the challenge of the distributed control · the challenge of the distributed control infrastructure...
TRANSCRIPT
The Challenge of theThe Challenge of theDistributed ControlDistributed Control
Workshop of the Workshop of the Future Media InternetFuture Media Internet
DraDra. . Carmen GuerreroCarmen [email protected]@uc3m.es
CONTENT CONTENT NoENoEUniversity Carlos III MadridUniversity Carlos III Madrid
Brussels, 22Brussels, 22ndnd January 2008 January 2008
NETWORKED MEDIA LONG TERM RESEARCH
TASK FORCE
Outline
CONTENT NoEThe Challenge of Distributed Control Infrastructure and Future Media InternetConclusions
CONTENT CONTENT NoENoE
CONTENT NETWORKS AND SERVICES FOR HOME USERSCONTENT NETWORKS AND SERVICES FOR HOME USERS
NoE IST-2006-38423Networked Audiovisual SystemsStart date: 1/07/2006 End date: 30/06/2009Project duration: 36 monthsTotal budget: 2,648,000 €11 partners from 10 countrieswww.ist-content.eu
IAB Board: Philips Research, Eindhoven, NL - Agilent Laboratories, UK - Danet GmgH, DE -Thales Broadcast & Multimedia, FR - Ericcson Research, IR - Telefónica, Publicidad e Informacion, ES - TECHMATH – Blue – Order, UK - Alcatel-Lucent, BE - Nokia Research, FI -Telefonica Research Lab Barcelona, ES - Google Switzerland GmbH, CH - TANDBERG ASA, NO
Why CDNs?
Technical development:
Digital representation is dominantAffordable devices allow direct production
MPEG or DV basedFor broadcasting, movie theatres, domestic use, etc.Not just by professionals.
There is an explosion of digital multimedia content
Why CDNs? (cont.)
Content Distribution Network (CDN) is an infrastructure designed to efficiently deliver content to the users
E.g. broadcast networks for television
Role of IP networks as a basic communication platform for content distribution is becoming more important
More digital multimedia contentBetter access technology (xDSL, cable, etc)
Content production Content consumption
Separation in space and/or time
1 : N
Generations of CDNs
1st generation CDNs (today)Mostly content presentation
Simple video streamingFairly static structure, manual content placement, etc
Better computing, networking, compression and synchronization technologies allow new applications Users can then select media elements and control applicationNeed for 2nd generation of CDNs – Future
Media Internet
Media Internet for Communities# Users
# Assets
From 1st tier providers
Opportunity for innovative media internet services
We seek to integrate our research expertise on the different levels of (overlay) networks and services for AV content and put special emphasis on the new challenges that arise from community networks.
Content Networks for Communities
HostBackbone
AccessConsumer
HostBackboneAccess
Consumer
ContentHost
BackboneAccess
ConsumerCommunity
Past Present Future
CONTENT Research Framework
TA2:
Ove
rlay
Net
wor
ks
Service Management
File download VoD Streaming Live Streaming
Chunk Management
Replication Management
Incentive System
Peer Selection
Search, Lookup
Trust & Reputation
Accounting & Charging
Content Protection
P2P
Pla
ttfor
ms
End-to-End QoS
QoS in WLAN and Internet
Mobility and Nomadicity
TA1: CommunityNetworking
TA3: Content
Networks
Quality of Experience
CONTENT Network Architecture and Service Framework
Benchm
arking
Monitoring &
Measurem
ents
EvaluationTools
Outline
CONTENT NoEThe Challenge of Distributed Control Infrastructure and Future Media InternetConclusions
Expected Networked Media of the Future
Built upon three axes:True broadbandPersonalized mediaDistributed control
Developments on these axes will lead:New models of rich media interactionsEnhanced AV content distribution at community levelPervasive personalized media
The Challenge of the DistributedControl Infrastructure
Research question : what is the challenge of distributed control. (CONTENT NoEquestion)The infrastructures and services which are based on the composition and interaction of many subsystems and entities, relaying in self-organization and self-interaction technologies
Key research directionsInfrastructure
Networked media infrastructureContent service infrastructure
ContentInteractive, scalable, multi-model content searchContent summarizationContent adaptation, aggregation and digestionNew paradigms of interactions with abstract entitiesMetadata and indexing
Today Media Internet
social networks knowledge sharing
personalized browsers, content sharingsearch engines
Today Media Internet (ii)And much more almost every day…
Evolution of Internet Traffic
1993 – academic traffic1995 – web in scene2000 – ~ 75% web traffic and data content2004 – 70% P2P traffic and media content
