arcadia project - scientific workshop report - eucnc 2016

21
©2016, ARCADIA project. http://www.arcadia-framework.eu. Deliverable D6.11 Scientific Workshop Report V1 Editor(s): Matteo Repetto Responsible Organization(s): CNIT Version: 1.0 Status: Final Date: 28/07/2016 Dissemination level: Public Ref. Ares(2016)4467999 - 17/08/2016

Upload: eu-arcadia-project

Post on 22-Jan-2018

58 views

Category:

Technology


0 download

TRANSCRIPT

©2016,ARCADIAproject.http://www.arcadia-framework.eu.

DeliverableD6.11

ScientificWorkshopReportV1

Editor(s): MatteoRepetto

ResponsibleOrganization(s): CNIT

Version: 1.0

Status: Final

Date: 28/07/2016

Disseminationlevel: Public

Ref. Ares(2016)4467999 - 17/08/2016

D6.11–ScientificWorkshopReportV1

2/21

Deliverablefactsheet

TitleofDeliverable: D6.11–ScientificWorkshopReportV1

RelatedWP: WP6–Communication,DisseminationandExploitation

Duedateaccordingtocontract: M14

Editor(s): MatteoRepetto(CNIT)

Contributor(s): AnastasiosZafeiropoulos(UBITECH)

Reviewer(s): LukaszPorwol

Approvedby: AllPartners

D6.11–ScientificWorkshopReportV1

3/21

RevisionHistoryVersion Date Editor(s) Remark(s)

0.1 23/06/2016 M.Repetto(CNIT) Documenttemplate,TOC

0.2 04/07/2016 M.Repetto(CNIT) Added short summaryof topics discussedattheworkshop

0.3 18/07/2016 M.Repetto(CNIT) Added main discussions items in finalpanelandconclusions

1.0 28/07/2016 M.Repetto(CNIT) Changedaccordingtointernalreview

Statementoforiginality:This deliverable contains original unpublished work except where clearly indicated otherwise.Acknowledgementofpreviouslypublishedmaterialandof thework ofothershasbeenmade throughappropriatecitation,quotationorboth.

D6.11–ScientificWorkshopReportV1

4/21

ExecutiveSummaryThisdeliverableprovidesareportfromthefirstARCADIAworkshop,heldinAthensduringEuCNC2016.Theworkshop gave insights on the ARCADIA architecture, on-going work, prototype implementation, andestablishedconnectionswithotherrelatedinitiatives.

The workshop included presentation of the ARCADIA project, presentations from the industry, andpresentationsfromotherEUprojects.Followingtheinvitedspeeches,apanelwasorganizedtodiscusstherelevanceoftheARCADIAframeworktotheevolvingcloudlandscape,tohighlightnoveltrendsthathaveemerged after the project proposal, to reinforce collaboration with parallel EU projects, and to gathersuggestions, ideas, remarks inorder to improvetheproject.Tothisaim,membersof theAdvisoryBoardwereinvitedtoattendtheworkshop.

Thisdocumentbrieflysummarizesthemaincontentoftheinvitedspeechesandthemaintopicsdiscussedduring the panel. It also summarizes the main directions and suggestions that emerged during thediscussion.

D6.11–ScientificWorkshopReportV1

5/21

TableofContentsREVISIONHISTORY......................................................................................................................................................................3EXECUTIVESUMMARY................................................................................................................................................................4TABLEOFCONTENTS..................................................................................................................................................................51 INTRODUCTION.....................................................................................................................................................................61.1 SCOPE.........................................................................................................................................................................................................61.2 AUDIENCE..................................................................................................................................................................................................61.3 STRUCTURE................................................................................................................................................................................................6

2 WORKSHOPFACTSHEET....................................................................................................................................................73 OBJECTIVESANDMOTIVATIONS.....................................................................................................................................83.1 OBJECTIVES................................................................................................................................................................................................83.2 MOTIVATIONS............................................................................................................................................................................................83.3 SCOPE.........................................................................................................................................................................................................83.4 ORGANIZATION.........................................................................................................................................................................................9

