ofstored produce allowing to monitor down to the square footthrough a dense sensor deployment thereby almost doublingthe productivity [50]ndash[52] Companies like Microsoft (Farm-Beats project [53] [54]) Climate Corp [55] ATampT [56] andMonsanto [57] are promoting agricultural IoT IIoT can alsosignificantly impact the healthcare field In hospitals human ortechnological errors caused by false alarms slow response andinaccurate information are still a major reason of preventabledeath and patient suffering By connecting distributed medicaldevices using IIoT technologies hospitals can significantlyovercome such limitations thereby improving patient safetyand experiences and more efficiently using the resources

IIoT also provides opportunities to enhance efficiencysafety and working conditions for workers For exampleusing unmanned aerial vehicles (UAVs) allows inspecting oilpipelines monitoring food safety using sensors and mini-mizing workersrsquo exposure to noise and hazardous gases orchemicals in industrial environments Schlumberger for ex-ample is now monitoring subsea conditions using unmannedmarine vehicles which can travel across oceans collecting datafor up to a year without fuel or crew moving under powergenerated from wave energy [58] Through remote monitoringand sensing powered by IIoT mining industries can dramati-cally decrease safety-related incidents while making mining inharsh locations more economical and productive For exampleRio Tinto a leading mining company intends its automatedoperations in Australia to preview a more efficient future forall of its mines to reduce the need for human miners [59]

Despite the great promise there are many challenges inrealizing the opportunities offered by IIoT which should be




addressed in the future research The key challenges stemfrom the requirements in energy-efficient operation real-timeperformance in dynamic environments the need for coexis-tence and interoperability and maintaining the security of theapplications and usersrsquo privacy as described below

A Energy Efficiency

Many IIoT applications need to run for years on batteriesThis calls for the design of low-power sensors which do notneed battery replacement over their lifetimes This creates ademand for energy-efficient designs To complement such de-signs upper-layer approaches can play important roles throughenergy-efficient operation Many energy efficient schemes forwireless sensor network (WSN) have been proposed in recentyears [60] but those approaches are not immediately applica-ble to IIoT IIoT applications typically need a dense deploy-ment of numerous devices Sensed data can be sent in queriedform or in a continuous form which in a dense deploymentcan consume a significant amount of energy Green networkingis thus crucial in IIoT to reduce power consumption andoperational costs It will lessen pollution and emissions andmake the most of surveillance and environmental conservationLPWAN IoT technologies achieve low-power operation usingseveral energy-efficient design approaches First they usuallyform a star topology which eliminates the energy consumedthrough packet routing in multi-hop networks Second theykeep the node design simple by offloading the complexitiesto the gateway Third they use narrowband channels therebydecreasing the noise level and extending the transmissionrange [35] [61]

Although there are numerous methods to achieve energyefficiency such as using lightweight communication protocolsor adopting low-power radio transceivers as described abovethe recent technology trend in energy harvesting providesanother fundamental method to prolong battery-life Thusenergy harvesting is a promising approach for the emergingIIoT Practically energy can be harvested from environmentalsources namely thermal solar vibration and wireless radio-frequency (RF) energy sources Harvesting from such envi-ronmental sources is dependent on the presence of the corre-sponding energy source However RF energy harvesting mayprovide benefits in terms of being wireless readily available inthe form of transmitted energy (TVradio broadcasters mobilebase stations and hand-held radios) low cost and in terms ofsmall form factor of devices

B Real-Time Performance

IIoT devices are typically deployed in noisy environmentsfor supporting mission- and safety-critical applications andhave stringent timing and reliability requirements on timelycollection of environmental data and proper delivery of controldecisions The QoS offered by IIoT is thus often measured byhow well it satisfies the end-to-end (e2e) deadlines of the real-time sensing and control tasks executed in the system [62][63]

Time-slotted packet scheduling in IIoT plays a critical rolein achieving the desired QoS For example many industrial

wireless networks perform network resource management viastatic data link layer scheduling [64]ndash[71] to achieve de-terministic e2e real-time communication Such approachestypically take a periodic approach to gathering the networkhealth status and then recompute and distribute the updatednetwork schedule information This process however is slownot scalable and incurs considerable network overhead Theexplosive growth of IIoT applications especially in terms oftheir scale and complexity has dramatically increased the levelof difficulty in ensuring the desired real-time performance Thefact that most IIoT must deal with unexpected disturbancesfurther aggravate the problem

Unexpected disturbances can be classified into externaldisturbances from the environment being monitored and con-trolled (eg detection of an emergency sudden pressure ortemperature changes) and internal disturbances within thenetwork infrastructure (eg link failure due to multi-userinterference or weather related changes in channel SNR) Inresponse to various internal disturbances many centralizedscheduling approaches [72]ndash[77] have been proposed Thereare also a few works on adapting to external disturbances incritical control systems For example rate-adaptive and rhyth-mic task models are introduced in [78] and [79] respectivelywhich allow tasks to change periods and relative deadlines insome control systems such as automotive systems

Given the requirement of meeting e2e deadlines the afore-mentioned approaches for handling unexpected disturbancesare almost all built on a centralized architecture Hencemost of them have limited scalability [80] The concept ofdistributed resource management is not new In fact distributedapproaches have been investigated fairly well in the wirelessnetwork community (eg [81]ndash[85]) However these studiestypically are not concerned with real-time e2e constraintsA few which consider real-time constraints mainly focuson soft real-time requirements and do not consider externaldisturbances that IIoT must have to deal with Only recentlywe have started to see some hybrid and fully distributedresource management approaches for IIoT [86] [87] Howeverhow to ensure bounded response time to handle concurrentdisturbances is still an open problem

