Information-Centric Networking in Wireless/Mobile Networks

IFIP/ACM LANC 20127th Latin America Networking Conference

Information-Centric Networking in Wireless/Mobile Networks

Torsten Braun, Universität Bern, [email protected], cds.unibe.ch

mailto:[email protected]

http://www.iam.unibe.ch/

Torsten Braun: Information-Centric Networking in Wireless/Mobile Networks

2

Overview

> Future Internet> Information-Centric Networking (ICN)

— ICN Approaches, e.g., CCN and PSI— Service-Centric Networking — ICN and Cloud Computing

> ICN in Wireless / Mobile Networks— Mobility— Delay/Disruption-Tolerant and Opportunistic Networks— Network Coding and Energy Efficiency

> Conclusions

Medellin, October 5, 2012


3

Future Internet

> The current Internet is based on principles of the 1960s with the main purpose to interconnect large computers.

> Today, Internet users are mainly interestedin retrieving content and accessing serviceswith increasing demand for mobility&security

> Increasing bandwidth demand due to— (personalized) multimedia data streams— Cloud computing and storage,

visualization on (mobile / high-resolution) end systems— Huge amounts of data from measurements and scientific

experiments— Internet of Things, Internet of Services, …

> New approaches, e.g., Information-Centric Networking, need more research, development, testing etc.



4

Motivation for Information-Centric Networking

> Today’s network traffic is dominated by information retrieval rather than point-to-point communication between machines or humans.

> Circuit communication model is not considered as appropriate any more.

> Future communication architecture should focus on information objects instead of nodes.

> Today, wires and memories solve complimentary aspects of the same problem:— Wires move information in space.— Memories move information in time.

> Future communications architecture should unify both issues.> Flash crowds are difficult to avoid in IP-based networks.



5

Traditional Web Retrieval / Web Services


web server / web service

DNS server

search engine /service registry

user’s end system

Key Principles and Functions of ICN

> Naming of content rather than hosts/interfaces— Content independent of devices that store it— Naming is location independent (mobility support !)

> Receivers (subscribers) request content.> Senders (publishers) advertise and deliver content.> Receivers and senders do not have to be aware of each

other and are decoupled in time.> Functions needed

— Name resolution to match subscriptions and publications— Routing and path formation— Forwarding content from publisher to subscriber

[Pentikousis, K.; Chemouil, P.; Nichols, K.; Pavlou, G.; Massey, D.; "Information-centric networking [Guest editorial]", IEEE Communications Magazine, vol.50, no.7, pp.22-25, July 2012



6


7

Content Distribution with ICN


/unibe.ch/braun/lecture/20120405

1

32

4 5 76

Naming Approaches

> Human-readable, hierarchical names— supports aggregation— needs coordination— Example: CCN

> Flat (self-certifying) names— Often based on hashing content name and/or owner’s public key— Aggregation more difficult but more flexible. — Examples: PSI, NetInf



8


9

Name Resolution and Data Transport

> Decoupled— Name resolution and data transport are independent of each

other, cf. DNS, with possibly different nodes for resolution and data transport

— allows different, possibly already existing transport mechanisms, also multi-path

— Examples: PSI, NetInf> Coupled

— Nodes for both name resolution and data transport with inverse data path compared to search path

— rather disruptive technology— Local routing procedures advantageous in case of short link

disruptions— Examples: CCN, NetInf



10

Content-Centric Networking (CCN)

> Combination of content lookup and message routing> Idea: describe the user’s interests in the message header,

but not where to get it. > Messages (using XML encoding)

— Interest: content name, selector— Data: content name, signature (info), data

> Hierarchical content names— Example: /unibe.ch/braun/lecture/20120405

> Related Projects— NDN = Named Data Networking, www.named-data.net — CCNx = Open Source Core Software Project for