4 WORKSHOPSUMMARY....................................................................................................................................................104.1 WELCOMEANDINTRODUCTION..........................................................................................................................................................104.2 NETWORKSOFTWARIZATION..............................................................................................................................................................104.3 NFVANDSDNINTERPLAYANDORCHESTRATIONCHALLENGES...................................................................................................124.4 THEARCADIAFRAMEWORK:ACOMPLETEFRAMEWORKFORSOFTWAREDEVELOPMENT,ORCHESTRATION,DEPLOYMENT,ANDEXECUTION..................................................................................................................................................................................................134.5 BRINGINGCOMPUTINGATTHENETWORKEDGEFORVIRTUALIZATIONOFDEVICESANDTHINGS:THEINPUTINFRASTRUCTUREFORFOGCOMPUTING.........................................................................................................................................................144.6 PANELDISCUSSION................................................................................................................................................................................154.6.1 DesignofSDN/NFVsolutions....................................................................................................................................................164.6.2 Energyefficiencyin5Ginfrastructures................................................................................................................................17

5 MAINOUTCOMESFROMTHEWORKSHOP................................................................................................................186 CONCLUSIONS......................................................................................................................................................................20ANNEXI:REFERENCES.............................................................................................................................................................21

D6.11–ScientificWorkshopReportV1

6/21

1 Introduction

1.1 ScopeAspartofitsdisseminationstrategy,ARCADIAplannedtoorganizethreescientificworkshops,targetedtothescientificand industrialcommunities in thedomainsofsoftwaredevelopment,cloudcomputing,andvirtualization[1].

Thisreportdocumentsthe1stARCADIAworkshop,plannedforpresentingtheproject’sintermediateresult,namely the definition of the ARCADIA ContextModel and the design and specification of the ARCADIAFramework.

1.2 AudienceThisdeliverableisintendedforpublicdissemination(includingtheCommissionServices).

1.3 StructureThedocumentisstructuredasfollows:

§ Section1introducesthedeliverable,itsscope,anditsstructure.

§ Section2providesaquickfactsheetoftheworkshop.

§ Section 3 explains the main objectives for this workshop and the overall strategy behind itsorganization.

§ Section Error! Reference source not found. reports a brief summary of the speeches and finaldiscussion.

§ Section6givestheconclusions.

D6.11–ScientificWorkshopReportV1

7/21

2 Workshopfactsheet

Title ARCADIA Special Session on Software engineering approaches aligned with theSoftwarizationofNetworksandServices

Event EuropeanConferenceonNetworksandCommunications(EuCNC2016)

http://www.eucnc.eu

Athens,Greece,June27-30,2016

Venue DivaniApollonPalace&Thalasso,RoomKleonikiB

AgiouNikolaou10,Vouliagmeni,Athens16671,Greece

Date Wednesday,29June2016,11:00-12:30

Chair RaffaeleBolla(CNIT,Italy)

Participants 40 participants attended the workshop, including people from the industry (telcos,vendors),peoplefromSMEsactiveinthe5Gdomain,aswellaspeoplefromtheacademiaandresearchinstitutes.

Programme 11:00-11:20“Networksoftwarization”,AntonioManzalini(TIM)

11:20-11:40 “NFV and SDN interplay and orchestration challenges”, Pedro A. ArandaGutiérrez(TELEFONICAI+D)

11:40-11:55“TheARCADIAFramework:acompleteframeworkforsoftwaredevelopment,orchestration,deployment,andexecution”,AnastasiosZafeiropoulos(UBITECH)

11:55-12:10 “Bringing computing at the network edge for virtualization of devices andthings:theINPUTinfrastructureforfogcomputing”,RobertoBruschi(CNIT)

12:10-12:30Panel:“Deployingdistributedapplicationsovernext-generationprogrammableinfrastructure: vision, challenges and research directions”, moderated by Dr. NikosKoutsouris(WINGS),Dr.AntonioManzalini(TELECOMITALIA),Dr.RobertoBruschi(CNIT).

D6.11–ScientificWorkshopReportV1

8/21

3 Objectivesandmotivations

3.1 ObjectivesThe1stARCADIAworkshopwasconceived forpresenting theproject’s intermediate resultsafter the firstyearofactivity,andtogatherfeedbacks,interests,suggestionsfromthecommunity.