C Coexistence and InteroperabilityWith the rapid growth of IIoT connectivity there will be

many coexisting devices deployed in close proximity in thelimited spectrum This brings forth the imminent challengeof coexistence in the crowded ISM bands Thus interferencebetween devices must be handled to keep them operationalExisting and near future IIoT devices will most likely havelimited memory and intelligence to combat interference orkeep it to a minimum While there exists much work on wire-less coexistence considering WiFi IEEE 802154 networksand Bluetooth (see surveys [88]ndash[91]) they will not work wellfor IIoT Due to their dense and large-scale deployments thesedevices can be subject to an unprecedented number of inter-ferers Technology-specific features of each IIoT technologymay introduce additional challenges

To ensure good coexistence it will become important thatfuture IIoT devices can detect classify and mitigate exter-




nal interference Recently some work regarding classifyinginterference via spectrum sensing [92] on IIoT devices hasbeen presented but most of the existing work fails sincea very long sampling window is needed and the proposedspectrum sensing methods need much more memory than whatis available in existing commercial IIoT devices Hence in[93] a promising method was presented and implemented inCrossbowrsquos TelosB mote CA2400 which is equipped withTexas Instrument CC2420 transceiver That method managesto classify external interference by using support vector ma-chines with a sensing duration below 300 ms Moreoverexisting devices based on IEEE 802154 standards do not haveany forward error correcting (FEC) capabilities to improvethe reliability of the data packet There exists some work thatinvestigated error control codes for industrial WSNs and theresults clearly show that FEC will improve reliability andthe coexistence [94]ndash[96] However most of the availableFEC methods are optimized for long packets Given thatIIoT communication will mainly consist of short packets(50-70 bytes) and many applications are time-critical moreresearch is needed to find good error correcting codes for IIoTcommunication [97] If the research of error correcting codesfor IIoT devices should be successful it is also important thatmore emphasis be given on investigating and understanding thecomplex radio environment where many of these IIoT deviceswill be deployed [98] [99]

The rapid growth of IIoT technologies also brings forththe requirements of interoperability Namely in the future afully functional digital ecosystem will require seamless datasharing between machines and other physical systems fromdifferent manufacturers The lack of interoperability amongIIoT devices will significantly increase the complexity andcost of IIoT deployment and integration The drive towardsseamless interoperability will be further complicated by thelong life span of typical industrial equipment which wouldrequire costly retrofitting or replacement to work with thelatest technologies

The challenges of device diversity in IIoT can be addressedalong three dimensions multimode radios software flexibil-ity cross-technology-communication [100] Multimode radiosallow diverse IIoT devices to talk to each other Softwareflexibility enables support for multiple protocols connectivityframeworks and cloud services Recently cross-technology-communication [101] without the assistance of additionalhardware has been studied for communication across WiFiZigBee and Bluetooth devices Such approaches are specificto technologies and thus future research is needed to enablecross-technology-communication in IIoT devices

D Security and Privacy

Besides the requirements of energy-efficiency and real-time performance security is another critical concern in IIoTIn general IIoT is a resource-constrained communicationnetwork which largely relies on low-bandwidth channels forcommunication among lightweight devices regarding CPUmemory and energy consumption [102] For this reasontraditional protection mechanisms are not sufficient to secure

the complex IIoT systems such as secure protocols [103]lightweight cryptography [104] and privacy assurance [105]To secure the IIoT infrastructure existing encryption tech-niques from industrial WSNs may be reviewed before appliedto build IIoT secure protocols For instance scarce computingand memory resources prevent the use of resource-demandingcrypto-primitives eg Public-Key Cryptography (PKC) Thischallenge is more critical in the applications of massive dataexchanged with real-time requirements To address privacy andsecurity threats in IIoT one can argue for a holistic approachas pointed out in [106] This means that aspects such asplatform security secure engineering security managementidentity management and industrial rights management mustbe taken into account throughout the whole life cycle of thesystems and products

There exist several security properties to consider whendesigning secure IIoT infrastructure [107]

1) IIoT devices need to be tamper resistant against potentialphysical attacks such as unauthorized re-programmingand passive secret stealing while allowing the authorizedusers to update the security firmware on the device

2) The storage of IIoT device should be protected againstadversary by keeping the data encrypted to keep theconfidentiality

3) The communication network among the IIoT devicesshould be secured to keep confidentiality and integrity

4) The IIoT infrastructure needs efficient identification andauthorization mechanisms so that only authorized enti-ties can access the IIoT resource

5) The system should be available within normal opera-tion even with the physical damage to the devices bymalicious users This guarantees the robustness of IIoT

Typically symmetric-key cryptography can provide alightweight solution for IIoT devices However both the keystorage and the key management are big issues if usingsymmetric-key encryption especially when considering low-capacity devices

Additionally if one device in IIoT is compromised it mayleak all other keys Public-key cryptography generally providesmore secure features and low storage requirements but suffersfrom high computational overhead due to complex encryptionThus reducing the overhead of complex security protocols forpublic-key cryptosystems remains a major challenge for IIoTsecurity In PKC Elliptic-Curve Cryptography (ECC) providesa lightweight solution regarding computational resources Itprovides a smaller key size reducing storage and transmissionrequirements

In IIoT systems it is important to provide the identificationto get the legal access The secure IIoT infrastructure mustensure the object identification regarding the integrity ofrecords used in the naming systems such as Domain NameSystem (DNS) The DNS system can provide name translationservices to the Internet user however it is in an insecure waywhich remains vulnerable to various attacks by deliberatedadversary [108] This challenge stays valid even for a boundedand closed environment Thus without the integrity protectionof the identification the whole naming system is still insecureSecurity extensions to DNS like Domain Name Service




Security Extension (DNSSEC) increases security and is doc-umented in IETF RFC4033 [109] However due to its highcomputation and communication overhead it is challenging todirectly apply DNSSEC to the IIoT infrastructure

