An information architecture to enable track-and-trace capability in Physical Internet ports

Delft University of Technology

An information architecture to enable track-and-trace capability in Physical Internet ports

Fahim, Patrick B.M.; An, Rowoon; Rezaei, Jafar; Pang, Yusong; Montreuil, Benoit; Tavasszy, Lorant

DOI

10.1016/j.compind.2021.103443

Publication date

2021

Document Version

Final published version

Published in

Computers in Industry

Citation (APA)

Fahim, P. B. M., An, R., Rezaei, J., Pang, Y., Montreuil, B., & Tavasszy, L. (2021). An information

architecture to enable track-and-trace capability in Physical Internet ports. Computers in Industry, 129,

[103443]. https://doi.org/10.1016/j.compind.2021.103443

Important note

To cite this publication, please use the final published version (if applicable).

Please check the document version above.

Copyright

Other than for strictly personal use, it is not permitted to download, forward or distribute the text or part of it, without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license such as Creative Commons. Takedown policy

Please contact us and provide details if you believe this document breaches copyrights. We will remove access to the work immediately and investigate your claim.

This work is downloaded from Delft University of Technology.

ContentslistsavailableatScienceDirect

Computers

in

Industry

j ou rn a l h om ep a g e :w w w . e l s e v i e r . c o m / l o c a t e / c o m p i n d

An

information

architecture

to

enable

track-and-trace

capability

in

Physical

Internet

ports

Patrick

B.M.

Fahim

_,

_Rowoon

_An

_,

_Jafar

_Rezaei

_,

_Yusong

_Pang

_,

_Benoit

_Montreuil

_,

Lorant

Tavasszy

a_{TransportandLogisticsGroup,FacultyofTechnology,PolicyandManagement,DelftUniversityofTechnology,2628BX,Delft,TheNetherlands}

b_{TransportEngineeringandLogisticsGroup,FacultyofMechanical,MaritimeandMaterialsEngineering,DelftUniversityofTechnology,2628BX,Delft,The}

Netherlands

c_{PhysicalInternetCenter,H.MiltonStewartSchoolofIndustrialandSystemsEngineering,GeorgiaInstituteofTechnology,Atlanta,30332,GA,USA}

a

r

t

i

c

l

e

i

n

f

o

Articlehistory:

Received17March2020

Receivedinrevisedform25January2021

Accepted10March2021 Keywords: PhysicalInternet Logistics Track-and-trace Industry4.0 Informationarchitecture Interoperability

a

b

s

t

r

a

c

t

ThePhysicalInternet(PI),anewvisionforthefutureoftheglobalfreighttransportandlogistics sys-tem,describesageographicalhierarchyofinterconnectednetworksofnetworks,fromtheurban,tothe national,thecontinental,andthegloballevel.Liketoday,inPIthemaritimeportswillfulfilrolesas continentalandglobalhubs.Differentlythanportstoday,however,decisionstosplitandbundlecargo acrossshipsandothermodeswillnotbemadesolelyonthebasisoflong-termagreementsbyports,but ratherevermoredynamicallyandinreal-time,aimingtoreconsolidateshipmentswithintheportarea. Thisimpliesaneedtoreconsiderthecurrentlyusedinformationsystems(ISs),andtogain understand-ingoffuturerequirementstosatisfytheirneeds.Weexploitadesignscienceresearch(DSR)approach toshapetheserequirements.AmongthemanycomponentsoffutureISs,westudyports’ track-and-trace(T&T)capability.Theproposedinformationarchitecture(IA)enablestointegrateT&Tcapabilityin PIportsbymeansofinformationcarriedonPIcontainersintothelogisticschainviaanopeninterface platform,whichalsosupportsinteroperabilityamongthevariousactors’ISs.Thedesignisbasedonthe ReferenceArchitectureModelforIndustry4.0(RAMI4.0).ThismodelsupportstheanalysisofPIportsin keydimensionsalongwithhierarchicallogisticsentities,whichcouldbeusedasablueprintforIAsofPI ports,globally.Weprovideinsightsintotheapproach’sapplicabilitybymeansoftheillustrativecaseof Teesport,locatedinNortheastEngland(UnitedKingdom).

©2021TheAuthors.PublishedbyElsevierB.V.ThisisanopenaccessarticleundertheCCBYlicense (http://creativecommons.org/licenses/by/4.0/).

1. Introduction

Throughoutthepastcenturies,thefacilitationofinternational tradehasmadesignificantcontributionstothecurrentlevelof glob-alization,aswellastoglobalwelfareandeconomy.Currentglobal maritime tradevolumessurpass10 billiontonsannually,while 80%ofthetotalworldmerchandisetradeistransportedoversea

(Hoffmannetal.,2018).Beingthegatewaybetweenlandandsea,

maritimeportsfunctionascriticalenablersofinternationaltrade andglobalsupplychains.Portscanberegardedasdynamicand organicsystemsin nationalsocio-economic-politicalsystems as wellasintheglobalizedeconomicsystem(Haraldsonetal.,2020).

∗ Correspondingauthor.

E-mailaddresses:p.b.m.fahim@tudelft.nl(P.B.M.Fahim),r.an@tudelft.nl(R.An),

j.rezaei@tudelft.nl(J.Rezaei),y.pang@tudelft.nl(Y.Pang),

benoit.montreuil@isye.gatech.edu(B.Montreuil),l.a.tavasszy@tudelft.nl

(L.Tavasszy).

Therefore,portscontinuouslyneedtoevolvebyadaptingtotheir externalenvironment in termsof changingeconomic and trad-ingpatterns,newtechnologies,legislation,and portgovernance systems.

A system innovation that is already impacting the current economicandtradingpatterns,technologies,legislation,and gov-ernancesystems,isthePhysicalInternet(PI).In2011,Montreuil

(2011)introducedthevision ofthePIasoneofanopenglobal

freightlogisticssystemfoundedonphysical,digitalandoperational hyperconnectivitythrough encapsulation,interfaces, and proto-cols.ThePIproposesphysicalpackagestobemovedsimilarlyto thewaydatapacketsmoveintheDigitalInternet(Panetal.,2017). InthePI,goodsareencapsulatedinmodularlydimensioned easy-to-interlockintelligentcontainers,calledPIcontainers,whichare designedtooptimallyflowinhyperconnectedlogisticsnetworks

(Sallezetal.,2016).ThePIisexpectedtostrengthentheeconomic,

environmental,andsocietalsustainabilityandefficiencyofglobal logistics(Montreuiletal.,2012).

https://doi.org/10.1016/j.compind.2021.103443

Tohelpachievehyperconnectivityintheglobalfreight logis-tics system,ports needto becapable ofautonomously routing shipmentsofPIcontainers,basedonappropriatereal-time infor-mationavailability.FuturePIapplicationswillbedataintensive andwillrequirestrongsensing,communication,dataprocessing, anddecision-makingcapabilities.Inthedesignofintelligent sys-tems, sensing(information handling),which isthefocus ofour study, comes prior to thinking (problem notification),and act-ing(decision-making)(Meyeretal.,2009).InPIapplications,we considersensing astheprocessofachievingincreasedvisibility bymeansofenhancedtrack-and-trace(T&T)systems,supported byinformationarchitectures(IAs)thatallowforcommunication among the various internal and external logistics entities and actors.Aprimarymeanstocreatevisibilityofshipmentsforthe completelogisticschainistheT&Tcapabilityinports(McFarlane etal.,2016).PIportswillneedtobeabletoprocessinformation onanindividualshipmentleveltofacilitateoptimal(un)loading andde-and(re-)compositioningoperationsofPIcontainers.This impliesthatdataabouttheshipmentswithincontainerswillneed tobeaccessible.Inaddition,Calatayudetal.(2019)emphasizethe importanceofT&Tsystemsforpredictivedecision-making capabil-itiesofsupplychains.WearguethatinthePI,thisimportancewill growfurtherandrequireaccesstomoredetailedinformation.Inthe PIcontext,T&Tisthereal-timeabilitytolocateeveryindividualPI containerwithitscontentsandtoprovidetraceabilityinformation (e.g.weight,state,commoditytype,estimatedarrivaland depar-turetimes,originanddestination,andenvironmentalconditions) torelevantactors(Sallezetal.,2016).Today,however,port infor-mationsystems(ISs)onlysupportT&Tatcontainerlevel,typically 20and40footcontainers,andnotatthelevelofunderlying ship-mentunits.Ifportswantkeepanessentialexistenceinthefuture door-to-doorPIsystem,theyshouldadapttotheneedsofthePI andextendthecapabilitiesoftheT&Tsystems.Untilnow,there hasbeennoattentionintheliteratureonthisproblem.

Tohelpfillingthisgapinliterature,ourresearchquestionisthe following:

Whatistheproperarrangementofinformationflowsonshipments andtheircharacteristics,thatsupportsT&Tofgoodsinsideaport, withinthePIcontext?

Inordertoanswerthisresearchquestion,weuseadesign sci-enceresearch(DSR)approach(Weber,2018),bytheguidelinesof whichwedevelopafunctionaldesignofanISthatprovidesthe port withtherequiredT&Tcapabilities(i.e.includingshipment levelinformation).Thetaskofre-designingports’ISstosuitanew functionality isnottrivial.WithinanIS,thedifferentaspectsof information sharing,including dataelements, messageformats, communicationlines,shouldbedefinedinlinewiththenew busi-nessobjectives,andinaconsistentrelationtoeachother(Romero

andVernadat,2016).Inthisstudy,wedevelopsuchadesign.