Content-Centric Networking, www.ccnx.orgVan Jacobson, Diana K. Smetters, James D. Thornton, Michael F. Plass, Nicholas H. Briggs, and Rebecca L. Braynard: Networking named content, 5th international conference on Emerging networking experiments and technologies (CoNEXT '09). ACM, New York, NY, USA, 1-12.


http://www.named-data.net/

http://www.ccnx.org/


11

IP Model


FIB

FIB: Forwarding Information Base

SRC

DST


12

Interest Message Processing in CCN

1. Longest prefix match on content name in Content Store: returning data and discarding Interest

2. Pending Interest Table match: adding request to PIT and discarding Interest

3. Forwarding Information Base match: forwarding of Interest towards data— FIB population by announcements of content availability



13

Match in Content Store


FIB

CS: Content StoreFIB: Forwarding Information BasePIT: Pending Interest Table

Name

CS

Name

Data

Data

Name

PIT


14

Match in Forwarding Information Base


FIBCS

Name

PIT


15

Match in Pending Interest Table


FIBCS

Name

PIT

x


16

CCN Naming

> Hierarchical naming allowing aggregation> Examples

— /unibe.ch/braun/lecture/20120405— /unibe.ch/N10/R306/Projector

> Support for simple operations— %C1.org.ccnx.frobnicate~1~37— command in the namespace org.ccnx— operation is frobnicate, which takes 1 and 37 as arguments



17

IP / CCN Hour-Glass Models


WWW, email, VoiP

HTTP, SMTP, RTP

TCP, UDP

IP

Ethernet, WiFi

Fiber, radio, copper

browser, chat

file, stream

security

contentchunks

TCP, P2P, Broadcast

Fiber, radio, copper

Strategy


18

CCN Routing

> Longest Prefix Match Routing (as in IP)> FIB entries should be populated proactively for known content.> Alternatively, searching for content, e.g., using flooding


CCN Transport

> Stateless operation with receiver control> Reliability by local retransmissions in strategy layer> Hop-by-hop flow control> Sequence numbers in names



19

CCN Security

> Signing of names and data in each packet> Denial-of-Service attacks are difficult:

Multiple Interests → only 1 data packet per Interest



20


21

CCN Evaluation

> Advantages— Automatic content distribution— < 1 round-trip-time— Minimization of latency— Minimization of bandwidth— Local congestion control— Built-in security

> Challenges— Routing protocols— Hierarchical naming— Source mobility



22

Publish-Subscribe Internet (PSI) Naming

> Information items = files, streams, services> Each information item has its own name. > Names are unique (SID,RID) pairs

— Rendezvous Identifier (RID)– Fixed-length, (flat) bit string– produced by application

specific function, e.g., hashing

— Scope Identifier (SID)– Containers for information

items– Basis for access control

> Scope hierarchy with information belonging to different scopes.> www.fp7-pursuit.eu


SId1 SId2

SId1 SId1 SId2

SId3

RId1 RId2 RId3

RId1 RId2 RId3RId4

RId3

http://www.fp7-pursuit.eu/

PSI Network Primitives

> Subscribe— used to express interest in information items— Users can subscribe to information items or scopes.

(SID/RID) must be known. > Publish

— used to announce information items— sent if information items have been requested



23

PSI Operation

> Information producer publishes information item to rendezvous system consisting of rendezvous points.

> Rendezvous points are responsible for certain scopes.> Information consumer subscribes to information item.> Rendezvous system matches announcements and subscriptions

and triggers delivery from information producer to information consumer, e.g. using OpenFlow

> Various caching strategies: on-path, off-path, replication



24

publication

subscription

path establishment


25

Service-Centric Networking (SCN)

> ICN/CCN is content-centric and encodes a few operations on content as extensions of names.