IIoT devices should follow specific schemes and rules forauthentication to exchangepublish their data Due to the re-source constraints of the IIoT devices low-cost authenticationschemes have not been provided as much as needed [110]Although public-key cryptography systems provide the meth-ods for constructing authentication and authorization schemesit fails to provide a global root certification authority (globalroot CA) which largely hinders many theoretically feasibleschemes from actually being deployed Without providing theglobal root CA it becomes very challenging to design a secureauthentication system in IIoT Thus currently if we intend toprovide the secure authentication for IIoT devices we have touse the high-cost solutions which is a conflict with the maingoal of the lightweight principle of IIoT [111] Furthermoreit is a big challenge to issue a certification to each object inIIoT since the total number of objects could be huge

Privacy is a very broad and diverse concept Many defini-tions and perspectives have been provided in the literatureGenerally speaking privacy in IIoT is the threefold guaran-tee [112] for 1) awareness of privacy risks imposed by thingsand services 2) individual control over the collection andprocessing of information 3) awareness and control of subse-quent use and dissemination to any outside entity The majorchallenges for privacy lie in two aspects data collection pro-cess and data anonymization process Typically data collectionprocess deals with the collectible data and the access controlto these data during the data collection from smart thingsdata anonymization is a process to ensure data anonymitythrough both cryptographic protection and concealment of datarelations Due to the restrictions on the collection and storageof private information privacy preservation can be ensuredduring the data collection However given the diversity of thethings in data anonymization different cryptographic schemesmay be adopted which is a challenge to privacy preservingMeanwhile the collected information needs to be sharedamong the IIoT devices and the computation on encrypteddata is another challenge for data anonymization

V CONCLUSION

This paper presented an overview of the emerging IIoTsolutions What is proposed as a revolution for the consumermarket can be another step of the ever evolving industrialcommunications world Several technologies are involved andterms as IoT IIoT and Industry 40 are often misused Inthis paper we have provided a systematic overview of IIoTfocusing on the definition of its architecture and describing theprotocol ecosystem which is emerging from standardization ef-forts We have also discussed the challenges for its realizationBesides the QoS requirements that characterize industrial com-munications IIoT suffers from yet to be considered securitychallenges that stem from the high sensitivity of the managedinformation Furthermore typical IIoT applications have todeal with constrained resources (both power and computing)

and must be operative for extended periods of time ensuringavailability and reliability We have described the state-of-the-art research and standardization efforts and future researchdirections to address IIoT challenges

REFERENCES

[1] Ericsson ldquoCellular networks for massive iotrdquo January 2016 httpswwwericssoncomassetslocalpublicationswhite-paperswp iotpdf

[2] F Group ldquoWirelessHART specificationrdquo 2007 httpwwwhartcomm2org

[3] ldquoISA100 Wireless systems for automationrdquo httpwwwisaorgMSTemplatecfmMicrositeID=1134ampCommitteeID=6891

[4] M Gidlund T Lennvall and J Akerberg ldquoWill 5g become yet anotherwireless technology for industrial automationrdquo in IEEE InternationalConference on Industrial Technology (ICIT) 2017 pp 1319ndash1324

[5] J Akerberg M Gidlund and M Bjorkman ldquoFuture research chal-lenges in wireless sensor and actuator networks targeting industrialautomationrdquo in Proceedings of the 9th IEEE International Conferenceon Industrial Informatics 2011 pp 410ndash415

[6] M R Palattella M Dohler A Grieco G Rizzo J Torsner T Engeland L Ladid ldquoInternet of things in the 5G era Enablers architectureand business modelsrdquo IEEE Journal on Selected Areas in Communi-cations vol 34 no 3 pp 510ndash527 2016

[7] D Bandyopadhyay and J Sen ldquoInternet of things Applications andchallenges in technology and standardizationrdquo Wireless Personal Com-munications vol 58 no 1 pp 49ndash69 2011

[8] M R Palattella P Thubert X Vilajosana T Watteyne Q Wang andT Engel Internet of Things IoT Infrastructures Second InternationalSummit 2016

[9] L D Xu W He and S Li ldquoInternet of things in industries A surveyrdquovol 10 no 4 pp 2233ndash2243

[10] M Wollschlaeger T Sauter and J Jasperneite ldquoThe future of industrialcommunication Automation networks in the era of the internet ofthings and industry 40rdquo IEEE Industrial Electronics Magazine vol 11no 1 pp 17ndash27 2017

[11] W He and L Xu ldquoA state-of-the-art survey of cloud manufacturingrdquoInternational Journal of Computer Integrated Manufacturing vol 28no 3 pp 239ndash250 2015 [Online] Available httpsdoiorg1010800951192X2013874595

[12] I Lee ldquoAn exploratory study of the impact of the internetof things iot on business model innovation Building smartenterprises at fortune 500 companiesrdquo Int J Inf Syst SocChang vol 7 no 3 pp 1ndash15 Jul 2016 [Online] Availablehttpdxdoiorg104018IJISSC2016070101

[13] P OrsquoDonovan K Leahy K Bruton and D T J OrsquoSullivan ldquoAnindustrial big data pipeline for data-driven analytics maintenanceapplications in large-scale smart manufacturing facilitiesrdquo Journalof Big Data vol 2 no 1 p 25 Nov 2015 [Online] Availablehttpsdoiorg101186s40537-015-0034-z

[14] T Qu S P Lei Z Z Wang D X Nie X Chen and G Q HuangldquoIot-based real-time production logistics synchronization system undersmart cloud manufacturingrdquo The International Journal of AdvancedManufacturing Technology vol 84 no 1 pp 147ndash164 Apr 2016[Online] Available httpsdoiorg101007s00170-015-7220-1