There-fore,ourmaincontributionisthetractableandreproducibledesign ofanIAfortheT&Tfunctionalityofmaritimeportsina PI con-text.Thedesignof asharedinformation environmentthatlives up totheseconditions is calledan IA(Yaqoobet al.,2017).To keepthedifferentaspectsoftheinformationtractable,consistent and complete,weusea referencearchitecturemodel(RAM)for theIAdesign, whichprovidesguidancerelativetothedifferent elementsthatneedtobeincluded.ARAMcanbedefinedasan abstractsystemframeworkthatcontainsaminimalsetofunifying concepts,axioms,andrelationshipstounderstandtheinteractions betweenentitiesin andwithitsenvironment(VanGeestet al., 2021).WeusetheReferenceArchitectureModelforIndustry4.0 (RAMI4.0),awell-knownreferencemodelusedworldwideforIA designs(Bangemannetal.,2016).Assuch,ourmainresearch con-tributionisthetractableandreproducibledesignofanIAforthe T&TfunctionalityofmaritimeportsinaPIcontext.

Therestofthepaperisbuiltupasfollows.Anoverviewofthe relevantport,PI,and IAliteratureis providedinSection2. Sec-tion3introducesthemethodology.Section4presentsareal-world case,whichisfollowedbyconceptualdesignoftheIAinSection

5.Section6providesadiscussion, whileSection7presentsthe conclusionsofourworkandrecommendationsforfutureresearch.

2. Literaturereview

T&Thasbeenrecognizedasanimportantelementwithinsupply chainmanagementingeneral,andportsinspecific.Onestreamof literatureaddressesthisfromadescriptiveportevolution perspec-tive;anotherfromanormativedesignapproachfocusingonthe globalPIasanultimatevision.Inaddition,thesetwostreamsof literature,wereviewtheliteratureofinnovativeRAMsandIAsand theirapplications,whichalsoincludeInternet-of-Things(IoT)and blockchainapplication,designedfortheIndustry4.0movement. Weconcludethissectionbyidentifyingaconvergingresearchgap asthestartingpointforourwork.

2.1. Maritimeportevolutionanddevelopments

Inthemaritimeportlogisticsliterature,theevolutionarypathof portshasbeendescribedthroughseveralgenerations(LeeandLam, 2016).Ports,overtime,haveevolvedfromfirstgenerationports (1GPs),whichmerelyservedasgatewaysbetweenlandandsea, andarenowmovingintofifthgenerationports(5GPs),whichare consideredhighlycomplexanddynamicmulti-actorsystemswith advanced(information)technologiesandawiderangeof (value-added)services,inadditiontothetraditionalones.LeeandLam

(2016)emphasize thekey rolesofnew informationtechnology

(IT)in themostmodern5GPs, notably contrastingtheirIT fea-turesversusthoseoffourth-generationports(4GPs).Essentially, ITin4GPsfocusesonprovidingcargoclearanceandT&Tservices oncontainerlevel, whereasITin 5GPsgoesonestepfurtherby offeringitsusers asinglewindow(SW)bymeansofPort Com-munitySystems(PCSs)forinformationexchangeaboutT&Tofnot onlymaritimecontainersbutalsoitscontents(onashipmentlevel), deliveryinformation,andperformancemeasurement(LeeandLam, 2016).Anothermorerecentlydevelopedconceptthatexplains cur-rentandfuturepractices,andiscloselylinkedwithPCSs,isPort CollaborativeDecision-Making(PortCDM).Bymakingtheforeland operationsaspredictableandreal-timeaspossible,PortCDMmakes notonlyprocessesinoneportmoreefficient,butwillalso con-tributetoanincreaseintheefficienciesofotherportsandvessels (Lindetal.,2020).

AdistinctioncanbemadebetweeninternalT&Tsystemsinside aparticular(local)logisticssystem,suchasaport,andexternalT&T systemsacrossthesupplychain.In5GPs,PCSsfulfilthefunctionof, amongothers,T&Tacrossthesupplychain(EPCSA,2011a).APCS canbedefinedasaneutralandopenelectronicplatform,enabling intelligentandsecureexchangeofinformationbetweenpublicand privateactorstoimprovethecompetitivenessofportcommunities

(EPCSA, 2011b).PCSs aim to contribute to optimizing,

manag-ing,andautomatingportandlogisticsprocessesthroughasingle submissionofdataandconnectingsupplychains(IPCSA,2018). Globally,variousPCSswitharangeoffunctionalitieshaveemerged overtheyears(e.g.DakosyinGermany,LoginkinChina,Maqtain UnitedArabEmirates,PortbaseintheNetherlands).Inaddition, initiativesarebeingtakentoexpandtheknowledgecapacityand enhanceusabilityofthesesystemsamongitsactors,oftenledbythe EuropeanandInternationalPCSAssociations(EPCSAandIPCSA), andUnitedNations.InlinewiththeobjectiveofthePIbecomingan openglobalfreighttransportandlogisticssystemthroughphysical, digitalandoperationalhyperconnectivity(Montreuil,2011),future

PCSsaimtosupportT&Tcapabilitiesandinteroperabilityacross supplychains(UNESCAP,2018).However,thePIhasnotbeen con-sideredinthePCSliteraturewhatsoever.Therequirementsofthe PIconcerningT&Tcapabilitiesofaportshouldbeknowntobeable todevelopPCSsinlinewiththe5GPvision.

2.2. PhysicalInternet(PI)

Montreuil(2011)definedthevisionofthePIasanopenlogistics

systemthatiscapableofbeingaccessedbyallactorsinalogistics chainataglobalscale.Montreuiletal.(2012)suggestaframework ofPIfoundationsrepresentingthePI’sbuildingblocksandtheir systematicrelationships,organizedinlayers,including commodi-ties,shipments,loadunits,carriers,andinfrastructurenetworks. AtthecoreofthePIarethefundamentalgoalsofimproving eco-nomic, environmental,andsocietalefficiency and sustainability

(Ballotetal.,2014).Toachievethesegoals,hyperconnectivityat

thephysical,digital,operational,transactional,legal,andpersonal levelsisaprerequisite(Montreuiletal.,2016).This hyperconnec-tivityisenabledbythreekeyPIfeatures:encapsulation,interfaces, andprotocols(Montreuiletal.,2013).

2.2.1. Encapsulation

ThePIencapsulatesfreight intomodular(PI)containersthat are easy to handle, store and transport, smart and connected, and eco-friendly(Montreuil, 2011).Montreuil etal.(2016) pro-poseathree-layertypologyofPIcontainers:packagingcontainers (P-containers), handling containers (H-containers), and transport containers(T-containers).P-containersdirectlyencloseandprotect thephysicalobjectsintheinnermostcomposition. P-containers can beembeddedinH-containers designed for usein handling and operationswithinthePI.H-containerscanbeembeddedin T-containers,whicharefunctionallysimilartothemaritime ship-pingcontainersthatarecurrentlyused,exploitableacrossmultiple modesoftransportation.

2.2.2. Interfaces

Inordertoprovidetransportandlogisticsservices,thePI sys-temneedstoconsiderbothphysical(operational)interfacesaswell asinformationandcommunication(I&C)interfaces,asemphasized

inMontreuiletal.(2012)andsynthesizedinTable1.The

interac-tionsandtheexchangingdatasourcesbetweenthetwointerfaces providethenewcontextforincreasingthevisibilityintransport chains.Whilethehigh-levelinterfacesfocusonlogisticsservices, thelow-levelinterfacesfocusonthePIcontainersatwhichthe informationiscarried.

2.2.3. Protocols

ThePIenablestheinterconnectedexploitationoflogistics net-worksthroughcooperativeprotocolsagreeduponandexploited bythevarietyofactorsinthelogisticschains.PIprotocolsnotonly ensuretheintegrationoflogisticsentitiesbutalsotheir perfor-mance,resilience,andreliabilityinPInetworks(Montreuil,2011). StandardizedPIroutingprotocolswillfacilitatedynamicroutingof PIcontainersacrossmultiplemodesoftransportinthePInetwork. Toconnectlogisticsnetworksandservicesbymeansofprotocolsin

thePI,Montreuiletal.(2012)proposedtheOpenLogistics

Intercon-nection(OLI)modelasthePI’sequivalenttotheDigitalInternet’s OpenSystemsInterconnection(OSI)model.Fig.1depictstheOLI modelwithitssevenlayersandrespectiveprotocols.Thelayered protocolsoftheOLImodelprovideaframeworkforexploiting phys-ical,digital,financial,human,andorganizationalmeansofthePI

(Ballotetal.,2014).Oneachlayer,aninstanceprovidesservices

toaninstanceonahigherlayer,whilereceivingservicesfroman instanceonalowerlayer.Simultaneously,instancesonthesame layercanalsoprovideandreceiveservicestoandfromeachother.

Notethat,fromtheOLIperspective,aT&Tfunctionalitywithina portwillprimarilyconducttheoperationswithinL1,L2,andL3, whilesupportingroutingandshipmentdecisionsatL4andL5.A port,asahub,allowsforroutingdecisions,therearrangementof productsbymeansofPIcontainers,andtheirassignmenttoservice classes.InlinewiththeOLI,thetobedesignedIAconsidershow dataistransmittedbetweendifferentlayers.