3.2 MotivationsThe first year of activity in ARCADIA was mainly devoted to framework specifications, including theidentification of relevant use cases, the collection of requirements, the outlining of the overarchingARCADIA framework (software development paradigm and orchestration environment), the preliminarydefinitionof theContextModel for developing anddeployingARCADIA-compliant applications.All theseactivitieswerecarriedoutinWorkPackage2(WP2)anddocumentedbytherelativedeliverables1.

Theworkundertakenduring the firstyearwasspecifically focusedtoblur thedemarcation linebetweeninfrastructureand softwaredevelopment,byunderstandinghow theemergingprogrammabilityof cloudenvironments could be reflected and exploited by programmers in order to make their applications“ductile”,namelyadaptable tomutableexecutionenvironmentsandconditions. In this respect, the ideabeyond the first workshop was to foster the discussion about the synergy between novel softwaredevelopmentparadigmsandtheincreasinglevelsofinfrastructureprogrammability;inparticular,ARCADIAenvisions a still largely unexplored potential of network programmability in the implementation anddeploymentofcomplexsoftwareecosystem,includingtheconceptofSoftware-DefinedNetworking(SDN).

Giventheaboveconsiderations,theProjectdecidedtopresenttheARCADIAvisionandpreliminaryresultsinanetworking/communicationcontext,deferringthedebatewiththesoftwaredevelopmentcommunityto the followingworkshops, whenmore focused technical achievements would have been available fordiscussion.

3.3 ScopeThegrowofInternet-centricparadigms(liketheInternetofServices,theInternetofThings)isboosting

theneedformoredistributedandmodularizedapplications.Newcomputingarchitecturesareemergingtotackletheseparadigms,likecloudfederationsand“fog”computing.Thisevolutionbringsanewdimensiontothesoftwaredevelopmentprocess,whichshalltakeintoaccounttheprominentroleofnetworkingandits implications for distributed systems (QoS, security, dependability, ...) as well as the increasing trendtowardsthesoftwarizationofnetworksandservices.

Given the heterogeneity of the computing and networking solutions, it is vital to introduce novelsoftware development paradigms that facilitate the automatic deployment and reconfiguration of theapplicationsoverdistributedinfrastructures,bytakingadvantageoftheincreasingprogrammabilityofsuchunderlying infrastructures. To this aim, principles and techniques already available in recent softwareengineering paradigms shall also consider ground-breaking concepts in the networking domain, likeNetworkFunctionVirtualization(NFV)andSoftwareDefinedNetworking(SDN).

1PublicdeliverablesareavailableontheARCADIAwebsite.URL:http://www.arcadia-framework.eu/wp/documentation/deliverables/.

D6.11–ScientificWorkshopReportV1

9/21

Theintentionoftheworkshopwastogatherprominentexpertsfrombothacademicandindustryandpromote an open discussion on the development of Reconfigurable-by-design applications, fostering atightercollaborationbetweentheSoftwareDevelopment,theComputingandtheNetworkingdomains.

The workshop provided an overview of on-going work in the EU projects ARCADIA and INPUT, andpresentation of related initiatives in the SDN/NFV domain. The main goal was to share bleeding-edgeknowledge, to cluster complementary activities, and to stimulate the synergy and cooperation fromdifferenttechnologydomainsandresearchproject.

3.4 OrganizationGiven themotivation outlined in Section 3.2, theworkshopwas organized to foster the discussionwithrelevant experts andparallel initiatives from the networking domain. Key topics for the discussionwerecutting-edgetechnologiesinthisfield,namelyNetworkFunctionVirtualization(NFV)andSoftware-DefinedNetworking. Invited speakerswere selected inorder todepict a clear andauthoritative visionabout themain trends in the industryand in the research community. Further, a representative from theAdvisoryBoardwashighlyrecommended.

Thefollowingpeoplewereinvitedforaspeech:

• Dr.Antonio Manzalini, Innovation/Future Centre at TIM (Italy). His interests include innovativenetworkingtechnologiesandarchitectures,networkvirtualization,5G,pervasiverobotics,InternetofThings.HeisstronglyinvolvedinSDNstandardization.