[15] S G Pease R Trueman C Davies J Grosberg K H Yau N KaurP Conway and A West ldquoAn intelligent real-time cyber-physicaltoolset for energy and process prediction and optimisation in thefuture industrial internet of thingsrdquo Future Generation ComputerSystems vol 79 pp 815 ndash 829 2018 [Online] AvailablehttpwwwsciencedirectcomsciencearticlepiiS0167739X1630382X

[16] T H Szymanski ldquoSupporting consumer services in a deterministicindustrial internet core networkrdquo IEEE Communications Magazinevol 54 no 6 pp 110ndash117 June 2016

[17] M Weyrich and C Ebert ldquoReference architectures for the internet ofthingsrdquo IEEE Software vol 33 no 1 pp 112ndash116 2016

[18] X Jia Q Feng T Fan and Q Lei ldquoRfid technology and itsapplications in internet of things (iot)rdquo in Proceedings of the 2ndInternational Conference on Consumer Electronics Communicationsand Networks (CECNet) 2012 pp 1282ndash1285

[19] M C Domingo ldquoAn overview of the internet of things for people withdisabilitiesrdquo Journal of Network and Computer Applications vol 35no 2 pp 584ndash596 2012

[20] L Atzori A Iera and G Morabito ldquoThe internet of things A surveyrdquoComputer networks vol 54 no 15 pp 2787ndash2805 2010




[21] C H Liu B Yang and T Liu ldquoEfficient naming addressing andprofile services in internet-of-things sensory environmentsrdquo Ad HocNetworks vol 18 pp 85ndash101 2014

[22] L Da Xu W He and S Li ldquoInternet of things in industries A surveyrdquoIEEE Transactions on industrial informatics vol 10 no 4 pp 2233ndash2243 2014

[23] H Flatt S Schriegel J Jasperneite H Trsek and H AdamczykldquoAnalysis of the cyber-security of industry 40 technologies based onrami 40 and identification of requirementsrdquo in IEEE 21st Int Confon Emerging Tech and Factory Automation 2016 pp 1ndash4

[24] ldquoIndustrial internet reference architecturerdquo httpwwwiiconsortiumorgIIRAhtm

[25] IoT 2020 Smart and Secure IoT Platform International Electrotech-nical Commission 2016

[26] J Kiljander A Delia F Morandi P Hyttinen J Takalo-MattilaA Ylisaukko-Oja J P Soininen and T S Cinotti ldquoSemantic interop-erability architecture for pervasive computing and internet of thingsrdquoIEEE Access vol 2 pp 856ndash873 2014

[27] httpwwwindustrial-iporgenindustrial-ipethernet-ipethernetip-infographic

[28] D Ismail M Rahman and A Saifullah ldquoLow-power wide-areanetworks Opportunities challenges and directionsrdquo in Proceedingsof the Workshop Program of the 19th International Conference onDistributed Computing and Networking ser Workshops ICDCN rsquo182018 pp 81ndash86

[29] Sigfox ldquoSigfox - the global communications service provider for theinternet of things (iot)rdquo httpsigfoxcom

[30] lora alliance ldquoLoRaWANrdquo httpswwwlora-allianceorg[31] W Yang M Wang J Zhang J Zou M Hua T Xia and X You

ldquoNarrowband wireless access for low-power massive internet of thingsA bandwidth perspectiverdquo IEEE Wireless Communications vol 24no 3 pp 138ndash145 2017

[32] P Ferrari A Flammini M Rizzi E Sisinni and M Gidlund ldquoOnthe evaluation of lorawan virtual channels orthogonality for densedistributed systemsrdquo in IEEE International Workshop on Measurementand Networking (MampN) 2017 pp 1ndash6

[33] M Rizzi P Ferrari A Flammini and E Sisinni ldquoEvaluation of theiot lorawan solution for distributed measurement applicationsrdquo IEEETransactions on Instrumentation and Measurement vol 66 no 12 pp3340ndash3349 Dec 2017

[34] M Rizzi P Ferrari A Flammini E Sisinni and M Gidlund ldquoUsinglora for industrial wireless networksrdquo in IEEE 13th InternationalWorkshop on Factory Communication Systems (WFCS) 2017 pp 1ndash4

[35] A Saifullah M Rahman D Ismail C Lu R Chandra and J LiuldquoSNOW Sensor network over white spacesrdquo in The 14th ACM Confon Embedded Network Sensor Systems (SenSys) 2016 pp 272ndash285

[36] A Saifullah M Rahman D Ismail C Lu J Liu and R ChandraldquoEnabling reliable asynchronous and bidirectional communication insensor networks over white spacesrdquo in The 15th ACM Conference onEmbedded Network Sensor Systems (SenSys) 2017 pp 1ndash14

[37] M Rahman and A Saifullah ldquoIntegrating low-power wide-area net-works in white spacesrdquo in ACMIEEE Conference on Internet-of-Things Design and Implementation (IoTDI) 2018

[38] A Saifullah M Rahman D Ismail C Lu J Liu and R ChandraldquoLow-power wide-area networks over white spacesrdquo ACMIEEE Trans-actions on Networking 2018

[39] Y D Beyene R Jantti O Tirkkonen K Ruttik S Iraji A LarmoT Tirronen and a J Torsner ldquoNb-iot technology overview and experi-ence from cloud-ran implementationrdquo IEEE Wireless Communicationsvol 24 no 3 pp 26ndash32 2017

[40] GSMA ldquo3gpp low power wide area technologiesrdquo October2016 httpswwwgsmacomiotwp-contentuploads2016103GPP-Low-Power-Wide-Area-Technologies-GSMA-White-Paperpdf

[41] u blox ldquoLte cat m1rdquo httpswwwu-bloxcomenlte-cat-m1[42] Bluetooth-SIG ldquoBluetooth core specification version 50rdquo 2016[43] R Rondon M Gidlund and K Landernas ldquoEvaluating bluetooth

low energy suitability for time-critical industrial iot applicationsrdquoInternational Journal of Wireless Information Networks vol 24 no 3pp 278ndash290 Sep 2017