FromaPIcontainerperspective,Sallezetal.(2016)exhibited itsroleinhyperconnectedPInetworks.Theyidentifiedfour cate-goriestoclassifyPIcontainerusersinalogisticschain.Asimplified logisticschainofaPIcontainerincludestheseusers:shippersand receivers, PI transport service providers, PI hubs, PI coordinators. Followingthiscategorization,maritimeportscanclearlybe cat-egorizedintothePIhubcategory,whereas,basedontheearlier provideddefinitionanddescriptionofPCSs,thesecouldbeastrong candidatefortheroleofaPIcoordinator.Furthermore,Sallezetal.

(2016)listedidentification,T&T,statemonitoring,datacompatibility

andinteroperability,andconfidentialityasinformationalaspectsof PIcontainers.Smartcontainershaveanembeddedsetofsensors, enablingittocommunicaterealtimeinformationwithitsusers onlocation, dooropening and closing, vibrations,temperature, humidity,andanymeasuredphysicalparametersofthe surround-ingenvironment(Bechaetal.,2020).Althoughourprimaryfocus isonT&Tsystemsinsidetheport,theotherinformationalaspects areimportanttoconsideraswell.TheModulushcaprojectwasthe firstprojectonaEuropeanlevelthatendeavoredtocontributeto therealizationofthePIbyfocusingonthedevelopmentofaset ofexchangeablemodularlogisticsunits,i.e.PIcontainers,inthe fast-movingconsumergoodsindustry(ModulushcaProject,2017). FromtheperspectiveofPIhubs,Ballotetal.(2012)andMeller

etal.(2012)proposefunctionaldesignsofaroad-railhubanda

road-basedtransithub,respectively. Ballotetal.(2014)present somegenericdesignsofuni-andmultimodalhubs,androadand rail hubs, while Sallez et al. (2015) proposeda hybrid control architecturefortherouting ofPI containersin road-rail (cross-docking)PIhubs.Walha etal.(2016)investigated anallocation problemin thecontextof thePIwiththeobjectivetoimprove rail-roadPIhubefficiencybyoptimizingthetravelleddistances. Summarizing,Montreuiletal.(2018)morerecentlyarguedthat exploitinghyperconnectivityandmodularityprovidesseven fun-damentaltransformationstoparcellogisticshubdesign:(1)hubs are toreceive and ship modular containers encapsulating par-celconsolidatedbynextjointdestination;(2)hubsaretoexploit pre-consolidation;(3)hubsaretohavelessdirectsourcesand des-tinationsasthecurrent;(4)hubsaretobeevermoremulti-actor andmulti-modalserviceproviders;(5)hubsaretobemoreagile throughreal-timedynamicandresponsiveshippingtimes;(6)hubs aretobecapableofconductingsmart,real-timedynamicdecisions onthecontainerconsolidation and internal flow orchestration; and(7)hubsaretobeactiveagentsinthePInetwork, dynam-icallyexchangingreal-timeinformationonthestatusofparcels, containers,vehicles,routes,andtheotherhubs.

ForamorecomprehensivereviewofthePIliterature,werefer

toTreiblmaieretal.(2020).

2.3. InformationArchitectures(IA)

Morerecently,withtheintroductionofIoTandblockchainas enablersforawiderangeofapplicationsinlogisticsandsupply chainmanagement (GalatiandBigliardi,2019),variousIAshave beenproposed withspecific applications(Yaqoobet al., 2017).

Bisognoetal.(2015)createdanintegratedinformationflowsmodel

forPCSstoimproveintelligentlogisticsservicesbymeansof adopt-ingacasestudyapproach,investigatingthePortofSalerno.Lietal.

(2016)arguedthatincurrentlogistics,thereisalackofdevices

infor-Table1

TypesandLevelsofInterfaces.

Typeofinterface Levelofinterface Interface

Physical(Operational)Interfaces Low ComplementaryphysicalfixturesthatallowPIcontainersto

interlockwithoneanother,andtobesnappedtostorage structure.

High LogisticsPI-nodesthatareavailableforsmoothlogistics

services(e.g.transferfromunimodaltomultimodal transportation)byappropriatelyallocatingfreightwithinthe PInetwork.

Information&Communication(I&C)Interfaces Low SmarttagsonPIcontainerscapableofidentification,routing, traceability,conditioningofeachmodularcontainer.

High Digitalmiddlewareplatformsthatprovideanopenmarketfor

logisticsservicesinPIbyconnectinghumanandthePI’s components.

Fig.1. Theseven-layerOLImodelwithrespectiveinter-layerservicedescription(adoptedfrom:Montreuiletal.,2012).

mationofcontainerfreight.Hence,theyconstructedaT&Tdevice architecturebasedonIoTtechnologyincombinationwitha multi-sensordevicetoprovidereal-timein-transitvisibility.Tian(2016)

studiedtheutilizationofradiofrequencyidentification(RFID)and blockchaintechnologyinbuildinganagri-foodsupplychain trace-abilitysystem.Theydevelopedasystemthatrealizestraceability withtrustedinformation,whichwouldeffectivelyguaranteethe foodsafetybygathering,transferring,andsharingdatain produc-tion,processing,warehousinganddistribution.Raapetal.(2016)

proposedanarchitectureforanintegrationplatformthatsupports theautomatedcollectionofreal-timecontainertrackingdatafor thepurposeofmoreefficientplanningbylogisticsserviceproviders (LSPs).Byunetal.(2017)developedasystemarchitecturethat con-tributestotheirgraph-orientedpersistenceapproachtoachieve efficientand privacy-enhancedobjecttraceabilitybasedon uni-fiedandlinkedelectronicproductcodeinformationservices.Betti

etal.(2019)andHasanetal.(2020)bothfocusedonexploiting

blockchainwithinaPIcontext.WhileBettietal.(2019)proposed smartcontractstoimprovePItrustabilityandcybersecurity,Hasan

et al. (2020) presented two permissioned blockchain

architec-turesthat providedecentralization,privacy,trust, immutability, andtransparencyinPInetworks.Alsointhefoodsupplychainarea,

Mondaletal.(2019)proposedablockchaininspiredIoT

architec-tureforthepurposeofenhancingtransparency.Thearchitecture wasbasedontheintegrationofRFID-basedsensorataphysical layer,whileapplyingblockchaintechnologyatthecyberlayer.Van

Geestetal.(2021)presentedagenericbusinessprocessmodelfor

smartwarehouses,whilesimultaneouslymodellingitsreference architecture.

TheIS literaturehasrecently evolvedin terms of providing RAMsforinnovativeIAdesigns.SimilartothePI,Industry4.0has thepotentialtoimpactentireindustriesbytransformingtheway goodsaredesigned,manufactured,delivered,andpaid(Hofmann

theproductionandlogisticsdomainandtheuseofweb-based ser-vicesinindustrialprocesses(GalatiandBigliardi,2019).Lasietal.

(2014)and Boyeset al.(2018)arguesimilarly thatIndustry4.0

demandsarchitectureswhichsupportitsimplementationin dif-ferentareas,fromthedesignofproductstothedistributionwith theparticipationofactorsconnectedbyacollaborativenetworkin adistributedenvironment.WeyrichandEbert(2015)proposefive RAMsthataresuitableforIoTapplications:RAMI4.0;Industrial InternetReferenceArchitecture(IIRA);IoT-Architecture;Standard for anArchitecturalFrameworkfor IoT;and Arrowhead Frame-work.AlthougheachoftheRAMshasitsmerits,RAMI4.0provides theextendedabilitytofocusonmultiplesystemlayers,while con-sideringhierarchicallevels,lifecyclesandvaluestreams(Pisching etal.,2018).Inaddition,RAMI4.0allowsforthedescriptionand implementationofhighlyflexibleconceptsinastandardizedway, whereas otherRAMs have a strongfocus on specificuse cases

(Adolphsetal.,2015).Inessence,RAMI4.0providesa“basic

ref-erencearchitecture”forIndustry4.0(Bangemannetal.,2016),and hence,manymajorcompaniesandinstitutionsinvariousindustries

useRAMI4.0(WeyrichandEbert,2015).

2.4. Literaturegapsandcontribution

Theliteratureonmaritimeportsandcargohubsisstartingto recognizetheimportanceandcomplexityoftheexchangeofdata acrossactorstoservetheusersoftheport(Watsonetal.,2020). Additionally,IThasbeenrecognizedasanenablerforportusers tosecurelyexchangedataandprovidevisibilitytothebenefitof theactorsandoperationsthroughoutthelogisticschains.Although designsofISsareemergingtoservenewneedsinports,suchas forsynchronizationofcontainers’movementsacrossmodes(Raap et al.,2016), we observethat therestill is a generalpaucity of ISresearchandliteratureonthe(maritime)shippingindustry.In addition,althoughresearchwithinthePIhasbeenmovingtowards design-orientedwork,currentworksarenotablyonthephysical layoutandactivitiesofPIhubsandtoamuchlesserextentontheir IA,wheremorerecentresearchofBettietal.(2019)andHasanetal.

(2020)canbecountedasexceptions.Theydonotdesignforthe

T&Tfunctionalityexplicitly,however.WeconcludethatanIAfor maritimePIports,withafocusonT&Ttosupportglobal hyper-connectivityata PIcontainerlevel,isstill lacking.Bymeansof designingatractableandreproducibleIAfortheT&Tfunctionality ofmaritimeportsinaPIcontextinthispaper,weaimtocontribute withafirststeptowardsasolutiontothisproblemandfillingthe aforementionedgapsinliterature.

