Carmelita Görg Universität BremenUniversität Bremen … · Carmelita Görg Universität...

Looking 4ward t th F t I t tto the Future Internet

Carmelita GörgUniversität BremenUniversität Bremen

ContentsContents

Introduction – who, why and what?

Migration and Innovation – 1000 networks bloomg

How can we build these networks?

What do we want or connect in these networks –people, things, information ...

How to transport data? What is a path?

How can we manage these networks?How can we manage these networks?

A Summary

23/11/2008© 4WARD Consortium Confidential

PM/Slide 2

Geographical distribution

EricssonKTHSICS

EricssonNokia-Siemens-Networks

VTT Alcatel-LucentDeutsche Telecom

EricssonEricssonNokia-Siemens-Networks

TU BerlinUniv. of Bremen

Univ. of KarlsruheUNiv. of PaderbornFinland

NECUniv. of

Lancasterf S

Canada

NorwaySweden

WITTelekomunikacja Polska

Univ. of SurreyEricsson

UK GermanyPoland

Austria

WITSiemensTPUCN

Alcatel-LucentFrance Telecom

GET INT

Romania

Ireland

Switzerland

Spain

France

Italia

Israel

Technion

GET-INTLIP6

Univ. of Basel

IST-TULPTIN

Telcom ItaliaRobotiker-Tecnalia

TelefonicaUS

© 4WARD Consortium Rutgers

University

Acknowledgements

4WARD Consortium4WARD i f d d b th E C i i4WARD is funded by the European Commissionunder the EU‘s 7th Framework Programmewww 4ward project euwww.4ward-project.eu

Norbert Niebert, The Way 4WARD to the Creation of a Future Internet, ICT Mobile Summit 2008, Stockholm, June 2008

CEWIT 2008

Network (R)Evolution – How?Network (R)Evolution – How?

I it j t IP 7 IP 8 ti lt ti ?Is it just IPv7 or IPv8 … or creating an alternative?By adding and patching we run into the risk of “obesity and a weak immune system”obesity and a weak immune systemOr should we even dare to think of tailor-made networks, fit for the purpose and reliable?fit for the purpose and reliable?Will migration work?

How can we solve this?


PM/Slide 5

The Facets of 4WARD

Business Innovation

Combination of clean slate

Architecture Framework

clean-slate researchapproachesppto address theNetwork of theFuture EndpointFutureSize: Roughly 23 M€

oldi

ng P

oint

Forwarder

Time 2 years

Fo

© 4WARD Consortium

Network Virtualisation – Main InnovationsNetwork Virtualisation – Main Innovations

Network virtualisation as a meta architecture in a commercial settingNetwork virtualisation as a meta-architecture in a commercial setting– Enable co-existence of diverse network architectures– Enable deployment of innovative approaches– Enable new business roles and playersEnable new business roles and players

• Allow split of infrastructure-/network-/service-providers• Lower barriers of entry• „Market place“ for shareable network resources

Provisioning and virtualisation management framework– On-demand instantiation of virtual networks at large scale

Virtualisation of diverse resources in a common framework– Routers, links, servers – can all be done today but need a unifying e2e

approachapproach– Extension on the virtualisation of the wireless infrastructure and spectrum– Folding points providing interworking between virtual networks


PM/Slide 7

ee

Folie 7

eednni1 Need a slide on folding points. The rest is doable. Too much text on the slide - I deleted a few subb ullets.Norbert Niebert; 17.09.2008

Possible Provider Roles in a Virtualised Environment

End-to-End Deployment

Virtual Network Operator

y

VNet VNet VNet

VNet VNet VNet

Phys. Net

InfrastructureBroker

(optional)Substrate SubstrateLegacy

SubstrateSubstrateSubstrate

Legacy Operator

Vertically Integrated Operators(virtualised networks) Infrastructure Providers


PM/Slide 8

Virtualisation of diverse resources in a common frameworkDynamic management of virtual networksEnable co‐existence of diverse network architecturesEnable new business roles and playersV t O t D i

Virtualised Networks

Vnet Operator DomainFolding Points

Vnet Provider Domain

Infrastructure Provider DomainInf‐Provider AVnet‐Provider A

Vnet‐Provider B

Routers ServersRouters, Servers, ...

Links, ...


PM/Slide 9

Inf‐Provider BInf‐Provider CWireless, spectrum, ...

Virtual Radio Framework Slice 1 Slice 2 Slice 3 Slice 4Slice 1 Slice 2 Slice 3 Slice 4

c-plane u-plane c-plane u-plane c-plane u-plane c-plane u-plane

Control

L2L2L2L2

HL HLHLHL Control ControlControl

SliceMgr.