[44] G Patti L Leonardi and L L Bello ldquoA bluetooth low energy real-time protocol for industrial wireless mesh networksrdquo in IECON 2016- 42nd Annual Conference of the IEEE Industrial Electronics SocietyOct 2016 pp 4627ndash4632

[45] M Marinoni A Biondi P Buonocunto G Franchino D Cesarini andG Buttazzo ldquoReal-time analysis and design of a dual protocol supportfor bluetooth le devicesrdquo IEEE Transactions on Industrial Informaticsvol 13 no 1 pp 80ndash91 Feb 2017

[46] A Al-Fuqaha A Khreishah M Guizani A Rayes and M Moham-madi ldquoToward better horizontal integration among iot servicesrdquo IEEECommunications Magazine vol 53 no 9 pp 72ndash79 2015

[47] A Al-Fuqaha M Guizani M Mohammadi M Aledhari andM Ayyash ldquoInternet of things A survey on enabling technologiesprotocols and applicationsrdquo IEEE Communications Surveys Tutorialsvol 17 no 4 pp 2347ndash2376 2015

[48] J P Tomas ldquoThames water rolls out smart meterproject in londonrdquo 2017 httpswiprodigitalcomcasesprogressive-metering-a-utilitys-strategic-move-into-predictive-planning

[49] httpenhartcommorghcptechapplicationsapplications successmitsubishi chemicalhtml

[50] M H Almarshadi and S M Ismail ldquoEffects of precision irrigation onproductivity and water use efficiency of alfalfa under different irrigationmethods in arid climatesrdquo Journal of Applied Sciences Research vol 7no 3 pp 299ndash308 2011

[51] H-J Kim K A Sudduth and J W Hummel ldquoSoil macronutrientsensing for precision agriculturerdquo Journal of Environmental Monitor-ing vol 11 no 10 pp 1810ndash1824 2009

[52] N D Mueller J S Gerber M Johnston D K Ray N Ramankuttyand J A Foley ldquoClosing yield gaps through nutrient and watermanagementrdquo Nature vol 490 no 7419 pp 254ndash257 2012

[53] D Vasisht Z Kapetanovic J Won X Jin R Chandra S SinhaA Kapoor M Sudarshan and S Stratman ldquoFarmbeats An iotplatform for data-driven agriculturerdquo in 14th USENIX Symp on NetSyst Design and Implementation (NSDI) 2017 pp 515ndash529

[54] Microsoft ldquoFarmBeats IoT for agriculturerdquo httpswwwmicrosoftcomen-usresearchprojectfarmbeats-iot-agriculture

[55] C Corporation ldquoData-driven agricultural decisions and insights tomaximize every acrerdquo httpswwwclimatecom

[56] ATampT M2X ldquoAgriculture iot software as a service (saas)rdquo httpsm2xattcomiotindustry-solutionsiot-dataagriculture

[57] J Hawn ldquoAgricultural iot promises to reshapefarmingrdquo RCR Wireless News November 2015httpswwwrcrwirelesscom20151111internet-of-thingsagricultural-internet-of-things-promises-to-reshape-farming-tag15

[58] Schlumberger ldquoSchlumberger robotics servicesrdquo httpwwwslbcomservicesadditionalrobotics-servicesaspx

[59] T Simonite ldquoMining 24 hours a day with robotsrdquo MIT TechnologyReview December 2016 httpswwwtechnologyreviewcoms603170mining-24-hours-a-day-with-robots

[60] T Rault A Bouabdallah and Y Challal ldquoEnergy efficiency in wirelesssensor networks a top-down surveyrdquo vol 67 pp 104ndash122 07 2014

[61] 3GPP ldquoStandardization of NB-IOT completedrdquo June 2016 httpwww3gpporgnews-events3gpp-news1785-nb iot complete

[62] P Ferrari A Flammini E Sisinni D Brando and M Rocha ldquoDelayestimation of industrial iot applications based on messaging protocolsrdquoIEEE Transactions on Instrumentation and Measurement pp 1ndash122018

[63] T Zheng M Gidlund and J Akerberg ldquoWirarb A new mac protocolfor time critical industrial wireless sensor network applicationsrdquo IEEESensors Journal vol 16 no 7 pp 2127ndash2139 April 2016

[64] S Han X Zhu D Chen A K Mok and M Nixon ldquoReliableand real-time communication in industrial wireless mesh networksrdquoin Proceedings of IEEE Real-Time and Embedded Technology andApplications Symposium (RTAS) 2011 pp 3ndash12

[65] Q Leng Y-H Wei S Han A Mok W Zhang and M TomizukaldquoImproving control performance by minimizing jitter in rt-wifi net-worksrdquo in IEEE Real-Time Sys Symp (RTSS) 2014 pp 63ndash73

[66] A Saifullah C Lu Y Xu and Y Chen ldquoReal-time scheduling forWirelessHART networksrdquo in Proceedings of IEEE Real-Time SystemsSymposium (RTSS) 2010 pp 150ndash159

[67] J Song S Han A Mok D Chen M Lucas M Nixon and W PrattldquoWirelesshart Applying wireless technology in real-time industrialprocess controlrdquo in Proceedings of IEEE Real-Time and EmbeddedTechnology and Applications Symposium (RTAS) 2008 pp 377ndash386

[68] Y-H Wei Q Leng S Han A K Mok W Zhang and M TomizukaldquoRT-WiFi Real-time high-speed communication protocol for wirelesscyber-physical control applicationsrdquo in Proceedings of IEEE Real-TimeSystems Symposium (RTSS) 2013 pp 140ndash149