Inthenextsection,weintroduceourmainapproach.InaDSR context,wedesignanIAfortheT&TfunctionofPIports.Theuseofa RAMI4.0allowsustoconsiderseverallayersandhierarchicallevels inISdesign,fromassetsprovidingthedatatothefunctionallevelof informationexchangebetweenactors.Weuseanillustrativecaseof areal-worldlogisticschaintoshowthepracticabilityofthedesign approach,notablyinderivingrequirements.

3. Methodology

The design of aninnovativePI T&TIA preliminarilyaims to achieve appropriatePIcontainerinformationaccessibility, qual-ityandusefulnessthroughopeninterfacesandglobalprotocols. PortISsneedtoprocessT&TinformationonPIcontainerlevelto facilitateeffective,dynamicandreal-time(un)loading,de-and (re-)compositioningofcontainersatports.Usingdesignasresearch activityimpliesaDSRapproach,incontrasttotheclassicalresearch approachfocusingontheorydevelopmentandtesting.

ThefocusofthedesignproblemissummarizedinFig.2.ThePI hasawellelaboratedsystemarchitecture,theOLImodel,relating

Fig.2. DesignFocus.

totheactivities,decisionsandcomponentsunderlyingthedemand for,anddeliveryof,freighttransportservices.Thisdomainmodel ofthePIalsospecifiesaninformationneed.TheIAforthesystemto satisfythisneedcanbedesignedbasedonaRAM,bydefiningthe componentsofthemodelinthedomaincontext.Together,these sketchthedesignproblem,whereourfocusliesonthedesignof thePIPortT&TIA.

3.1. DesignScienceResearch(DSR)

ResearchwithinthefieldofISsisconsideredtobeadiscipline thatcombinestechnicalresearchonIT,theapplicationandbusiness usesofIT,aswellasitsnatural,social,andbehaviouralscientific dimensions(GregorandHevner,2013).AccordingtoWeber(2018), withintheISresearchdiscipline,traditionallytherearetwotypes ofresearch:(1)classicalresearch,and(2)DSR.Theclassicaltype ofresearchfocuses onbuildingand testing theories,while DSR focusesonbuildingartefactsthatcouldbeusefultoaparticular actorcommunity.DSRhasitsrootsinengineeringand fundamen-tallyworksaccordingtoaproblem-solvingparadigm(Baskerville etal.,2018).DSRinvolvestheconstructionofawiderangeof socio-technicalartefacts,suchasdecisionsupportsystems,modelling tools,methodsforISevaluationandchangeintervention,and gov-ernancestrategies(GregorandHevner,2013).AccordingtoHevner

(2007),everyDSRprojectshouldhave(1)itsproblem,(2)its

(bene-fitting)environment,(3)thetobedesignedartefact,and(4)clearly identifiedanddefinedcontributiontoknowledge.Baskervilleetal.

(2018)summarizethattheDSRparadigmcombinespractical

rel-evanceandscientificrigortoISresearch,throughitsemphasison designingusefulartefactsandformulatingdesigntheories.

InlinewithHaraldsonetal.(2020),wearguethatthefreight

transport and logistics system can be considered as a large-scale socio-technical system that consists of various functional subsystemsandoperatesinacomplexenvironment,which corre-spondinglyincludesalargesetofparticipatingactors.Ourresearch canbepositionedinthelightofthefourmainDSRelementsofas follows:

1TheproblemisthatcurrentISofportsarenotabletoprovidethe necessaryvisibilityandinteroperability,intermsofT&Tof logis-ticsoperations,tofullyoperateinahyperconnectedPInetwork withitsmodularPIcontainers.

2The(benefitting)environmentconsistsofactorsinthelogistics chainthatareinvolvedintheshippingandtradingofgoods.As summarizedbySallezetal.(2016),theseactorscanbe catego-rizedinto:shippersandreceivers,transportserviceproviders,hubs, andcoordinators.

3ThetobedesignedartefactisaninnovativeIA,whichisbasedon theRAMI4.0,fortheT&TfunctionofmaritimeportsinthePI.A

Fig. 3.Reference Architecture Model for Industry(RAMI) 4.0 (adaptedfrom:

Adolphsetal.,2015).

suitablewaytotesttheapplicationofRAMI4.0isthroughause

case(Adolphsetal.,2015).Hence,tokeepthedesignrootedin

areal-worldsituation,inSection4,weshowtheapplicabilityof thetobedesignedartefactthroughanillustrativeusecase. 4 Themaincontributiontoknowledgeofourresearchisthedesign

ofatractableand reproducibleIAfor theT&Tfunctionality of maritimeportsinaPIcontext.

3.2. ReferenceArchitectureModelforIndustry4.0(RAMI4.0) Asmentionedearlier,inasimilarmannerasthePI,Industry 4.0hasthepotentialtoimpactentireindustries(Oesterreichand

Teuteberg,2016).InlinewithIndustry4.0,RAMI4.0wasintroduced

byAdolphsetal.(2015).InRAMI4.0,thedesignofobjectsofthe

physicalanddigitalworldarecombinedintoaholisticapproach bymeansofdifferentlayers.Itstructuresexistingstandards, iden-tifiesmissing(linksbetween)standards,andhighlightsareasthat needstandardization(Weyrichand Ebert,2015),whileoverlaps andredundanciesbecomevisibleandopentodiscussion(Adolphs

etal.,2015).

As can be observed from Fig. 3, RAMI 4.0 comprises three dimensionsthatareusedtoviewoneparticular(sub)systemfrom differentangles(Fleischmannetal.,2016):

• Layersseparatetheconcernofinteroperability,and understand-ingofsyntaxandsemanticsfromdifferentviews.Also,thelayers serveasinterfacesbetweenthephysicalanddigitalworld. • Hierarchy Levelsenable a functional allocation of (sub)system

components, and therefore, this dimension can beused as a guidelinetoallocatethedifferentmodulesofasystem.Fromthe perspectiveofthisdimension,theRAMI4.0derivesits classifica-tionfromtheIEC62,264andIEC61,512standards.

• LifeCycle&ValueStream(LC&VS)allowstheclassificationofa par-ticularstateinwhichthe(sub)systemcurrentlyfindsitselfinthe LC&VS.Fromthestandardizationperspectiveofthisdimension, theRAMI4.0derivestheLC&VSfromtheIEC62,890standards. 3.3. ScopingofRAMI4.0forthedesignproblem

Firstly,whenconsideringthethreedimensionsofthe frame-work, our focus will lie on the Layers and Hierarchy Levels dimensionsfordesigningtheRAMofaPIport’sT&Tsystemunder practical conditions. Although the dimension of LC&VS, which concernsitselfwiththedynamicprocessofmigrationand imple-mentationfromtheworld oftoday intothat ofthefuture, isa significantone,ourprimaryobjectiveistoproposeadesignfor

anIAoftheT&TsystemofPIports.Hence,wewillconsiderthe singleandconstantpointintimeofanimplementedPI.

Secondly,althoughtheCommunicationandIntegrationlayersare includedintheRAMI4.0,thesemainlyconcerntheITtechnologies thatcombineandtransmitinformationfromtheAssetlayerintothe Informationlayer.Itisatthislevelthattechnologicaloptionssuchas blockchainenterthedesignofthesystem.Inourdesign,however, wemakethechoicetoabstainfromspecifyingthesetechnologies, aswebelievethatthesechoices arenotessentialtosketchthe functionalityoftheIS,andwillevendistractusfromdoingso.For readersthatareinterestedinthesespecifictwolayersina logis-ticscontext,werefertoLietal.(2016).Theemphasisofourpaper liesonthedesignoftheAsset,Information,Functional,andBusiness layersoftheIA.

4. Teesportasillustrativeusecase

Intheprevioussection,weintroducedRAMI4.0todesignanIA forPIports’T&Tsystems.Inthissection,weintroducetheTeesport asan illustrative usecase through which we aimto showthe applicabilityofourmethodology. Inaddition,we aimtoderive requirementsfromtheTeesport case tousefor theconceptual designoftheIAoftheT&Tsysteminthenextsection.

TeesportcanbeconsideredasanexampleofthePortCentric Logistics(PCL)paradigm.PCLcanbedefinedasproviding value-addedservices(VAS),suchasproductlocalization,warehousing anddistribution,labelling,qualityinspections,light manufactur-ingandfinalassembly,withinportperimeters(Moniosetal.,2018). IntegratingVASatportsenableslogisticsnetworkstobeless com-plexand,amongothers,removesthenecessityofmakinganextra stopatotherdedicatedlogisticscenters.PCLhasbeenarguedtobe themainconceptofthenextgeneration(intheevolution)ofports

(Moniosetal.,2018).Fromthisperspective,PCLcanberegardedas

anearlygenerationPIport,whichisexpectedtobeanincreasingly dominant,activeandintelligentagentinthelogisticschainthrough thedynamic exchangeofgoodsandinformation withitsactors

(Montreuiletal.,2018).WeinvestigatedtheconceptofPCLand

itscurrentpracticalimplementationstounderstandpotential use-fulcontributionsofthethreePIcomponentsof(1)encapsulation, (2)interfaces,and(3)protocols.Encapsulationthrough modular-izationisexpectedto,amongothers,contributetodecreasingthe numberofusedcontainersthroughimprovedspaceutilization.By theuseofinterfacesandstandardprotocols(inT&Tsystems),both visibilityofPIcontainersin-andoutsidetheport,and intercon-nectivitybetweenports,andbetweenportsandotheractorsinthe logisticschainareexpectedtobeenhanced.