RAC ( li bl k ll ti )

PHYPHYPHYPHY

Virtualisation of Wirelessharmonised resource access

RAC (per slice resource block allocation)

Slices can implement their ownVirtualisation of Wireless Resources & Efficient Spectrum SharingFlexible and cost-efficient d l t f di

Slices can implement their own protocols/methods – routing, mobility management,

namingdeployment of new radio technologiesHarmonised access of slices to a common radio resource block

g– radio protocols, channel

coding, smart antenna steering– cross layer optimisation


PM/Slide 10

a common radio resource block cross layer optimisationScheduling and isolation

Virtual Router PerformanceVirtual Router Performance

• High-performance virtual router platform for modern commodity hardware• Evaluation of virtualised forwarding planes in terms of isolation and fairness


PM/Slide 11

Evaluation of virtualised forwarding planes in terms of isolation and fairness

VNet Instantiation ProcessVNet Instantiation Process

VNet Provider

Install & Configure

GrantRequestOffer

(specific network descr.+

Query(abstracted network descr.

+

MappingCommit

Initiate Inter-Provider Links

(access IDs +certificates)

(Specific network description)+cost and other parameters)query language)

Mapping+

Commit-Request

Commit+

EmbedCommit

+Embed

VMGW VMGWVMGW

Commit+

Embed

PeeringPoint Peering

Point


PM/Slide 12

Infrastructure Provider AInfrastructure Provider BInfrastructure Provider C

How far have we come?How far have we come?

D ft hit t VN customers

Query Language VN requirements

Language/Semantic/Ide/Syntaxe.g.: topolgy description and objectives

• Draft architecture• Scalable mapping algorithms using data mining

technologyVN Provisioning and Management Architecture

VN Request analysis and processing

VN Requests

VN Provider

P

Queue

Scheduling

• Initial definition of signaling and control interfaces• First version resource description language

– Modelling of resources and networks

g g

VN Mapping Framework(VN mapping algorithms and

protocols)

Resource Monitoring &

Discovery

Resource Allocator &

configuratorResource

InformationDatabase

Accounting

Security Management- Authorisation- Authentication

Mobility Mgt, QoS Mgt….

rovisioning & M

anageme

Layer

Modelling of resources and networks– XML-based – Used for request and offer Resource

description(Language/Semantic/Identifiers/Names…)

Virtual Resources(virtual nodes/links, CPU,

memory, bandwidth…)

Distributed Database

ent

WP3-Task3.2Resource

Virtualisation

Virtualisation Management

Configuration Agents

Update Action

Virtual Resou

Layer

– Additional query language for complex requests

• Virtual Radio concept

VirtualNode

Substrate Node

SubstratePath

Virtual Networks

Virtual

Abstract view of substrate from VN provisioning & management point of view

Physical Reso

Layer

Virtualisation

Probes/monitors

urces

p• Early prototyping and testbeds• Controlled Interworking concept

Node

Substrate Link

Substrate Network

VirtualLink

ources


PM/Slide 13

But how to design a network architecture?

B i d t hi i ti t l f t tBy reusing and patching existing protocols we forgot todevelop tools for clean slate design esp. for the moredetailed network architecture specificationsCan such a design toolkit and process be easily developedand used?Where can we find reusable components other than protocolWhere can we find reusable components other than protocolspecs and implementations?How can we ensure interoperability?p y


PM/Slide 14

Requirements (Meta) Modeling

Overview of 4WARD’s Network Design ProcessRequirements

…(Meta) Modeling

Model Driven Design Process(including iterations)

“Blue Print”of network

Building blocks,Netlets

Architecture(selected netlets/strata)

Network architect(needs to have

Inter-operability

Compositionof functionality

Netlets,Strata

Netlets, Abstract strata,Architecture

patterns

(needs to haveknowledge aboutnetworks – worksoff-line, before (!)

operabilityy(CFI)

Strata

4WARD Architectural Framework

operation of the network)

I l t ti

Repository PrototypeRepository


PM/Slide 15

4WARD Architectural Framework Implementation

What a new network architecture could interconnect

W d t thi k t kWe are used to think a networkconsisting of nodes(end + forwarding) and links(end forwarding) and linksWhat if we start to network theinformation we are looking for?Triggered by Van Jacobsen andothers a new view on interconnectingi f ti h dinformation has emerged ...that could change the way we engineer networksfundamentallyfundamentallyThe Networking of Information is looking into this from a systems perspective


PM/Slide 16

y p p

Information Objects and Data ObjectsInformation Objects and Data Objects

AggregationIOs

All AboutSong1 Paris

ServiceIO

EncodingIO

Song1 EiffelVirtual EntityIO

DataObj t

IOIO

Song1 mp3 Song1 wav Lyrics txt Service1Eiffel jpg

IO

S 1 3

Objects Song1.mp3

S 1

Song1.wav Lyrics.txt Service1Eiffel.jpg

ffSong1.mp3Service1

Files Song1.mp3Song1.wav

Song1.wavLyrics.txt

Lyrics.txtLyrics.txt Service1

Eiffel.jpgEiffel.jpg


PM/Slide 17

How to transport?How to transport?