emailftp
other
Inside P2P
Internet Study 2007 in Germany (similar in rest of the world)www.ipoque.com
Internet Traffic Today (ii)
30%-70% traffic P2P~ half Bittorrent~ 70% media contentImplications:
Economics of ISPsImpact on traffic engineering and routingInteraction between applications and networkNext generation of CDNs
Client-Server Economics
Transit ISP Transit ISP
INTERNET
Access ISP Access ISPAccess ISP
€€ €€ €€
P2P Economics
Transit ISP Transit ISP
INTERNET
Access ISP Access ISPAccess ISP
€€€€ €€ €€
Filtering out P2P trafficBut P2P increasesaccess demands (€€)
Add cachesEnrich connectivity
Peering betweenaccess ISPs
P2P Economics
Transit ISP Transit ISP
INTERNET
Access ISP Access ISPAccess ISP
€€
Solutions:
That means:FlatteningInternetImpact on thedesign of Internet
Storedcontent
Livecontent
1st generation: main issues streaming and presentation→ static content, static management
Evolution to Future Media Internet
Today
Dynamiccontent
Future
• More interactivitywith the content
• Combining and customizingof content
• Users push contentinto the CDN
• Editing of content• Harder management
tasks
Requirements on next generation(s): • efficient read access, navigation, search and retrieval
• efficient support for distr. customization, composition andediting
• efficient management
The Need for FlexibilityFuture Media Internet:
No more static specialized infrastructures, but …Flexible, dynamic, multi-service content-centered infrastructures, supporting:
The concept of dynamic user communitiesThe entire life cycle of each piece of content, and of combinations of themMobility of usersA variety of networking technologies
Emerging paradigms well fitting these needs:P2P and OverlaysSelf-describing, composable servicesSelf-describing, composable pieces of content (metadata)
Services to add metadata,combine content elements
dynamic,automatic, selforganizingmanagement of resources
Traditional CDN forVoD & WWW
Content deliveryCachingReplicationPrefetchingMulticastOverlay routing
CDN managementContent management
From CDNs to Future Internet Media
Internet infrastructureFlattened StructureBig increase in traffic volume and number of terminalsExplosion of heterogeneous hand-held and wearable devicesWireless Access: Future Wireless InternetEnd-to-end paradigm: consider limited-capabilities of hand-held terminalsMobility (not just nomadicity) is a mustDominance of P2P and overlay over client-serverQoS and multicast issues to deal with…
Challenges of Future Media Internet
Challenges of the Future Media Internet
Internet is sociologicaly flatEmpowering peopleEqualizing experimental opportunities
Technologicaly flatHierarchical structure
Flat structureNetwork resource management
End-host/appilication resource managementComplexity
Simplicity
Challenges of Future Media Internet
The three challenges are not independent, and rather reinforce each other
DistributedControl
IncreasedBandwidth
PersonalizedMedia
Challenges of Future Media Internet
The three challenges imply very fundamental techno-socio-economic alterations
Anyone can broadcast content,
merge content, produce content, etc
Anyone can receive and deliverhigh quality media
Anyone receiveswhat he wants and NOT
what the broadcaster wants
Challenges of Future Media Internet
Technologies for ContentDistribution
Networks
Technologies forInterconnected
BroadbandAccess Networks
Technologies for highly automatic
Personalization
Overall Conclusions are that in 5 to 10 years we should see:Fundamental reshaping of the mass media business structure Fundamental reshaping of the Internet
Emerging phenomenon: Future Media InternetFuture Media Internet
Challenges of Future Media Internet
Technologies for ContentDistribution
Networks
Technologies forInterconnected
BroadbandAccess Networks
Technologies for highly automatic
Personalization
Technologies forMedia Wireless
Devices
Technologies forMedia Adaptation
Technologies toScale InternetIn traffic and
number of terminals
Overall Conclusions are that in 5 to 10 years we should see:Fundamental reshaping of the mass media business structure Fundamental reshaping of the Internet
Emerging phenomenon: Future Media InternetFuture Media Internet
Other technologiesfor Emerging uses
The Challenge of theThe Challenge of theDistributed ControlDistributed Control
Workshop of the Workshop of the Future Media InternetFuture Media Internet
DraDra. . Carmen GuerreroCarmen [email protected]@uc3m.es
CONTENT CONTENT NoENoEUniversity Carlos III MadridUniversity Carlos III Madrid
Brussels, 22Brussels, 22ndnd January 2008 January 2008
NETWORKED MEDIA LONG TERM RESEARCH
TASK FORCE