4.1. PositionofTeesportinthelogisticschain

Fig.4showsthatManufacturerX,whichisaShanghaibased T-shirtandswimsuitmanufacturer,andManufacturerY,whichis aHongKong basedtelevisionmanufacturer,shiptheirproducts throughthePortofShanghaiandPortofHongKong,respectively, bymeansofmaritimecontainertransporttoTeesport,theportof discharge.OncearrivedatTeesport,theshipmentswillbe reposi-tionedaccordingtotheirnextorfinaldestination,asforexample theRetailer’sdistributioncenter(DC),andwillcontinuetheir jour-ney.

4.2. EnvisionedoperationsatTeesport

Fig.5showsamoredetailedschematicofanexampleof envi-sioneddecompositioning and (re-)compositioningoperations at Teesport.Two T-containersarrive at Teesportfrom thePort of Shanghai and Port of Hong Kong. As indicated by the orange, green,andbluerectangles,oncetheinboundT-containersarriveat

Fig.4. ThelogisticschainoftheTeesportcase.

Fig.5.Decompositioningand(re-)compositioningoperationsatTeesport.

Teesport,theyaredecomposedintheDecompositionphase.Next, P-containersandH-containers,are,again,composed(or consoli-dated)intoH-containersandT-containersinthe(re-)Composition phaseaccordingtotheiroptimalroutingandconsolidation oppor-tunities,whicharedetermined,amongothers,bythevariablesof finaldestinationanddesiredtime-window.Here,P-containersand H-containersarecomposedintoaT-containerinsuchawaythat spaceisoptimallyutilized,andtheyarereadytobedispatchedto theretailer’sDC.Inthemeantime,the“leftover”P-containersand H-containersarestoreduntilthereareenoughforanext destina-tioninadesiredtimewindowtobeconsolidatedanddispatched. 4.3. EnvisionedT&Tsystem

WhenweconsidertheTeesportcase,wearguethat,by imple-mentingtheproposedIA,enhancedvisibilitywillbegainedonthe twoinboundcontainersbymeansoftheT&Tsystem,throughbeing abletoaccesslocalandglobaldatawhichhasbeenprovidedby logisticsactorsthroughthePI’sOpenInterface(PI OI).Thisdata allowsTeesporttoplanitsoperationsinadvanceanddynamically accordingtotheoptimalcontainer(re-)configurationsbefore out-bounddispatching.ModularP/H/T-containersmight,forexample, havechangingstates,routes,andestimateddepartureandarrival times.Inaddition,intermsofenhancedinterconnectivity,changes in relevantlocalandglobaldataarerequiredtobedetectedby

Teesport’sT&Tsystemand sharedwithotherrelevantactorsin thelogisticschain(e.g.vessels,shippinglines,transportsuppliers, consignees)throughthePIOI.

ThefollowingrequirementsforPIportsanditsT&Tsystemcan bederivedfromtheTeesportcase:

• Theportneedsphysicalanddigitalaccessibilityonallthreetiers ofmodularcontainerstoincreasinglybecomeadominant,active andintelligentagentinthelogisticschainthroughthedynamic exchangeofgoodsandinformation;

• Theportneedstobeabletoretrievehighqualityandusefuldata (e.g.weight,state,commoditytype,estimatedarrivaland depar-turetimes,originanddestination,andenvironmentalconditions) abouttheincomingshipmentstobeabletodetermineoptimal (re-)compositioningconfigurationsfortheutilizationofspace, consideringoptimalroutesanddeliverytimewindows;and • Theportneedstohavereal-timeaccesstobothlocalandglobal

dataonmodularPIcontainersinthePIOI,andviceversa.

5. Conceptualdesign

AfterhavingintroducedthemethodologyinSection3and hav-ingpresentedtheillustrativeTeesportcaseinSection4,thissection proposestheconceptualdesignofthePIPortT&TIS’IA.However, tosupportthisdesign,wefirstdefineaminimalscopeforourIA

design whichobviatesthedefinitionofspecifictechnologiesfor hardwareandsoftware.

5.1. DesignscopinginrelationtothefullRAMI4.0framework InlinewiththescopingofourresearchinSection3,inspiredby

Fleischmannetal.(2016),ourdesignwilloperationaliseareduced

versionofRAMI4.0(seeFig.6),whichincludesthedimensionsof layersandhierarchylevels.Aswewanttoemphasizetheexchange ofinformationandstayclearfromadiscussionofspecific technolo-giestostoreandexchangeinformation,ourfocusisonthedesignof theAsset,InformationandFunctionallayersoftheIA,giventheneeds identifiedintheBusinesslayer.Wearguethatourdesignisneutral totechnologychoicesmadeintheIntegrationandCommunication layers.Inasecond-rounddesign,afollow-uponthisresearchwill beneededtocontemplatealternatives,evaluatethem(basedonthe abilitytosupportthisIAandoncriterialiketechnologyreadiness), andspecifytheselayersindetail.

Inthisframework,dataofthelogisticsentitiesfortheT&T func-tionalityisacquiredontheAssetlayer,wheretheinformationflows startfrom.Thedataisacquiredviaalow-levelinterfacebymeans of aFieldDevice,suchassmarttags (e.g.RFID).TheT&T Engine and theWEBEnginearealsopartoftheAssetlayer.Aftergoing throughtheIntegrationandCommunicationlayer,whichallowsfor thetransitionfromthephysicalanddigitalworld,onthe Informa-tionlayer,firstly,theinternallocaldataflowsareacquiredbythe middlewareplatformofthehigh-levelinterfacetosupportthePI PortT&TIS.Thiscanbedonebyconnectinglocalportentitiesvia localdataflows.Secondly,thePortT&TISenablestheexchange oflocaldataflowsandexternaldatafromexternallogisticsentities throughcollaborativeagreementsbetweenactorsinthelogistics chainbymeansoftheInterconnectionmodulebyexploiting inter-facesandstandardizedPIprotocolsinthePIOpenInterface(PIOI). ThePIOIrepresentstheinterfaceandinterconnectionwithallother relevantactorsinthePInetwork.ThePIcontainers’T&T informa-tionandtheinterconnectivityoftheISofPIportsareimplemented ontheFunctionallayer,whichcontainsallthenecessaryfunctions. ThehighestBusinesslayercontainstheoverallbusinessmodel, reg-ulatoryframeworkandrespectiveoperations.

Wecandefinefourmodulesalongwiththehierarchylevelsin theadaptedversionofRAMI4.0.ThePerceptionmoduleservesto perceivelocaldatafromthephysical(logistics)entitiesduringthe operations.TheProcessingmodulegeneratestheT&Tdatabymeans oftheT&TEngine,whereastheHuman-MachineInterfacemodule enablesthecommunicationwithclientsbymeansofaWEBEngine. The Interconnectionmoduleconnects theport’sISwithexternal logisticsentities’ISbymeansofthePIOItofacilitateinformation exchange.Theoverallfunctionofthefourmodulesdeterminesthe informationflowsofthePIcontainers’T&TdatawithinPIportswith respecttothefouraddressedlayerswithintheRAMI4.0.

Below we describe these four layers to operationalize the depictedreferenceframeworkinFig.6forourspecificpurposes, leadingtotheIAdesign.Inatop-downsequencewedescribethe BusinesslayertohavetherequirementsclearfromthePI,and sub-sequentlyturntotheFunctional,InformationandAssetlayers. 5.2. Businesslayer

The Business layer refers to the business processes, and describesthelogisticsoperationsaswouldhappeninthePI,tohave aclearstartingpointforthedesignoftheunderlyinginformation processes.Here,wefurtherbuilduponthefoundationsofthe busi-nessprocessesandlogisticsoperationsthathavebeenillustratedin theTeesportcaseofSection4.Fig.7visualizestheoperational pro-cessesofapartofalogisticschaininthePI,usingaBusinessProcess ModelandNotation(BPMN)diagramthatstartsataportterminal

andendsataconsignee.Amajordifferenceintheprocesseswith thetoday’ssituationisthepresenceofvariouslevelsofPI contain-ers(P/H/T-containers)inPIports.Anothermajordifference,asalso illustratedbeforeintheTeesportcase,istheabsorptionof(some of)theVAS,suchasdecompositioningand(re-)compositioinngof PIcontainersbyPIports.Theblue-highlightedoperationsinFig.7

specifythenewandPIspecificoperationsattheport.Thefollowing assumptionsholdinthisdesignoftheoperationalprocesses: • LoadingunitsinPIportscanbeP/H/T-containers;

• T&TsystemsarelinkedwiththePIOIformultilateralinformation exchange;

• Modularcontainers areembeddedwithsmarttags capableof providingdatatoPIports;and

• Thede-and(re)compositioningofPIcontainerstakesplaceatthe port.

Fig.7showsthatthebusinessprocessesincludenewoperational activitiesrelatedtodecisionmaking,theacquisitionof decision-makinginformation,andthepublishingofupdated information thatresultsfromtheimplementationofthesedecisions.Allthese servethede-andrecompositioningofPIcontainers,atdifferent levelsofmodularity,astheneedsarise.

5.3. Functionallayer

TheFunctionallayerisaformaldescriptionoftheinformation processingfunctionsoftheinternalT&TfunctionsforthePI con-tainers,togetherwiththeinteractionswithexternalISsbymeans ofthePIOI.ThesefunctionsarederivedfromtheBusinesslayer,so thatintheIA,themodelworkflowsanddataflowsintersectwith logisticsactivities.TheperformanceoftheFunctionallayerhasa newmeaninginPIportscomparedtothecurrentsystems,asit nowalsorepresentstheintegrationbetweeninternalT&Tsystems ofPIportsandthePIOI.