• Dr.DiegoR.R. Lopez,SeniorTechnologyExpert,Telefonica I+D(Spain).He is responsible for thedefinitionandcoordinationofresearchprojects intheareasofnewnetworkingtechnologiesandnetwork infrastructures.He is currentlyChairof theETSINFVTechnical SteeringCommittee,Co-chair of the NFVRG. Dr. Lopez was replaced by Dr. Pedro A. Aranda Gutiérrez, TechnologyExploration, Network Innovation & Virtualisation, Telefonica I+D. He is the coordinator of theNetIdeproject.

• Dr.RobertoBruschi,ResearchScientistatCNIT(Italy).HeistheProjectCoordinatorof INPUT,anH2020researchprojecttargetingnovelvirtualizationparadigmsatthenetworkedge.

Dr.AntonioManzaliniisalsoamemberoftheARCADIAAdvisoryBoard.

D6.11–ScientificWorkshopReportV1

10/21

4 WorkshopsummaryThis Section reports a brief summary from each speech. It is not intended to be a mere copy of thepresentations, rather toextrapolate thekeymessagesand tobring them to theattentionof theprojectConsortiumandallinterestedreaders.

ThefullpresentationsareavailableontheARCADIAwebsites(onlyforConsortiummembersandtheEC).

4.1 WelcomeandintroductionProf.RaffaeleBolla,chairoftheworkshop,gaveaninitialintroductiontothemainscopeoftheworkshop,aswellastheprogramthatisgoingtobefollowed.

4.2 NetworksoftwarizationSPEAKER:AntonioManzalini,TIM

D6.11–ScientificWorkshopReportV1

11/21

Networksareundergoinganunprecedentedtechno-economictransformationdrivenbytheevergrowingdeployment of IT equipment within communication infrastructures, with the prime aim to realizemostlegacy functions in software. This trendwill ultimately deploy a large amount of computing and storageresources into the network, building a vast programmable execution environment that open theopportunities for far more services beyond mere voice/data transport (5G networks). Telecoms arecarefully looking at this evolution, because of the higher revenuemargins for value-added services overtheirtraditionalbusinessmodel.

The key driver in this evolution is the definition of some kind of middleware that is able to runheterogeneous services on heterogeneous and distributed hardware, in a transparent, seamless andautomatedway.Thismiddlewarecanbedesignedinasimilarwaytoamodernoperatingsystem,offeringstandardized interfaces toapplicationsandservices (similar tosystemcallsand libraries),abstracting theresource diversity (through some abstraction similar to device drivers). The Network Operating System(NOS) should provide a platform for automated operation, bringing together communicationinfrastructures, IT equipment, terminals. The definition of a NOS entails several on-going initiatives,including theEverything-as-a-Service (XaaS)unificationparadigm, service interfacingand templates (e.g.,IETF intent framework, OASIS TOSCA/YANG), virtualization environments (e.g., OpenStack, NFV, ONOS),programmingandconfigurationprotocols(OpenFlow,Netconf,…),commoditizationofTelcocentralofficeand disaggregation into edge/fog components (Telecom Infra Project, Open Compute, Mobile EdgeComputing,etc.).

D6.11–ScientificWorkshopReportV1

12/21

Thetopichasbeendiscussedforseveralyears,andtherearemanycomponentsthatcouldbeassembledtogether in a common architecture (e.g., CORD2 ). However, such a distributed and heterogeneousenvironmentwillbecharacterizedbyahigherlevelofcomplexitythantoday,andthiswillbeachallengingissueforproperautomatedcontrolandmanagement.

ArtificialIntelligence(AI)maybetherightsolutionforcomplexity.AIhasbeenalreadyusedsuccessfullyfordetection of cyber-attacks, and its usage is under investigation in different sectors: car driving, smartambient,financialandmarketprediction,geneticsandDNAprogramming.Butthechallengingquestionis:couldAIbethe“endofcode”?Mayweenvisionanewparadigm,i.e.,AI-definednetworkformakinglarge-infrastructuresandbig-data“actionable”?