[69] A Saifullah Y Xu C Lu and Y Chen ldquoEnd-to-end communicationdelay analysis in industrial wireless networksrdquo IEEE Transactions onComputers vol 64 no 5 pp 1361ndash1374 2014

[70] A Saifullah D Gunatilaka P Tiwari M Sha C Lu B Li C Wuand Y Chen ldquoSchedulability analysis under graph routing in Wire-lessHART networksrdquo in Proceedings of the IEEE Real-Time SystemsSymposium (RTSS) 2015 pp 165ndash174




[71] A Saifullah S Sankar J Liu C Lu B Priyantha and R ChandraldquoCapNet A real-time wireless management network for data centerpower cappingrdquo in Proceedings of the IEEE Real-Time Systems Sym-posium (RTSS) 2014 pp 334ndash345

[72] O Chipara C Lu and G-C Roman ldquoReal-time query scheduling forwireless sensor networksrdquo IEEE transactions on computers vol 62no 9 pp 1850ndash1865 2013

[73] O Chipara C Wu C Lu and W Griswold ldquoInterference-awarereal-time flow scheduling for wireless sensor networksrdquo in the 23rdEuromicro Conf on Real-Time Sys (ECRTS) 2011 pp 67ndash77

[74] T L Crenshaw S Hoke A Tirumala and M Caccamo ldquoRobustimplicit edf A wireless mac protocol for collaborative real-timesystemsrdquo ACM Trans on Embed Comp Sys (TECS) vol 6 no 4p 28 2007

[75] A Saifullah C Wu P Tiwari Y Xu Y Fu C Lu and Y Chen ldquoNearoptimal rate selection for wireless control systemsrdquo ACM Transactionson Embedded Computing Systems vol 13 no 4s pp 1ndash25 2013

[76] W Shen T Zhang M Gidlund and F Dobslaw ldquoSas-tdma a sourceaware scheduling algorithm for real-time communication in industrialwireless sensor networksrdquo Wireless networks vol 19 no 6 pp 1155ndash1170 2013

[77] F Dobslaw T Zhang and M Gidlund ldquoEnd-to-end reliability-awarescheduling for wireless sensor networksrdquo IEEE Transactions on Indus-trial Informatics vol 12 no 2 pp 758ndash767 2016

[78] G C Buttazzo E Bini and D Buttle ldquoRate-adaptive tasks Modelanalysis and design issuesrdquo in Design Automation amp Test in EuropeConference amp Exhibition (DATE) IEEE 2014 pp 1ndash6

[79] J Kim K Lakshmanan and R Rajkumar ldquoRhythmic tasks A new taskmodel with continually varying periods for cyber-physical systemsrdquo inIEEEACM Third International Conference on Cyber-Physical Systems(ICCPS) 2012 pp 55ndash64

[80] C Lu A Saifullah B Li M Sha H Gonzalez D Gunatilaka C WuL Nie and Y Chen ldquoReal-time wireless sensor-actuator networks forindustrial cyber-physical systemsrdquo Proceedings of the IEEE vol 104no 5 pp 1013ndash1024 2016

[81] A Gupta X Lin and R Srikant ldquoLow-complexity distributed schedul-ing algorithms for wireless networksrdquo IEEEACM Transactions onNetworking (TON) vol 17 no 6 pp 1846ndash1859 2009

[82] X Lin and S B Rasool ldquoConstant-time distributed scheduling poli-cies for ad hoc wireless networksrdquo IEEE Transactions on AutomaticControl vol 54 no 2 pp 231ndash242 2009

[83] N Vaidya A Dugar S Gupta and P Bahl ldquoDistributed fair schedulingin a wireless lanrdquo IEEE Transactions on Mobile Computing vol 4no 6 pp 616ndash629 2005

[84] K S Vijayalayan A Harwood and S Karunasekera ldquoDistributedscheduling schemes for wireless mesh networks A surveyrdquo ACMComputing Surveys (CSUR) vol 46 no 1 p 14 2013

[85] X Wu R Srikant and J R Perkins ldquoScheduling efficiency ofdistributed greedy scheduling algorithms in wireless networksrdquo IEEETransactions on Mobile Computing vol 6 no 6 pp 595ndash605 2007

[86] T Zhang T Gong C Gu H Ji S Han Q Deng and X S HuldquoDistributed dynamic packet scheduling for handling disturbances inreal-time wireless networksrdquo in IEEE Real-Time and Embed Tech andApp Symp (RTAS) 2017 pp 261ndash272

[87] T Zhang T Gong Z Yun S Han Q Deng and X S Hu ldquoFd-pas Afully distributed packet scheduling framework for handling disturbancesin real-time wireless networksrdquo in IEEE Real-Time and Embed Techand App Symp (RTAS) 2018 pp 1ndash12

[88] D Yang Y Xu and M Gidlund ldquoCoexistence of ieee802154 basednetworks A surveyrdquo in Proceedings of the 36th Annual Conference onIEEE Industrial Electronics Society (IECON) 2010 pp 2107ndash2113

[89] mdashmdash ldquoWireless coexistence between ieee 80211- and ieee 802154-based networks A surveyrdquo International Journal of Distributed SensorNetworks vol 7 no 1 p 912152 2011

[90] A Sikora and V F Groza ldquoCoexistence of ieee802154 with other sys-tems in the 24 ghz-ism-bandrdquo in Proceedings of IEEE Instrumentationand Measurement Technology Conference vol 3 2005 pp 1786ndash1791

[91] L L Bello and E Toscano ldquoCoexistence issues of multiple co-locatedieee 802154zigbee networks running on adjacent radio channels inindustrial environmentsrdquo IEEE Transactions on Industrial Informaticsvol 5 no 2 pp 157ndash167 2009