ThislayerismodelledbymeansofanActivityDiagram,asshown inFig.8.AimingatthemajorT&Tfunctions,thefigureshowsthe internalelementsoftheT&Tsystemsinports,theexternalelements ofthePIOI,andtheuserof thePIOI.Inaddition,itshowsthe interactionbetweendifferentelementsinsidetheT&Tsystemand thePIOI,andbetweenthesesystems.Informationflowsareused asprimaryinputoftheseactivitiesandinteractions.

Asoneofthenotabledifferencesfromthecurrentsystems,the FunctionallayerofPIportsincludesthePIOI.ThePIOIcomprises threeprimary components:(1) Databaseserver, (2)Application ProgrammingInterface(API),and(3)Interface(web).Regarding therequestsfromusers,thefront-endinterfacegraspstherequests andcallstheAPItoprocessthemwithauthentication.Byrequest, theAPIcanfeedinformationintothedatabase(DB),or alterna-tivelyretrieveinformationfromit.DBshavebeensimplifiedinthe laneoftheDBserverasPIDBanduserDB.ThePIDBcorresponds withtheDBofthevessel,transportationsupplier,PIcontainers, and transport status. Anotherdifference canbe pointed out as aconsequenceoftheintelligenceofPIcontainers.Whereas cur-rentlythefunctionofinformationhandlinganddecisionmaking isdistributedovermultipleactorsinthelogisticschain,PI con-tainerswill,bymeansofsmarttags,havethecapabilitytocollect therelevantinformationthemselvesandmakingtheirown deci-sionsaccordingtothelatestknownstateofthesystem(Sallezetal.,

2016).

TheFunctional layer of RAMI 4.0 hashighlighted interoper-ability betweentheT&T systemand thePI OIwitha focuson informationexchange.Incontrastwiththereciprocal communica-tionincurrentportsystems,thePIOIenablesallrelevantactorsto exchangetheirinformationinamultilateralmannerwiththe sup-portofanAPIandDBserver.Inthenextsubsection,wedescribe

Fig.6. AdaptedversionofRAMI4.0fortheT&TsystemofPIPorts.

Fig.7. BPMN-diagramofEnvisionedOperationalProcessfromtheBusinessLayer.

howdataisusedtocomposetheinformationelementssupportthe Functionallayer.

5.4. Informationlayer

In the Informationlayer, therelevant attributes and opera-tionsofshipmentsarerecordedandstoredasdigitalsourcesand exchangedindataflows.TheInformationlayerelaboratesonthe informationexchangeandtheprovisionofstructureddatavia

ser-viceinterfaces fromoneentity toanother,whileensuring data integrity,consistentintegrationofdata,and obtainingnewand high-qualitydata.

Fig.9showsthecontextdiagramoftheInformationlayer,and providesaformaldescriptionofrulesandtheexecutionof event-relatedrules.Theserulesinitiateprocessingofinformationinthe Functionallayer.Inourcase,localandexternaldataflowsbetween internal and external logisticsactors and entitiesare the main subjectofthislayer.Thedataflowsreflecttheinterdependencies

Fig.8. ActivitydiagramfortheFunctionalLayerofPIPorts’T&TInformationArchitecture.

Fig.9.ContextdiagramfortheInformationLayerofPIPorts’T&TInformationArchitecture.

betweentheT&Tsystemofaport,thePIOI(WebPlatform),andthe otherlogisticsentities.Incontrastwithcurrentsystems,PIports sendusercredentialinformationtothePIOI(WebPlatform)for authenticationandauthorizationofdata,andnotdirectlytoother logisticsactorsinabilateralmanner.InthePIOI(WebPlatform), theAPIauthenticatesthePIportandretrievesdatafromthePIOI’s DBserver.TheinformationfromtheDBserveristransferredthe otherwayaroundfromtheDBtothePIOIthroughtheAPI.

Reflecting ontheT&Tsystem,theretrieved information can beusedasinputforT&Tinformationtoforexampleoptimizeits decompositioningand(re-)compositioningoperations,asalso indi-catedintheTeesportcase.TheundertakenoperationsinPIports

canderecordedintheDBofoperations.Inturn,PIports’T&T infor-mationis alsotransmittedintothePIOIfor theuseofexternal logisticsactorsandentities.Depending onactors’specifictasks andinvolvement inaparticular shipment’slogisticschain,they willreceive respectiveauthorizationtodatain thePI OI. Ship-pers,forexample,canreceivetheT&Tinformation onalllevels ofPIcontainers,inwhichtheirshipmentisencapsulated.LSPsand transportationsuppliersaresimilarlyauthorizedtoalltypesofT&T informationofmodularcontainers,dependingontheirspecifictask andinvolvementinthelogisticschain.Incontrast,shippinglines mostlydealwithT-containersinshippingoperations,and there-fore,mostlikelytobeauthorizedtoretrievedataonT-container

level. Customs agencies willagain be authorized to beable to receivethemostdetailedinformationaboutcontainersonevery level.

5.5. Assetlayer

TheAssetlayerwithinRAMI4.0describestheattributesofthe physicalassets,suchas,forexample,components,machinesand factoriesofasystem.Inourcase,itisdesignedtoclarifythe char-acteristicsandrelationshipsoflogisticsentitiessuchasvessels,PI containersandvarioustypesofterminalequipment,suchasquay cranes,yardcranes,andotherinternalvehicles.Webuildonthe entitiesaswe envisionthemin aPIport,tobeabletosupport allthehigher-levellayersoftheAIinaPIcontext.Assumingthat T&TsystemsofPIportsareinterconnectedwiththePIOIasa web-basedplatform,portsareenabledtocommunicatetheinternalT&T informationofPIcontainerswithotherlogisticsactorsand enti-ties.InformationflowsofPIcontainersfulfilthefunctionsofT&T viathelocalT&Tinterface,wherethePIDBandtheUserDBarethe intermediatestepsoftheinformationflowsthroughthePIOI(see

Fig.8).

Comparedtoanon-PIenvironment,theAssetlayerwillneed tocaptureincreasedinteractions inoperations andinformation exchange within ports, as well as new attributes of contain-ers.TheAssetlayerreflectsthephysicaldifferencewithcurrent T&TsystemsthroughtheuseofPIcontainers,which ultimately are expected to contribute to more efficient space utilization, enhanced visibility, and seamless multimodal multi-party flow throughenhancedinterconnectivity.Informationrelatedtoweight, currentlocation,originanddestination,routing,estimatedarrival anddeparturetimes,andstateshouldberegisteredbythePI con-tainersandmadeavailabletothePIOIandtherebyotherrelevant actors.EmbeddedsmartsensorsinPIcontainers,whicharealso partofthislayer,areusedforthepurposeofretrievingthisdata.

6. Discussion

Withrespecttotheoverallapproach,wepositionedourresearch in thelight ofthefourmainelementsof DSRin Section3.We designed a newartefact in terms of an IA based onRAMI 4.0, whichbenefitstheactorsinthelogisticschain,andsatisfiesthe PI’srequirements.WeshowthattheuseofRAMI4.0facilitates sys-tematicreasoninganditsapplicabilitybymeansoftheTeesport case. TheIApresentedinthispaperhighlightstheorganization, functions,interactivityof,andinteractionbetweentheinformation flowsinsidetheportandwiththePIOI.Themaindesignlimitations oftheworkaretwofold.Asthedetailsofitsimplementationarenot demonstratedyetinreallife,theperformanceoftheproposedIA cannotbevalidatedandevaluated.However,thefunctional illus-tration oftheIAintheTeesportcaseprovidesinsightsintothe functioningoftheIAinpracticeanditsbenefitsforPIportsand itsactors.Anotherclearlimitationofourdesign,althoughfor jus-tifiablereasonswhichareexplainedinSection3,istheexclusion oftheCommunicationandIntegrationlayers.

Fromanoperationsperspective,Montreuiletal.(2018)pointed outthattherearesevenfundamentaltransformationsfrom cur-rentintohyperconnectedlogisticscityhubsinPI.Wearguethat ourdesignsupportsthefollowingthreetransformationsthatare of majorimportance to maritime PI ports: (1) becoming more agilethroughreal-timedynamicandresponsiveshippingtimes; (2) becoming capable of conducting smart, real-time dynamic decisionsonthecontainerconsolidationandinternalflow orches-tration; and (3) becoming active agents in the PI network, dynamically exchanging real-time information on thestatus of parcels, containers, vehicles, routes, and theother hubs. These

majortransformationsareaimedatoptimizingportoperationsto minimizevesselcongestion times,and achieve ofeconomies of scale,handlingefficiencies,andenhancedsecurity.

Fromaninformationalperspective, Sallezetal.(2016)listed identification,T&T,statemonitoring,datacompatibilityand inter-operability, and confidentiality as being essential in the PI.

McFarlane et al. (2016) emphasizes that the value of T&T can

becapturedin accessibility,quality, andusefulness of informa-tionthroughoutthelogisticschain,thusimpactingtheoperational efficiencyandstrategiccompetenciesofthesupplychainandits multipleparticipatingactors.Lindetal.(2020)introducedthe con-ceptof PortCDMwhich will benefit all actors in the maritime logisticschainbymoreefficientdatadistributionandusage.By implementingourproposedIA,theaforementionedvalueofT&T canberealizedandacontributiontotherealizationofPortCDM canbemade.Inthissense,ourdesignoftheIAalsoextendsthe commondatamodeloftheModulushcaprojecttothePI contain-ersbyfocusingontheAsset,Information,FunctionalandBusiness layers.