4.3 NFVandSDNinterplayandorchestrationchallengesSPEAKER:PedroA.ArandaGutiérrez,TELEFONICAI+D

NFV and SDN are promising technologies for networking. However, their real potential is currentlyoverlooked.Asamatteroffact,manytraditionalnetworkingguysarepronetoseeNetworkFunctionsasthe “open”versionof their legacymiddle-boxesand theSDNcontrolleras thedelightful replacementoftheir patch panels. NFV/SDN bring far more opportunities than deploying network services whilecomfortablysittingattheowndesk!

Thefine-grainedcontroloverpacketforwardingenabledbySDNisveryusefultooptimizeservicegraphsinNFVandtocomposeapplicationsandservices.NFVstronglyreliesontheconceptofrecursivecomposition,startingfromelementarysoftwarecomponents,toVNFsandNetworkServices.Thenotionofapplicationcompositiongoesfarerthanmereservicechaining,envisioningthepossibilitytoperforminparallelseveraltasks(e.g.,firewalling,natting,monitoring),toassembletheresultandtoapplyittothepacket.

From another perspective, SDN is enabling DevOps development cycles which benefit from priorexperiences in software development, such as Agile development. The standard software cyclic lifecyclemade of design, implementation, test, deployment could be easily adopted for network services andapplications,withgreatbenefitsintermsofreleasetimesanddevelopmentcosts.

Finally, another issue in NFV/SDN converge concerns packaging. NFV is already defining its packagingformats (e.g., OpenMANO, TOSCA), while SDN applications are usually shipped in the controller-specificformat(e.g.,jarfile,pythonscript).Thereisstillamissinggapinbringingthesetechnologiestogetherandconvergingtoacommonandoverarchingstandard.

Goingfurther,thelastremarkisforcurrentdescriptionlanguages.Forexample,currentwayofdescribingServiceGraphs inNFV-MANOishuman-readable,but.. is ithuman-understandable?“Intent” is themagicwordinthiscontext;itindicatesthepossibilitytospecifywhat isexpectedfromthenetwork,ratherthanhowtoconfiguretheunderlyinginfrastructure.Currenteffortsinsuchdirectionarestillattheearlystage,andwithseveraldrawbacks(e.g.,FlowObjectives,Boulder,NEMO).

Thefinalmessageisawarningtoavoidbeing“seduced”bythebuzzrecurrentwordsinNFV/SDN:thesearevery powerful concepts, though the implementation of their full potential is very challenging andmanyaspectsstillneedinvestigation.

2CentralOfficeRe-architectedasaDatacenter(CORD),OpenNetworkingLabincollaborationwithAT&T.[Online]Available:http://onrc.stanford.edu/protected%20files/PDF/ONRC-CORD-Larry.pdf.

D6.11–ScientificWorkshopReportV1

13/21

4.4 TheARCADIAFramework:acompleteframeworkforsoftwaredevelopment,orchestration,deployment,andexecution

SPEAKER:AnastasiosZafeiropoulos,UBITECH

TheARCADIAprojectistargetingthealignmentofsoftwaredevelopmentparadigmswiththeopportunitiesofferedbyprogrammableinfrastructure,inordertoenableapplicationsandservicestodynamicallyadapttothecurrentexecutionenvironment.Themaingoalisatightintegrationbetweenthedevelopmentandoperations processes, enabling automatic deployment and life-cycle management through integratedannotationsintotheapplications.

The ARCADIA development paradigm aims at building reactive systems by including more context-awareness inapplicationsandservices; theapproach isbasedon themodularization, theorganization inmicro-servicesandthedelegationofthecontrolandmanagement logictoanexternalentity,namelytheorchestrator. The main driver underpinning the project approach is the availability of programmableinfrastructures, includingbothcloudinstallationsandcommunicationsystems;inthisrespect,ARCADIAisexplicitlytakingintoconsiderationsemergingtechnologieslikeSDNandNFV.

The ARCADIA framework is an overarching systems encompassing both software development,deployment,andruntimemanagement.Softwaredevelopmentisbasedontwomainelements:themicro-servicesarchitecture(whichallowsflexiblecompositionofapplicationsandservices)andaContextModel(whichallows the specificationof relationships,policies, requirements, constraints,performancemetrics,etc). Thedeployment and runtimemanagement system is basedon a smart orchestratorwhich deploys

D6.11–ScientificWorkshopReportV1

14/21

applicationsovermultipleIaaS,byinterpretingtheContextModelandbyadaptingapplications/servicestothecurrentexecutionenvironment.