> Proposal: Service-Centric Networking— Extension of content-centric networking to support services,

possibly operating on content. — Description of a service using content naming scheme,

e.g., /google.com/file-service— Service request (response) in Interest (Data) message

> Services— Infrastructure services, e.g., cloud computing services— Client-oriented services, e.g., web services— Continuous content retrieval and streaming services,

e.g., A/V conferencing, streaming— Event services, e.g., exceeding sensor data thresholds or stock data— In-network services, e.g., aggregation and filtering of sensor data

Braun, T.; Hilt, V.; Hofmann, M.; Rimac, I.; Steiner, M.; Varvello, M.: Service-Centric Networking, 2011 IEEE International Conference on Communications Workshops (ICC), pp.1-6, Kyoto, 5-9 June 2011Medellin, October 5, 2012


26

Service-Centric Network


Service request/response

SCN routerSCN network


27

Advantages of SCN

> No service lookup and service registry

> Caching of service data

> Extended caching of multimedia data (transcoding)

> Location-based services

> Optimized service selection



28

Uniform Naming for Services (Functions) and Content (Data)

> Services perform (data) processing and are represented by functions to be invoked. Content stores for data.

> Service-centric networking should support both data and functions.> Object-orientated programming paradigm integrates

both functions and data into objects. Method calls among objects to invoke functions.

> Proposal: Object names for both services (functions) and content (data), e.g., — /youtube.com/rendering — /unibe.ch/braun/lecture/20120405

> Advantages of object-oriented approach— Uniform naming— Services can be implemented as a set of cooperating objects



29

SCN Object Types

contentdata

read

1: Content Object 2: Service Object

contentdata

read

3: Content/Service Object

function3function1

function2

function1

function2



30

Example: Video Rendering


render

video-file1

read

client

video-file2

read

video-file3

read writewritewrite


31

Example: Real-time Audio Conferencing


Echo cancellation

Trans-coding

Sent audio data

Received audio data

Mixing


32

SCN Prototype Evaluation


user1

ccnd1

client1

user2

ccnd2

client2 ccndrconversion service & repository

CCN router

ccndp

publisher

1/8

2/711/12

3

5

6

4

13/14

18/19

10

159

16

17/20

1-8 9-16 17-200

20406080

100120140160180200

186 31 4

time [s]

1-8: Client 1 retrieves bit map image (40 MB)9-16: Client 2 retrieves JPEG image17-20: Client 1 retrieves JPEG image


33

ICN and Cloud Computing

> ICN might be used to— find/call cloud services (computation, storage)— deploy cloud services— support caching of cloud data



34

Upcoming EU FP7 Integrated Project: Mobile Cloud Network

> Integration of— Radio Access Network— Mobile Core Network— Data Centre (computing, storage, applications)

into (decentralized) cloud computing infrastructure (management!)> Advantages: elasticity, costs, performance, e.g., by

exploitation of user location for service provisioning → Follow Me Cloud



35

MobileCloudNetwork Architecture



36

Benefits of ICN in Wireless/Mobile Networks

> Broadcast property of wireless medium → single Interest to many neighbor nodes

> No beaconing required to learn about neighbor nodes> Reduced delay and network load due to possible caching> Request/Response based routing in mobile ad-hoc networks

(MANETs) could be replaced by Interest/Data exchange, cf. Directed Diffusion

> Support of short mobile nodes’ encounters> ICN architectures inherently support receiver mobility !



37

Source Mobility

> Problem— Finding a source’s location at the beginning of and when source is moving during

communication. > Solutions

— Routing-based approaches– (Proactive) updating of routing tables in case of moving sources

→ issues on scalability and convergence

— Indirection – (Home) agent forwards all Content Requests to mobile source, cf. Mobile IP.– This requires location-based identifiers, cf. location/identity separation.

— Resolution– Receiver requests location-based identifier for a content name specified in a Content Request. – Approach adds resolution phase and requires resolution entities in visited networks.