PCSs positively impact port community performance by connectingITsystemsofeachofitsmembersandenabling commu-nication(Calderinhaetal.,2020).Althoughthisalsocountsforour design,ourIArepresentsthefunctionallevelofaportsystem,and isnotareplacementforPCSs.CurrentPCSsdonottrackandtrace onshipmentlevel,whiletheproposedIAdoes,however. Further-more,ourdesignstatestheneedforthePIOItoallowPIportsto exchangeinformationwithexternalactorsinthelogisticschainto increasevisibility,bothinsidetheportandthroughoutthelogistics chain.Alternatively,PCSscouldfulfiltheroleof“PIcoordinator” asspecifiedby Sallezet al.(2016), offeringglobal information-basedservicesforinteroperabilityandcoordinationofshipments. Dependingonitsrole inthePI,PCSscouldalsoadoptthe pro-posedIAanditsfunctionalities.Clearly,therearemanypotential interactionsbetweenportsandPCSsinthePI.

WhenconsideringtheOLImodel,itmustbekeptinmindthat, beingatranslation oftheOSI, itaddresses PIsystemprotocols, whileRAMI4.0focusesonthesupportingICT.Wepositionthe Busi-nesslayerasOLI’sreflectionintheIAbyshowinggeneralbusiness processesandoperationsofthePI.Inaddition,wenotethatthe operationsoftheOLI’sPhysicalLayer(L1),LinkLayer(L2),and Net-workLayer(L3)willbeconductedbytheport’sT&Tsystem,since theselayersdealwith(1)operatingandmovingphysicalelements, (2)detectionandcorrectionofeventsfromthephysicallayerby meansof adigital twin, and(3) interconnectivity,integrity and interoperabilitywithinthenetwork,respectively(Montreuiletal., 2012).Furthermore,theservicesoftheRoutingLayer(L4)andthe ShippingLayer(L5)areessentialtoPIports’T&Tsystemandits respectiveIAsincethesemonitorthePIcontainers’informationas theyflowacrossthenetwork,definetheshipmentcompositionof PIcontainers,anddecideontheirrouting.

7. Conclusionsandfutureresearch

Theproblemaddressedinthispaperisthatportsneedtoadapt theirT&Tsystemsiftheywanttobecomepartofandplayan essen-tialrole in theglobalPInetwork.Currently, portsonlysupport T&Tinformationatcontainerlevel,whileinthePI,theloadunits thatencapsulateindividualshipments,i.e.PIcontainers,including thesurroundingmodularloadsystem,becomerelevant.Untilnow, therehasbeennodesign-orientedworktoenablethefunctioningof portT&TsystemsforthePI.Ourmaincontributiontoresearchisthe tractableandreproducibledesignofanIAfortheT&Tfunctionality ofmaritimeportsinaPIcontext.

TheIAdesign approachallowsustoexplore thepotentialof keyPIelementsforportstocopewithfuturechallenges inthe

PI.The applicationoftheRAMI 4.0visualizesthelogisticsin PI ports,includingtheinformationflowsregardingtherequired logis-tics entities,the activitiesand interactions in T&Tsystemsand respectiveoperationalprocesses.Bymeansofencapsulationand modularity,spaceutilizationisenhancedbycreatingloadingunits through the three standardized levels (P,H, and T) of PI con-tainers. The PI OI platform allows PI ports to manage various informationalinteractions betweeninternal andexternal actors forpurposesofoptimizingoperations,andadditionally,increase visibilitythroughoutthelogisticschainontheseloadingunitsby linkingtheT&TsystemtoexternalISs.Theusedprotocolsinthe IA improvethevisibilityinPIportsbyproposingguidelinesfor PIportsandexternalactors.TheTeesportcasedemonstratesthe futurecapabilityofPIportstodecomposeand(re-)composetheir inboundshipmentsonthebasisofthestandardizedlevelsofPI con-tainerswithappropriateinformation accessibilityandimproved visibilityinportlogistics.

AsstandardizationandinvestmentsinglobalT&Tsystemsare keyprerequisitesforagloballyfunctioningPI,werecommendthat future work exploresITaspects of logisticsoperations in more depth.Inparallelwithourdesigncase,wealsofindthatthePImay requirediversedesignmodelsthat,forconsistencypurposes,can bebasedonthesamereferenceframework.Theseshouldbeinline withpracticalsituationstosupportlogisticschainvisibilityneeds intheoryandpractice.

Newresearchcouldapplymoreextensivetestingofthe infor-mation flows and the architecture, along with the various PI logisticsentities.Quantitativemethodsincombinationwith sim-ulationsonPIportscouldbeconductedtoevaluatehowthethree PIcomponentsenhancespaceutilization,supplychainvisibility, andserviceofferingcapabilities,comparedtocurrentT&Tsystems. In addition,theintegrationoftheinformation flowswithinthe designedarchitectureintoexternalISs,bymeansofforexample PCSs, is alsoforms a potential futureresearch subject.Another avenueforfutureresearchwouldbethegeneralapplicabilityof ourdesigntoothertypesofPIhubs,suchasrail-,air-,androad hubs.Althoughinourdesign,wefocusedonspecificallymaritimePI ports,generalapplicabilityofourdesignisexpected,with appropri-ateextensionsandadaptations.Lastly,althoughweintentionally excludedtheCommunicationandIntegrationlayersinthedesign oftheIA,anextstepinthedesigncouldbetospecifytheexact technology(soft-andhardware)thatbestsupportsourdesign.

Authorstatement

We wouldliketothankyoufortheopportunitytosubmita revisedversionofourmanuscript,AnInformationArchitectureto EnableTrack-and-TraceCapabilityinPhysicalInternetPorts.

Weareverygratefulforthevaluablecommentsofthe review-ers, which we have incorporated into the revised version of ourmanuscript aswellaswecould.Inaddition,we attacheda responselettertothereviewers’comments,inwhichweexplain ourresponsetoeachofthereviewers’comment.

Thankyou,again,fortheopportunitytoimproveandsubmita revisedversion,andthankyouforyourconsiderationfor publica-tion.

DeclarationofCompetingInterest

Theauthorsdeclarethattheyhavenoknowncompeting finan-cialinterestsorpersonalrelationshipsthatcouldhaveappearedto influencetheworkreportedinthispaper.

Acknowledgements

The research is funded by NWO (Dutch Research Council), PortofRotterdamandGroningenSeaports(grantnumber NWO-4381525), and is conducted in partnership with the Georgia InstituteofTechnologyandtheUniversityofGroningen.

References

Adolphs,P.,Bedenbender,H.,Dirzus,D.,Ehlich,M.,Epple,U.,Hankel,M.,etal.,2015].

StatusReport-referenceArchitectureModelIndustrie4.0(rami4.0).VDI-Verein DeutscherIngenieureeVandZVEI-GermanElectricalandElectronic Manufac-turersAssociation,Tech.Rep.

Ballot,E.,Montreuil,B.,Thivierge,C.,2012].Functionaldesignofphysicalinternet

facilities:aRoad-railHub.In:12thIMHRCProceedings,Gardanne,France2012, p.13.

Ballot,E.,Montreuil,B.,Meller,R.D.,2014.ThePhysicalInternet:TheNetworkofthe

LogisticsNetworks.LaDocumentationFrancaise,Paris.

Bangemann,T.,Bauer,C.,Bedenbender,H.,Diesner,M.,Epple,U.,Elmas,F.,etal.,

2016].Industrie4.0-TechnicalAssets:BasicTerminologyConceptsLifeCycles

andAdministrationModels.VDI/VDEandZVEI.

Baskerville,R.,Baiyere,A.,Gregor,S.,Hevner,A.,Rossi,M.,2018].Designscience

researchcontributions:findingabalancebetweenartifactandtheory.J.Assoc. Inf.Syst.19(5),3.

Becha,H.,Schröder,M.,Voorspuij,J.,Frazier,T.,Lind,M.,2020].Globaldataexchange

standards:thebasisforfuturesmartcontainerdigitalservices.In:Maritime Informatics.Springer,Cham,pp.293–307.

Betti,Q.,Khoury,R.,Hallé,S.,Montreuil,B.,2019].Improvinghyperconnected

logis-ticswithblockchainsandsmartcontracts.ITProf.21(4),25–32.

Bisogno,M.,Nota,G.,Saccomanno,A.,Tommasetti,A.,2015].Improvingthe

effi-ciencyofPortCommunitySystemsthroughintegratedinformationflowsof logisticprocesses.Int.J.Digit.Account.Res.15.

Boyes,H.,Hallaq,B.,Cunningham,J.,Watson,T.,2018].Theindustrialinternetof

things(IIoT):ananalysisframework.Comput.Ind.101,1–12.

Byun,J.,Woo,S.,Kim,D.,2017].Efficientandprivacy-enhancedobjecttraceability

basedonunifiedandlinkedEPCISevents.Comput.Ind.89,35–49.

Calatayud,A.,Mangan,J.,Christopher,M.,2019].Theself-thinkingsupplychain.

SupplyChain.Manag.Int.J.24(1),22–38.