ARCADIA is currently developing several artefacts to build its framework: a web IDE for insertingannotations into applications and building service graphs, a software repository for ARCADIA-compliantcomponents and service graphs, a policy specification and enforcement framework, an optimizationalgorithm for distributing software components over multiple IaaS, resource management and overlaynetworking.

ARCADIA isparticularly interested inSDN/NFV.Ontheoneside,SDNenables theprogrammaticnetworkcontrol necessary todeployhighly-distributed applications; on theother side,VirtualNetwork Functionsareusefultobeintegratedintocomplexapplicationsandservices(e.g.,loadbalancers,firewalls,etc.).

4.5 Bringing computing at the network edge for virtualization of devices andthings:theINPUTinfrastructureforfogcomputing

SPEAKER:RobertoBruschi,CNIT

TheINPUTprojectistargetingfogcomputingparadigmsbyexploitingICTinstallationattheTelcos’networkedge.Themainmotivationistheincreasingvirtualizationtrend,withaneverlargernumberofapplicationsandservicesmovedto thecloud,and thedifficulty tomeetstringentQoSconstraints for interactiveandmultimedia applications.As amatter of fact, latency andbandwidth are key factors for applications likevoiceandvideoservices,on-linegaming,gridprotection,trading,remotecontrol.

D6.11–ScientificWorkshopReportV1

15/21

INPUTfocusesonaspecificusecase,namelythecreationofpersonalnetworkservices.Followingthetrendtomoveapplications to the cloud,even traditionalhomedevices couldbe (at leastpartially) virtualized,leaving only simpler hardware at home. Several reasons motivate this approach: morecomputation/storage resources available, more flexibility in upgrading the application, less energyconsumption at home. In this respect, this Use Case represents an evolution of the Home GatewayvirtualizationalreadyenvisionedbytelecomsandtheETSINFVframework.ThePersonalNetworkisaverychallengingservice,sincetheendusermusthavethe feelinghisdevicesarestill functioning locally,withnegligiblelatency.

ToimplementPersonalNetworkservices,INPUTisdefininganexecutionenvironmentatthenetworkedge.The execution environment provides virtualization capabilities (both as IaaS and PaaS) in order to runService_Apps,andexploitsin-networkprogrammability(throughSDN/NFV)tocreatethecorrectresourceslicing.ThemainINPUTfeatureisthecapabilitytofollowtheuser,bymovinghisService_Appsclosetohiscurrent point of attachment to the network. The INPUT orchestrator is responsible to decide whichapplicationsshouldmove,inordertoavoidbandwidthwasting.

As part of the project activities, INPUT is developingOpenVolcano, a complete virtualization frameworkincludingthecapabilityofprovidingvirtualcomputing,storageandnetworking,ahigh-performancevirtualswitchbasedonDPDK,andanSDNcontroller.

Two challenging use cases have already been identified to demonstrate the INPUT framework. The firstconcerns IoT: simple sensors are deployed at home, while their management/processing software isdeployed as virtual instance at the network edge and data are stored in a legacy cloud. The secondconcerns the virtualization of set-top-boxes: video is decoded at the network edge and the stream isdirectly sent to network-attached TVs. In both cases the purpose is to show how the INPUT frameworkenablesserviceproviderstodesigntheir(PersonalNetwork)services,todeploytheserviceforeachusers,andtoachieveverylow-latencyperformance.

4.6 PaneldiscussionFollowingthemainopinionsandworkspresentedintheprevioussessions,ashortdiscussionsessiontookintoconsiderationthecritical issuesstilltobetackled,andidentifiedsomeexpectedcontributionsbytheindustrialandscientificcommunity.

D6.11–ScientificWorkshopReportV1

16/21

Themainaspectsemergedfromthediscussionarebrieflydescribedinthefollowing.