[92] T M Chiwewe C F Mbuya and G P Hancke ldquoUsing cognitiveradio for interference-resistant industrial wireless sensor networks Anoverviewrdquo IEEE Transactions on Industrial Informatics vol 11 no 6pp 1466ndash1481 2015

[93] S Grimaldi A Mahmood and M Gidlund ldquoAn svm-based method forclassification of external interference in industrial wireless sensor and

actuator networksrdquo Journal of Sensor and Actuator Networks vol 6no 2 p 9 2017

[94] F Barac M Gidlund and T Zhang ldquoScrutinizing bit- and symbol-errors of ieee 802154 communication in industrial environmentsrdquoIEEE Transactions on Instrumentation and Measurement vol 63 no 7pp 1783ndash1794 2014

[95] Y H Yitbarek K Yu J Akerberg M Gidlund and M BjorkmanldquoImplementation and evaluation of error control schemes in industrialwireless sensor networksrdquo in 2014 IEEE International Conference onIndustrial Technology (ICIT) 2014 pp 730ndash735

[96] F Barac M Gidlund and T Zhang ldquoUbiquitous yet deceptiveHardware-based channel metrics on interfered wsn linksrdquo IEEE Trans-actions on Vehicular Technology vol 64 no 5 pp 1766ndash1778 2015

[97] F Barac S Caiola M Gidlund E Sisinni and T Zhang ldquoChanneldiagnostics for wireless sensor networks in harsh industrial environ-mentsrdquo IEEE Sensors Journal vol 14 no 11 pp 3983ndash3995 2014

[98] P Agrawal A Ahlen T Olofsson and M Gidlund ldquoLong termchannel characterization for energy efficient transmission in industrialenvironmentsrdquo IEEE Transactions on Communications vol 62 no 8pp 3004ndash3014 2014

[99] T Olofsson A Ahlen and M Gidlund ldquoModeling of the fading statis-tics of wireless sensor network channels in industrial environmentsrdquoIEEE Transactions on Signal Processing vol 64 no 12 pp 3021ndash3034 2016

[100] L Ascorti S Savazzi G Soatti M Nicoli E Sisinni and S Gal-imberti ldquoA wireless cloud network platform for industrial processautomation Critical data publishing and distributed sensingrdquo IEEETransactions on Instrumentation and Measurement vol 66 no 4 pp592ndash603 2017

[101] S M Kim and T He ldquoFreebee Cross-technology communication viafree side-channelrdquo in Proceedings of the 21st Annual InternationalConference on Mobile Computing and Networking ACM 2015 pp317ndash330

[102] T Heer O Garcia-Morchon R Hummen S L Keoh S S Kumarand K Wehrle ldquoSecurity challenges in the ip-based internet of thingsrdquoWireless Personal Communications vol 61 no 3 pp 527ndash542 2011

[103] J Gubbi R Buyya S Marusic and M Palaniswami ldquoInternet ofthings (iot) A vision architectural elements and future directionsrdquoFuture generation comp syst vol 29 no 7 pp 1645ndash1660 2013

[104] P H Cole and D C Ranasinghe Networked rfid Systems ampLightweight Cryptography Springer 2007

[105] H C Pohls V Angelakis S Suppan K Fischer G Oikonomou E ZTragos R D Rodriguez and T Mouroutis ldquoRerum Building a reli-able iot upon privacy-and security-enabled smart objectsrdquo in WirelessCommunications and Networking Conference Workshops (WCNCW)IEEE 2014 pp 122ndash127

[106] A-R Sadeghi C Wachsmann and M Waidner ldquoSecurity and privacychallenges in industrial internet of thingsrdquo in Proceedings of the 52ndannual design automation conference ACM 2015 p 54

[107] A W Atamli and A Martin ldquoThreat-based security analysis for theinternet of thingsrdquo in International Workshop on Secure Internet ofThings (SIoT) IEEE 2014 pp 35ndash43

[108] Z-K Zhang M C Y Cho C-W Wang C-W Hsu C-K Chen andS Shieh ldquoIot security ongoing challenges and research opportunitiesrdquoin IEEE 7th International Conference on Service-Oriented Computingand Applications (SOCA) 2014 pp 230ndash234

[109] R Arends R Austein M Larson D Massey and S Rose ldquoDnssecurity introduction and requirementsrdquo Tech Rep 2005

[110] G Baldini T Peirce M Botterman et al ldquoIot governance privacyand security issuesrdquo Position paper European Research Cluster on theInternet of Things 2015

[111] S Raza ldquoLightweight security solutions for the internet of thingsrdquoPhD dissertation Malardalen University Vasteras Sweden 2013

[112] J H Ziegeldorf O G Morchon and K Wehrle ldquoPrivacy in theinternet of things threats and challengesrdquo Security and CommunicationNetworks vol 7 no 12 pp 2728ndash2742 2014




definitions exist each one trying to capture one of its funda-mental characteristics It is often considered as a sort of webfor the machines highlighting the aim of allowing things toexchange data However application fields are so diverse thatsome requirements (especially those related to communicationaspects) can be very different depending on the intendedgoals and end-users the underlying business models and theadopted technological solutions What is usually addressed asIoT could be better named as consumer IoT as opposed toindustrial IoT [6] [7]

Consumer IoT is human-centered the ldquothingsrdquo are smartconsumer electronic devices interconnected with each otherin order to improve human awareness of the surroundingenvironment saving time and money In general consumerIoT communications can be classified as machine-to-user andin the form of client-server interactions

On the other hand in the industrial world we are assisting tothe advent of the digital and smart manufacturing which aimat integrating Operational Technology (OT) with InformationTechnology (IT) domains [8] In very few words the IIoT (thebasic pillar of digital manufacturing) is about connecting allthe industrial assets including machines and control systemswith the information systems and the business processes Asa consequence the large amount of data collected can feedanalytics solutions and lead to optimal industrial operationsOn the other hand smart manufacturing obviously focuses onthe manufacturing stage of (smart) products life-cycle with thegoal of quickly and dynamically respond to demand changesTherefore the IIoT affects all the industrial value chain andis a requirement for smart manufacturing