> Conclusions and Suggestions— Both content names and locators might be needed to efficiently support ICN source

mobility. — Flexible use of content names / locators — Possibly late binding of content names and locators

Vasilios Siris:, Torsten Braun: Content-Centric Networking Architectures for Moving Objects, COST Action IC0906, STSM Report, 2012, http://cost-winemo.org/ Medellin, October 5, 2012

http://cost-winemo.org/

http://cost-winemo.org/


38

CCN in Delay/Disruption-Tolerant (DTN) and Opportunistic Networks

Name Origin of Interest

Data

/feeds.sf.tv/podcast/10vor10 me -

/youtube.com/movies/ DesperateHousewives

mydaughter -

… … …

/unibe.ch/braun/lecture/ OperatingSystems

me,mystudent

os.mpg

… … …


Encounter

Possible message exchange at encounter1. Exchange of interests, possibly with priority of own / friends’ interests2. Exchange of relevant data


39

Content-Centric Networks and DTN/Opportunistic Networks

> Content-Centric Networks and Delay/Disruption-Tolerant Networks seem to be a good match. — No needs for beacons and neighbour lists— Exploitation of wireless broadcast media— Efficient exchange of Interests and Data

> Publish / Subscribe approaches might be less beneficial here.



40

Network Coding and CCN

> Regular case: — 4 Interests + 4 Data

messages> Network Coding:

— 3 Interests + 3 Data messages

> Problem: Naming> Proposal: Encapsulation

Alexander Striffeler: Network Coding in mobile CCN, Bachelor Thesis at Universität Bern


Interest A

Interest B

Interest A B

Data A

Data B

Data A B


41

Network Coding of CCN Data Messages



42

Energy Efficiency and CCN

> ICN can reduce packet transmissions on links> This comes at the expense of additional memory needed in

caches. > Assumptions for power usage

— DRAM: 3 W/GB — Hard disk: 0.02 W/GB— SSD: 0.01 W/GB— Energy for transmission: 15 W/Gbps = 15 Joules/Gigabit

> Analytical evaluation in example scenarios, options— no CCN support at all — CCN support in CCN router

Tuan Anh Trinh, Torsten Braun: Energy Efficiency in Information-Centric Networking, COST Action IC0804, STSM Report, 2012, http://www.irit.fr/cost804/


http://www.irit.fr/cost804/


43

Example Scenario



44

Energy Efficiency and CCN

> Required energy for serving K requests from original content source: Eread_from_source = K * N * link_energy * datasize.

> Required energy for serving request from cache:Eread_from_cache = K * M * link_energy * datasize + lifetime_in_cache * storage_power * datasize

> ICN can help to decrease energy costs, if Eread_from_cache < Eread_from_source

→ (N - M) * link_energy / storage_power > lifetime_in_cache / K— lifetime_in_cache: lifetime a content object is usually stored in the cache— storage_power: power required to store a certain amount of data— link_energy: energy to transport a certain amount of data across a link

> lifetime_in_cache / K = inter_access_time (avg. time between two content accesses)> access frequency f = 1/inter_access_time→ f > storage_power / ((N-M) * link_energy)

> Example— Link energy = 15 Joules / Gigabit = 120 Joules / GB— Storage power = 3 W/GB for storage in DRAM— M = 1, N = 11— ICN only reduces energy consumption if f > (3 W / GB) / ((11-1) * (120 Joules / GB)) = 1 / 400 sMedellin, October 5, 2012


45

Conclusions

> Information-Centric Networking attracts huge research interest.

> ICN has potential to — save bandwidth— reduce delay— save energy— support receiver mobility— operate efficiently in wireless networks— support DTN and opportunistic networks

> Challenges— storage in routers— routing strategies— source mobility— more general services



46

Thank You for Your Attention !

> [email protected] > http://cds.unibe.ch> http://www.slideshare.net/torstenbraun


mailto:[email protected]

http://cds.unibe.ch/

http://www.slideshare.net/torstenbraun

Information-Centric Networking in Wireless/Mobile Networks

Documents

named data networking

information retrievalrather

information base

hashing content

netinf medellin

ipbased networks

data transport decoupled

selector data