4.6.1 DesignofSDN/NFVsolutionsThere has been a great effort in understanding and developing the concepts of Software-DefinedNetworking andNetwork FunctionVirtualization in recent years.Many initiatives have beenundertakenboth in standardization fora, international research projects, and early industrial and commercialprototyping. Such broad range of interests and initiatives have led to many different architecturalapproachesandtechnicalsolutions,reflectivetheheterogeneityinscope,backgroundandobjectives.

Tomove fromconcept toreal-worldproductsandservices, it isnownecessary to finda largeconsensustowards common and shared architectures, interfaces, protocols, andmethodologies, in order to designinteroperablesolutionsthatnourishthelargeecosystemexpected.Inthisrespect,standardizationplaysakeyrole,andallstakeholdersarerequestedtocontribute.

Amongothers,EUprojectsarecarryingoutagreatworkonarchitectures,prototyping,servicedefinitionsinthisfield,butoftentheydonotputenoughemphasisonstandardization.MoreeffortinstandardizationisexpectedfromEUprojects,inordertocapitalizetheirwork,toachievewideconsensus,andtoboosttheconvergencetowardscommonandinteroperabletechnologiesandsystems.

SeveralNFVframeworkshavebeendesignedwiththeirowninternalAPIsandinterfaces,butthecommonbeliefisthatnoneofthemwillprevailontheothers,andtheupcomingindustrialsolutionswillbeanopen

D6.11–ScientificWorkshopReportV1

17/21

ecosystemmadeofinteroperablecomponentscomingfromdifferentinitiatives.Inthisrespects,commonopenAPIs(southbound,northbound,peerinterfaces)areexpectedformanagementofnetworkfunctions,service orchestration as well as support of automated deployment and orchestration of distributedapplications.

Withthelargeplethoraofdifferentsolutionsavailable,itwillbecrucialtocomparetheminastandardanduniformway,inordertoselectthebestapproachesandtofosterconvergenceamongthemostpromisingdirections.

4.6.2 Energyefficiencyin5GinfrastructuresICTinfrastructuresarerequestedtolowertheirpowerconsumptionforlongtermsustainability.Inrecentyears, several studies have demonstrated that a large share of power consumption is ascribable to lowefficiency in the way networking systems are operated. Re-design of networking components and low-powerhardwarehelps in cuttingdownenergy, but it is clear since several years that themost effectivepowerreductionwouldonlycomewithglobalstrategiesthatcoordinatethewholesystem.

SDN and NFV are moving towards new networking paradigms and are expected to bring a tightconvergencewithcomputinginfrastructures.Asamatteroffact,evermoreITequipmentisbeingdeployedin network infrastructure, and the virtualization trend is blurring the boundary between the networkingandtheITworlds.Inthissituation,thechallengeisabetterintegrationbetweennetworkanddatacentres,including peripheral installations at the very edge of the network. There must be a trade-off betweenperformance (in terms of QoS, reliability, availability, dependability) and the power consumed by theinfrastructure. The softwarization trend, which decouples applications from the underlying hardware,provides a great flexibility for workload consolidation and efficient hardware usage, but also bringsadditional complexity in understanding the power consumption and the energy efficiency of softwareinstances.

5G infrastructures are explicitly asked to face the above challenge, and to find sustainable andeffectivesolutionsforaneverincreasingconnectedword.

D6.11–ScientificWorkshopReportV1

18/21

5 MainoutcomesfromtheworkshopSummingupthemaintopicsoutlinesbypresentationsandthediscussedinthepanel,thefollowingmainissueshavebeenidentified:

• Issue 1. There is a need for the harmonization of the "chaos" that exists in the design anddevelopment of various SDN and NFV solutions, including the set of service/network functionsorchestrationapproaches.

• Issue2.ThereisaneedforcontributionsintherelevantstandardizationbodiesonbehalfoftheEUprojects, as well as collaboration among the standardization bodies for production of commonspecifications.

• Issue3.There isaneedforthedesignanddevelopmentofopenAPIs (southbound,northbound,peerinterfaces)formanagementofnetworkfunctions,serviceorchestrationaswellassupportofautomateddeploymentandorchestrationofdistributedapplications.

• Issue 4. There is a need for common benchmarkingmethodologies and set of key performanceindicatorsinordertohavecomparableperformanceevaluationresultsandidentifyadvantagesanddisadvantagesperapproach.