Calderinha,V.,Felício,J.A.,Salvador,A.S.,Nabais,J.,Pinho,T.,2020].Theimpact

ofportcommunitysystems(PCS)characteristicsonperformance.Res.Transp. Econ.,100818.

EPCSA,2011a].WHITEPAPER–theroleofPortCommunitysystemsinthe

devel-opmentofthesinglewindow.EuropeansPortCommunitySystemsAssociation EEIG.June15th_,2011.

EPCSA, 2011b]. How to Develop a Port Community System. European

Port Community System Association, Accessed on https://www. unece.org/fileadmin/DAM/trade/Trade FacilitationForum/BkgrdDocs/ HowToDevelopPortCommunitySystem-EPCSAGuide.pdfon10-02-2019.

Fleischmann,H.,Kohl,J.,Franke,J.,2016.Areferencearchitectureforthe

develop-mentofsocio-cyber-physicalconditionmonitoringsystems.In:June201611th SystemofSystemsEngineeringConference(SoSE),IEEE,pp.1–6.

Galati,F.,Bigliardi,B.,2019].Industry4.0:EmergingThemesandFutureResearch

AvenuesUsingaTextMiningApproach.Comput.Ind.109,100–113.

Gregor,S.,Hevner,A.R.,2013].Positioningandpresentingdesignscienceresearch

formaximumimpact.MisQ.,337–355.

Haraldson,S.,Lind,M.,Breitenbach,S.,Croston,J.C.,Karlsson,M.,Hirt,G.,2020].

Theportasasetofsocio-technicalsystems:amulti-organisationalview.In: MaritimeInformatics.Springer,Cham,pp.47–63.

Hasan,H.R.,Salah,K.,Jayaraman,R.,Yaqoob,Omar,M.,2020.Blockchain

Architec-turesforPhysicalInternet:AVision,Features,Requirements,andApplications. IEEENetwork.

Hevner,A.R.,2007].Athreecycleviewofdesignscienceresearch.Scand.J.Inform.

Syst.19(2),4.

Hoffmann,J.,Asariotis,R.,Assaf,M.,Benamara,H.,2018.UNCTADReviewofMaritime

Transport.,pp.2018.

Hofmann,E.,Rüsch,M.,2017].Industry4.0andthecurrentstatusaswellasfuture

prospectsonlogistics.Comput.Ind.89,23–34.

IPCSA,2018].SustainableFreightTransportinSupportofthe2030Agendafor

Sus-tainableDevelopment.UNCTADMultiyearExpertMeetingonTransport,Trade LogisticsandTradeFacilitation,Geneva,pp.21–23,November2018.

Lasi,H.,Fettke,P.,Kemper,H.-G.,Feld,T.,Hoffmann,M.,2014].Industry4.0.Bus.Inf.

Syst.Eng.6(4),239–242.

Lee,P.T.W.,Lam,J.S.L.,2016].Developingthefifthgenerationportsmodel.In:

DynamicShippingandPortDevelopmentintheGlobalizedEconomy.Palgrave Macmillan,London,pp.186–210.

Li,Q.,Cao,X.,Xu,H.,2016].In-transitStatusperceptionoffreightcontainerslogistics

basedonmulti-sensorinformation.In:InternationalConferenceonInternetand DistributedComputingSystems,Springer,Cham,pp.503–512.

Lind,M.,Ward,R.,Watson,R.T.,Haraldson,S.,Zerem,A.,Paulsen,S.,2020].Decision

supportforportvisits.In:MaritimeInformatics.Springer,Cham,pp.167–186.

McFarlane,D.,Giannikas,V.,Lu,W.,2016].Intelligentlogistics:involvingthe

Meller,R.D.,Montreuil,B.,Thivierge,C.,Montreuil,Z.,2012].Functionaldesignof physicalinternetfacilities:aRoad-basedtransithub.In:12thIMHRC Proceed-ings,Gardanne,France2012,p.42.

Meyer,G.G.,Främling,K.,Holmström,J.,2009].Intelligentproducts:asurvey.

Com-put.Ind.60(3),137–148.

ModulushcaProject(2017).http://www.modulushca.eu/.

Mondal,S.,Wijewardena,K.P.,Karuppuswami,S.,Kriti,N.,Kumar,D.,Chahal,P.,

2019].BlockchaininspiredRFID-basedinformationarchitectureforfoodsupply

chain.IEEEInternetThingsJ.6(3),5803–5813.

Monios,J.,Bergqvist,R.,Woxenius,J.,2018].Port-centriccities:theroleoffreight

distributionindefiningtheport-cityrelationship.J.Transp.Geogr.66,53–64.

Montreuil,B.,2011].Towardsaphysicalinternet:meetingthegloballogistics

sus-tainabilitygrandchallenge.Logist.Res.3(2–3),71–87.

Montreuil,B.,Ballot,E.,Fontane,E.,2012].AnOpenlogisticsinterconnectionmodel

forthephysicalinternet.In:Proceedingsof14thIFACSymposiumon Informa-tionControlProblemsinManufacturing(INCOM2012),Bucharest,Romania.

Montreuil,B.,Meller,R.D.,Ballot,E.,2013].Physicalinternetfoundations.In:Service

OrientationinHolonicandMultiAgentManufacturingandRobotics.Springer, Berlin,Heidelberg,pp.151–166.

Montreuil,B.,Ballot,E.,Tremblay,W.,2016].Modulardesignofphysicalinternet

transport,handlingandpackagingcontainers.In:Smith,J.(Ed.),Progressin MaterialHandlingResearch,Vol.13.MHI,Charlotte,NC,USA.

Montreuil,B.,Buckley,S.,Faugere,L.,Khir,R.,Derhami,S.,2018].In:Carrano,A.

(Ed.),UrbanParcelLogisticsHubandNetworkDesign:TheImpactof Modular-ityandHyperconnectivity,ProgressinMaterialHandlingResearch:2018.MHI, Charlotte,NC,USA.

Oesterreich,T.D.,Teuteberg,F.,2016].Understandingtheimplicationsofdigitisation

andautomationinthecontextofIndustry4.0:atriangulationapproachand elementsofaresearchagendafortheconstructionindustry.Comput.Ind.83, 121–139.

Pan,S.,Ballot,E.,Huang,G.Q.,Montreuil,B.,2017].PhysicalInternetand

inter-connectedlogisticsservices:researchandapplications.Int.J.Prod.Res.55(9), 2603–2609.

Pisching,M.A.,Pessoa,M.A.,Junqueira,F.,dosSantosFilho,D.J.,Miyagi,P.E.,2018].

AnarchitecturebasedonRAMI4.0todiscoverequipmenttoprocessoperations requiredbyproducts.Comput.Ind.Eng.125,574–591.

Raap,W.B.,Iacob,M.E.,vanSinderen,M.,Piest,S.,October2016.Anarchitecture

andcommondatamodelforopendata-basedcargo-trackinginsynchromodal logistics.In:OntheMovetoMeaningfulInternetSystems.Springer,Cham,pp. 327–343.

Romero,D.,Vernadat,F.,2016].Enterpriseinformationsystemsstateoftheart:past,

presentandfuturetrends.Comput.Ind.79,3–13.

Sallez,Y.,Berger,T.,Bonte,T.,Trentesaux,D.,2015].Propositionofahybridcontrol

architecturefortheroutinginaPhysicalInternetcross-dockinghub. IFACPaper-sOnLine48(3),1978–1983.

Sallez,Y.,Pan,S.,Montreuil,B.,Berger,T.,Ballot,E.,2016].Ontheactivenessof

intelligentPhysicalInternetcontainers.Comput.Ind.81,96–104.

Tian,F.,2016].Anagri-foodsupplychaintraceabilitysystemforChinabasedon

RFID&blockchaintechnology.In:201613thInternationalConferenceonService SystemsandServiceManagement(ICSSSM),IEEE,pp.1–6.

Treiblmaier,H.,Mirkovski,K.,Lowry,P.B.,Zacharia,Z.G.,2020].Thephysicalinternet

asanewsupplychainparadigm:asystematicliteraturereviewanda compre-hensiveframework.Int.J.Logist.Manag.31(2),239–287.

UNESCAP,2018.CapacityBuildingWorkshoponStrengtheningIntegrated

Inter-modalTransportConnectivityforSoutheastandSouth-SouthwestAsia.United NationsEconomicandSocialCommissionforAsiaandthePacific,Bangkok,9 March2018.

VanGeest,M.,Tekinerdogan,B.,Catal,C.,2021].Designofareferencearchitecture

fordevelopingsmartwarehousesinindustry4.0.Comput.Ind.124,103343.

Walha,F.,Bekrar,A.,Chaabane,S.,Loukil,T.M.,2016].Arail-roadPI-huballocation

problem:activeandreactiveapproaches.Comput.Ind.81,138–151.

Watson,R.T.,Lind,M.,Delmeire,N.,Liesa,F.,2020].Shipping:aself-organising

ecosystem.In:MaritimeInformatics.Springer,Cham,pp.13–32.

Weber,R.,2018.Design-scienceresearch.In:ResearchMethods:Information,

Sys-tems,andContexts.ChandosPublishing,pp.267–288.

Weyrich,M.,Ebert,C.,2015].Referencearchitecturesfortheinternetofthings.IEEE

Softw.33(1),112–116.

Yaqoob,I.,Ahmed,E.,Hashem,I.A.T.,Ahmed,A.I.A.,Gani,A.,Imran,M.,Guizani,

M.,2017].Internetofthingsarchitecture:recentadvances,taxonomy,