20 Dec 2012

Toward a contextualized

Information-Centric Architecture

www.huawei.com

Information-Centric Architecture

G.Q Wang

Innovation Center

Futurewei Technology

USA

HUAWEI TECHNOLOGIES Co., Ltd.

Mobile Users(“Internet Trends”, KPCB 12/3/2012, Mary Meeker)

1/7/2013

Emerging applications(“Internet Trends”, KPCB 12/3/2012, Mary Meeker)

1/7/2013

Info-genetic system connection: context-sensible

Bodynet / Homenet interworking1. Smart shoes for personal authentication

2. Smart shoes tells sofa to adjust comfort

3. i-Glasses tells TV to select YouTube channels

4. Phone call comes in

5. TV lowers the volume to let call going

6. TV raises the volume when call finished

7. i-Watch tells thermometer to adjust

1/7/2013

7. i-Watch tells thermometer to adjust

temperature

But we are still at the very beginning …

It is predicted 50 ~ 100 Billion Internet

connected devices by 2020.

Human brain has 86bn neurons

1/7/2013

Compared with human brain, Internet is far inferior to being:Context-awareness, thinkable, processing speed, memory size, energy efficiency, …

Future context-aware connections:from connecting “dumb” things to connecting “thinkable” things

Any

“thinkable”

things

Circuit-centricConnecting

Host-centric Connecting

Information

entities

Information-centric

Scale

Connecting

Cyber,

Biological,

Physical

Cognition-centric

Location

“Who”

“What”

1/7/2013

Connecting

wires

Connecting

nodes

entities

1870 1980 2010 Time2020+

It is predicted 50 ~ 100 Billion

Internet connected devices

by 2020, in which 95% are mobile

Internet Top Challenges: Scalability, Mobility and Security

by 2020, in which 95% are mobile

• Latency vs. Scale • Video: 10+ traffic growth in 5 years

• IoT: 50+ billion hyper-connected devices

• BGP size becomes routing bottleneck

• Mobility and Dynamism • Mobile Internet, Mobile IoT

• Heterogeneous access • Multi-radio, self organized connection

• Mobile IP drops 50% ~ 90% packets in V2V

1/7/2013

• Mobile IP drops 50% ~ 90% packets in V2V

• Security and Accessibility • DDos

• Mobility fails pipe-based security

• Integrate access control and security at

information level

ICN target: Faster, Smarter and Securer

Contextualized Information-Centric Networking Architecture

M2M Smart sensor

Smart Home

Net

Smart Car

Net

Smart City

Net

Smart DC

Net

Contextualized Information Bus (CIBUS)

DC

TelecomICN

SDN ~ 2015

~ 2020

Contextualized Information Bus (CIBUS)• Auto node/service discovery• Context-aware routing/forwarding• user empowered firewalling

• name-based security/mobility integration• Context-sensible multi-radio access• Content multi-homing and caching

1/7/2013

DC

DC

Internet

ICN Protocol and ApplicationsICN Protocol and Applications

1/7/2013

P-I-D: a naming schema for ICN

Entity

Security

bindingApplication

binding

Network

binding

Naming

Identifier DomainPrincipal

TopTopTop

Named-Entity Example:

{ (@!~&%) , (huawei/innovation-center/Able.iPhone) , (Skype/HLR) }

P I D

Huawei

Innovation-center

Abel.iPhone Abel.Desk-Phone

ATT.US

LTE WiFi

SF-GW-1 SF-GW-2

Identifier Locator

Skype

HLR

3rd party-domain

resolution

1/7/2013

Naming functions• P is used for data authentication

• P:I is used for information access

• D is used for cross-domain routing label

resolution and late-binding

Open Context-Aware Protocol (OCAP) PDU

Message type = Interest / Data Forwarding mode (m)

Source-entity-name (P:I:D) (o) Dest-entity-name (P:I:D) (m)

Checksum (o) TTL (m)