C th t t i i l d ti l di i tCan we assume that turning wireless and optical media intocopper will work forever?And why is it efficient to do transport innovations only asAnd why is it efficient to do transport innovations only asoverlays?What can be gained with a completelyWhat can be gained with a completelyfresh view on transport mechanisms?The Generic Path is an answer tothese questions...


PM/Slide 18

An architecture and a set of mechanismsAn architecture and a set of mechanisms

th G i P th hit tthe Generic Path architecture– a much richer class of data flows, beyond TCP, UDP– state within the network, as necessary but no more than necessary y y– common management interfaces, to set up and tear down flows and

to query their status– explicit identification, notably to facilitate control of multi-flow p , y

applications like videoconferencingmechanisms for assured performance and efficient operation– to exploit techniques like network coding and cooperative– to exploit techniques like network coding and cooperative

transmission– to choose the "best" paths for the considered transport

to ensure resource sharing is "fair" and meets application– to ensure resource sharing is fair and meets application requirements

– to manage the mobility of users, networks and information


PM/Slide 19

Management?!Management?!

Th t t d i d i ld iThe most urgent need in a dynamic world isSelf-ManagementAutomation of Management has been a research topic forAutomation of Management has been a research topic formany yearsDoes it provide in practice more than automated settings onDoes it provide in practice more than automated settings on FI routers?Can we rely on this?What are the new approaches in this area?


PM/Slide 2023/11/2008© 4WARD Consortium Confidential

PM/Slide 20

Monitoring and adaptationMonitoring and adaptation

INM f i i• INM functions aim at predictability– Local optimization loops reduce

0,4

0,6

0,8

1

of a

ggre

gatin

g no

des

h=1h=2

h=3h=4

time to react– Trade-off between accuracy,

timeliness and overhead

0

0,2

1 10 100 1000Updates/sec

frac

tion

o

– Tunable objectives in adaptation algorithms

– Anomaly detection to perform Objectives

isolate exceptions in the network

• Final objectiveNetInf

Final objective– Build full control control loops– Maintain service-level objectives

Enforcing required OAM functionsGP

VNet


PM/Slide 21

– Enforcing required OAM functions


PM/Slide 21

SummarySummary

4WARD f ll b f t h i l i ti4WARD follows a number of technical innovationapproaches in all areas of future networkingThey all start from a clean slate perspective definingThey all start from a clean-slate perspective, definingradically new solutions for the Network of the FutureAfter 11 month work first results look promisingAfter 11 month work first results look promisingNow the hard work of refining and integrating startstowards a new

Network of the Futureas a Family of Networksas a Family of Networks


PM/Slide 22

FISS´09FISS 09

F t I t t S S h l 2009Future Internet Summer School 2009Universität BremenJ l 20 24 2009July 20-24, 2009Mixture of courses, presentations and invited talks by 4ward in cooperation with other Future Internet projects e gin cooperation with other Future Internet projects, e.g. EuroNF, ANA, Chianti, …Courses for graduate students and researchersgSee 4ward website or www.comnets.uni-bremen.de

ComNets FFV 21. Nov. 08© 4WARD

Slide 23

ReferencesReferences

Ines Houidi, Wajdi Louati and Djamal Zeghlache, “A Distributed Virtual Network Mapping Algorithm,” In Proceedings of the 2008 IEEE International Conference on Communications (ICC 2008), May 19-23, 2008 Beijing China pp 5634 – 56402008, Beijing, China, pp. 5634 5640Ines Houidi, Wajdi Louati and Djamal Zeghlache, “A Distributed and Autonomic Virtual Network Mapping Framework,” In Proceedings of the 4th International Conference on Autonomic and Autonomous Systems (ICAS y (2008), IEEE Computer Society, March 16-21, 2008, Gosier, Guadeloupe, pp. 241-247N. Egi, A. Greenhalgh, M. Handley, M. Hoerdt, F. Huici, and L. Mathy; “Fairness Issues in Software Virtual Routers” in Proc PRESTO '08 ACMFairness Issues in Software Virtual Routers , in Proc. PRESTO 08, ACM SIGCOMM Workshops, August 2008, Seattle, USA.Joachim Sachs, Stephan Baucke, “Virtual Radio – A Framework for Configurable Radio Networks”, The Fourth International Wireless Internet C f (WICON 2008) N b 2008 M i USAConference (WICON 2008), November 2008, Maui, USA


PM/Slide 24

Carmelita Görg Universität BremenUniversität Bremen … · Carmelita Görg Universität...

Documents

Transcript of Carmelita Görg Universität BremenUniversität Bremen … · Carmelita Görg Universität...