As underlined in the following communication in IIoT ismachine oriented and can range across a large variety ofdifferent market sectors and activities The IIoT scenariosinclude legacy monitoring applications (eg process moni-toring in production plants) and innovative approaches forself-organizing systems (eg autonomic industrial plant thatrequires little if any human intervention) [9]

While the most general communication requirements of IoTand IIoT are similar eg support for the Internet ecosystemusing low-cost resource-constrained devices and network scal-ability many communication requirements are specific to eachdomain and can be very different eg Quality of Service(QoS) (in terms of determinism latency throughput etc) theavailability and reliability and the security and privacy IoTfocuses more on the design of new communication standardswhich can connect novel devices into the Internet ecosystemin a flexible and user-friendly way By contrast the currentdesign of IIoT emphasizes on possible integration and inter-connection of once isolated plants and working islands or evenmachineries thus offering a more efficient production and newservices [9] For this reason compared with IoT IIoT can beconsidered more an evolution rather than a revolution TableI gives a qualitative comparison of these technologies

Regarding the connectivity and criticality IoT is more flex-ible allowing ad hoc and mobile network structures and hav-ing less stringent timing and reliability requirements (exceptfor medical applications) On the other hand IIoT typicallyemploys fixed and infrastructure-based network solutions that

TABLE ICOMPARISON BETWEEN CONSUMER IOT AND INDUSTRIAL IOT

Consumer IoT Industrial IoT

Impact Revolution Evolution

Service Model Human-centered Machine-oriented

Current Status New devices and stan-dards

Existing devicesand standards

Connectivity Ad-Hoc (infrastructureis not tolerated nodescan be mobile)

Structured (nodesare fixed central-ized network man-agement)

Criticality Not stringent (exclud-ing medical applica-tions)

Mission critical(timing reliabilitysecurity privacy)

Data Volume Medium to High High to Very High

are well designed to match communication and coexistenceneeds In IIoT communications are in the form of machine-to-machine links that have to satisfy stringent requirements interms of timeliness and reliability Taking process automationas an example domain where process monitoring and controlapplications can be grouped into three sub-categories moni-toringsupervision closed loop control and interlocking andcontrol While monitoring and supervision applications are lesssensitive to packet loss and jitter and can tolerate transmissiondelay at second level closed loop control and interlocking andcontrol applications require bounded delay at millisecond level(10-100ms) and a transmission reliability of 9999 [5]

Comparing the data volume the generated data from IoT isheavily application dependent while IIoT currently targets atanalytics eg for predictive maintenance and improved logis-tics This implies that very large amount of data are exchangedin IIoT For example it is reported that the Rio Tinto minegenerates up to 24TB of data per minute according to CiscoGlobal Cloud Index

The concept of Industry 40 (where 40 represents the fourthindustrial revolution) arises when the IoT paradigm is mergedwith the Cyber-Physical Systems (CPSs) idea [10] Originallydefined in Germany the Industry 40 concept has gained aglobal visibility and it is nowadays universally adopted foraddressing the use of Internet technologies to improve produc-tion efficiency by means of smart services in smart factoriesCPSs extend real-world physical objects by interconnectingthem altogether and providing their digital descriptions Suchinformation stored in models and data objects that can beupdated in real time represents a second identity of the objectitself and constitutes a sort of ldquodigital twinrdquo Thanks to thedynamic nature of these digital twins innovative services thatwere not possible in the past can be implemented acrossthe whole product lifecycle from inception to disposal ofmanufactured products In summary IIoT is a subset of IoTwhich is specific to industrial applications The manufacturingphase of the product lifecycle is where the IoT and Industry40 meet originating to the IIoT Figure 1 shows intersectionsof IoT CPS IIoT and Industry 40

As a concluding remark it has to be highlighted thatthe IIoT paradigm is not intended for substituting traditionalautomation applications but aims at increasing the knowledge




IIoTIoT CPS

Ind

ustr

y 4

0













ENTERPRISE busines



archive UIs)

EDGE monitoring

control safety

domains

PLATFORM provides a


bus (performs data

collection and

transformation)

IIoT Edge

Gateways

Proximity

Networks

TIERS


Controllers

Networks

Service





NETWORKING

CONNECTIVITY

INFORMATION

Number of

Protocols

Protocol age

and mutability


































A Energy Efficiency




































V CONCLUSION



REFERENCES


























































































































IIoTIoT CPS

Ind

ustr

y 4

0













ENTERPRISE busines



archive UIs)

EDGE monitoring

control safety

domains

PLATFORM provides a


bus (performs data

collection and

transformation)

IIoT Edge

Gateways

Proximity

Networks

TIERS


Controllers

Networks

Service





NETWORKING

CONNECTIVITY

INFORMATION

Number of

Protocols

Protocol age

and mutability


































A Energy Efficiency




































V CONCLUSION



REFERENCES


























































































































ENTERPRISE busines



archive UIs)

EDGE monitoring

control safety

domains

PLATFORM provides a


bus (performs data

collection and

transformation)

IIoT Edge

Gateways

Proximity

Networks

TIERS


Controllers

Networks

Service





NETWORKING

CONNECTIVITY

INFORMATION

Number of

Protocols

Protocol age

and mutability


































A Energy Efficiency




































V CONCLUSION



REFERENCES




















































































































































A Energy Efficiency




































V CONCLUSION



REFERENCES















































































































































V CONCLUSION



REFERENCES























































































































Industrial Internet of Things: Challenges, Opportunities...

Documents

Transcript of Industrial Internet of Things: Challenges, Opportunities...