Source-entity-locator (o) Dest-entity-locator (o)

o: optional

m: mandatory

resolution and late-binding

• Locator is used for routing/forwarding (e.g.,

longest prefix match with a routing label)

Checksum (o) TTL (m)

Signature (o)

Context Array (o)

Device type LBS Selector Others

Payload (o)

Nonce(m)

1/7/2013

Naming: (P:I:D) Provenance , Identifier, Domain

�Address security-content binding

�Late-binding & scalable routing

Security: context-aware & info-centric

� scalable and flexible trust management

�Name-based key management

Mobility: context-sensible multi-radio access

�Rendezvous-based intra/inter domain Mobility

�Global/local name resolution & mobility control

Routing: context-guided multipath

�Extended OSPF & BGP for intra/inter domain content routing

�Self-organized disruptive-tolerant routing

Forwarding: content multi-homing

� Context-aware multi-path and opportunistic forwarding

� Hop-by-hop traffic engineering

Caching: context supervisory

� Collaborative caching and cognitive routing

D2D Service publishing & discovery

� Name-based service publishing and automatic discovery

ICN/SDN integrated

Optical Core

Wifi

Net

Wifi

Net

PoA1SRN1

SP/CPG-SAP1

Google

SRN3

Cloud

Service

SRN4

NetflixSRN6

Scenario-1: Network as a Data Center

Optical Core

LTE

Net

LTE

Net

PoA2H-AP1

WLANWLANLANLAN

Amazon

SRN5

PoA3 PoA4

H-AP2

SRN2G-SAP2

SP/CP

Generic Service Access Point (G-SAP) Platform

– Service Routing Node (SRN): ICN-enabled

– Heterogeneous Access Point (H-AP): access

transparent

– Point of Attachment (PoA): access specific

Distributed DCs provide edge-cloud services– Partitioning and replicating a service across multiple globally distributed cloud nodes

– Generic service access point integrated with heterogeneous mobility

– Integrated with a SDN-enabled and programmable optical transport

– Cross-layer control to utilize multiple path forwarding for resiliency and efficiency

– Standard inter-cloud interface

– Point of Attachment (PoA): access specific

– Service pool / Content pool (SP/CP)

1/7/2013

profile_1 (name1, dom1, type1, sig1)profile_2 (name2, dom2, type2, sig2)

ccnd(node name)

Service1

Service2

SPDPublish API

Local Service Publishing

Scenario-2: self-organized D-2-D Networking

ccnd(n1)

ccnd(n2)

SPD2SPD1

FIB: disc: fsdisc/pseduo2: f1disc/n1:f1

FIB: disc: fsdisc/pseduo1: f2disc/n2:f2

fs fs

f2 f1

I

I: n=disc/pseduo1, h=1

I D

D: n=disc/pseduo1, d=n2

D

Name-based Neighbor Discovery

ccnd(n1)

SPD1

ccnd(n2)

SPD2 ccnd(n3)

SPD3

ccnd(n4)

FIB: disc: fsdisc/n2:f2

App

Discovery API

fs

f2 f1

fsf3

f4

I: n=disc/n2/n1, h=2, dn=x

SPD3

I3: n=disc/n3/n2, h=1, dn=x

I4: n=disc/n4/n2, h=1, dn=x

D4: n=disc/n4/n2, d=profiles

I4

D4

I3

D3: n=disc/n3,/n2 d=profilesD

D

FIB: disc: fsdisc/n1:f1; disc/n3:f3; disc/n4:f4

Name-based Service Discovery

I

D = D2 + D3 + D4

D

D3

Per-device Service Publishing &

Discovery (SPD)• Auto device and service discovery

• Self-organized ad hoc and collaborative virtual

groups

• Self-controlled intra/inter-AS routing 1/7/2013

Q&A:Q&A:

So is ICN the Future ?

1/7/2013

20 Dec 2012

Thank You

www.huawei.com

Thank You

www.huawei.com

HUAWEI TECHNOLOGIES Co., Ltd.