• Issue5.Thereisaneedforthedesignanddeploymentofmechanismsthatcansupportinaholisticway energy efficiency in 5G infrastructures. The cost in terms of energy consumption of thesoftwarizationoftheInternethastobeexamined.

Consequently,alistofsuggestedactionstobeundertakenorstrengthenedbytheARCADIAConsortiuminthenextmonthshavebeenderived:

Currently,somerepresentativesfromARCADIAarealreadytakenpartintothefollowinginitiatives:

• theTechnicalCommitteeonOASISTopologyandOrchestrationSpecificationforCloudApplication(TOSCA)3;

• the ETSI Environmental Engineering committee (TC EE), part of the Energy Efficiency technologycluster4.

More effort is expected to align theARCADIA software framework to parallel initiatives and projects, inordertoidentifycommoninterfacesforserviceorchestrationandautomateddeploymentofapplications.The ARCADIA framework should also target the NVF environment, as a particular Use Case for serviceorchestrationanddeployment.

Themainobjective is toproduceevaluation results that canbeeasily comparedwith thoseproducesbyothersolutionsandarchitectures.

3https://www.oasis-open.org/committees/tc_home.php?wg_abbrev=tosca4http://www.etsi.org/technologies-clusters/technologies/energy-efficiency

Action 1 – Insist on standardization. The Consortium is encouraged to continue its effort instandardizationbodies.

Action2–Commonbenchmarking.TheConsortiumisencouragedtodefineevaluationmetricsfor itsUseCasedemonstratorsincompliancewithotherprojects/initiatives.

D6.11–ScientificWorkshopReportV1

19/21

EnergyefficiencyforsoftwaredeploymentisalreadyincludedinARCADIAUseCases.Thespecificobjectivefor theproject is toparticipate in theemergingdiscussionofenergyefficiency forsoftware,whichcouldprovideanenergy-awarecharacterizationforapplicationsandtheircomponents.

Action3–Energyefficiency.TheConsortiumisencouragedtotakeintoconsiderationenergyefficiency.

D6.11–ScientificWorkshopReportV1

20/21

6 ConclusionsThe1stARCADIAworkshopwasagreatopportunityfortheConsortiumtomeetexpertsfromtheSDN/NFVfields.Frominvitedspeechesandthefinalshortdiscussion,itwasclearthatSDNandNFVarestillevolving,and no prevailing architectures and solutions have clearly emerged yet. Upcoming networkinginfrastructureswillbeverycomplexsystemtomanage,andtheavailabilityofautomaticmechanisms forservicedeploymentandlifecyclemanagementisaninescapableneed.Thisbasicallyconfirmstheworthofthe ARCADIA approach, and also motivates to design quite general solutions, not still tailored to anyexistingframework.

Fromthe1stARCADIAworkshop,theProjecthasearnedageneralapprovalaboutitsarchitectureandtherelevance for theNFVusecase.However, theConsortiumhasalso receivedaclear indication thatmucheffort is expected in standardization and harmonization with other undergoing initiatives. ARCADIA isalreadyparticipating in standardization initiatives,both in the fieldof softwareorchestrationandenergyefficiency. However, the definition of common interfaces and interoperability issues should be explicitlytaken intoaccount. Inaddition,theProjectshouldalsoconsiderthedefinitionofcommonbenchmarkingmetricstoproduceevaluationresultsthatareeasilycomparablewithotherarchitectures.

The ARCADIA Consortium is therefore mainly committed to propose/adopt open interfaces in itsframework, and to take part into relevant standardization fora. The objective is to improve the currentarchitectureandprototypeimplementationbeforethenextworkshop.

Thefollowingtwoworkshopsenvisionedforthesecondandthirdyearsofactivitywilltargetthesoftwaredevelopment community. In such events, the ARCADIA framework will be presented to get preliminaryfeedback(2ndworkshop),thenimprovedandevaluatedinthethreeprojectUseCases(3rdworkshop).

D6.11–ScientificWorkshopReportV1

21/21

AnnexI:References[1] A novel reconfigurable-by-design Highly Distributed Applications development paradigm over

programmableinfrastructure(ARCADIA).EUH2020Programme,GrantAgreementNo.645372.