DEPT. OF INFO. & COMM., GIST Page 1 Adaptive High-quality Video Service for Adaptive High-quality Video Service for Network-based Multi-party Network-based Multi-party Collaboration Collaboration 19 th Oct 2005 Sangwoo Han [email protected] Networked Media Lab., Dept. of Information an d Communications Gwangju Institute of Science and Technology (GIST), Korea Graduate Workshop on Networking
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
DEPT. OF INFO. & COMM., GIST
Page 1
Adaptive High-quality Video Adaptive High-quality Video Service for Network-based Service for Network-based Multi-party CollaborationMulti-party Collaboration
19th Oct 2005Sangwoo Han
Networked Media Lab., Dept. of Information and CommunicationsGwangju Institute of Science and Technology (GIST), Korea
Graduate Workshop on Networking
DEPT. OF INFO. & COMM., GIST
Page 2
ContentsContents
• Introduction
• Background
• AG Media architecture
• Evaluation
• Conclusion
• Q & A
DEPT. OF INFO. & COMM., GIST
Page 3
IntroductionIntroduction
• Terminologies– Advanced Collaborative Environments (ACE)
• Bring together the right people and the right data at the right time to perform a task, solve a problem, or simply discuss something of common interest.
– Quality of Experience (QoE)• A factor to evaluate satisfaction of the users in the context of the collaborat
ion tasks.
• Motivations– How to improve QoE in ACE?
• High-quality video support
– How to remove barriers preventing high-quality video support, such as network variation and heterogeneity problem?
• Network monitoring and adaptation
DEPT. OF INFO. & COMM., GIST
Page 4
Target Application: Target Application: Access GridAccess Grid
• What is Access Grid (AG)? (from the Access Grid web site)– An ensemble of resources including multimedia large-format displays,
presentation and interactive environments, and interfaces to Grid middleware and to visualization environments.
– These resources are used to support group-to-group interactions.
– The Access Grid thus differs from desktop-to-desktop tools that focus on individual communication.
• Features– Multiple video/audio streams and unlimited number of participants (in theory)
– Open source software
– Centralized and (mostly) public meeting “locations”
– Uses multicast-enabled networks
– Shared applications, data, and services
DEPT. OF INFO. & COMM., GIST
Page 5
Access Grid in ActionAccess Grid in Action
Smart AG nodeAG Node architecture
NodeService
ServiceManager
ServiceManager
ServiceManager
VideoConsumer
Service
AudioService
Display AudioVideo
VideoProducerService
VideoProducerService
VideoProducerService
VideoProducerService
DEPT. OF INFO. & COMM., GIST
Page 6
Design of AG Media Design of AG Media ArchitectureArchitecture
• Design consideration– Service-capability negotiation
• defines a protocol of exchanging service capabilities between participants.
– Application-layer QoS control• resolves network problem points by rate control
• Main components– AG Media Arbitrator
• manages AG media interfaces, negotiates service capabilities between AG media arbitrators, and achieves QoS control
– AG Media Interface• accesses to the video resources provided by adaptive video applications on
a local machine
DEPT. OF INFO. & COMM., GIST
Page 7
AG Media ArchitectureAG Media Architecture
Access Grid Node 1
AG Media Interface
Video producer
AG Media Interface
Video consumer
Access Grid Node 2
Access Grid Node 3
Service capabilitiesAvailable service 1: HDV
Available service 2: DV
Service capabilitiesAvailable service 2: DVAvailable service 3: Stereo HD
Service capabilitiesAvailable service 1: HDV
Available service 2: DVAvailable service 3: Stereo HD
Capability negotiation data
Operation control command
Streaming video data
Multicast monitor
Node monitor
AG Media Interface
Video consumer
AG Media Interface
Video consumer
AG Media Interface
Video consumerAG Media Interface
Video consumer
AG Media Interface
Video consumer
AG Media Interface
Video consumer
AG Media Interface
Video consumerAG Media Interface
Video producer
AG Media Interface
Video producerAG Multicast Networks
AG Media Interface
Video consumer
AG MediaArbitrator
AG MediaArbitrator
AG Media Interface
Video producer
AG MediaArbitrator
AG Venue Server
Service capability negotiationMulticast monitoringApplication level adaptation
Congestion
Low performance
Service capability negotiationMulticast monitoring
Service capability negotiationMulticast monitoring
DEPT. OF INFO. & COMM., GIST
Page 8
AG Media Arbitrator and AG Media Arbitrator and InterfaceInterface
CapabilityNegotiation
QoS Control
Capability Negotiation
Multicast Monitor
AdaptationManager
ServiceController
SessionManager
QoS Control
AG Media ArbitratorAG Media Interface
Event H
andlerE
vent Generator
Event H
andlerE
vent Generator
NodeMonitor
ServiceDescriptor
ApplicationController
ServiceTable
AddressGenerator
ExternalNode Monitor
VideoApplication 1
VideoApplication 2
VideoApplication 3
AG
Media A
rbitrators
Operation C
ontrol
Capability
Negotiation
VenueServer
Potential C
onfiguration
DEPT. OF INFO. & COMM., GIST
Page 9
Adaptation Manager of AG Adaptation Manager of AG Media ArbitratorMedia Arbitrator
Multicast Monitors Node Monitors
Multicast Performance Table
System Performance Table
Problem Detector
Database
Network Network and System System
Service Requirements
Enabled Service List
ProblemDetection
AdaptationMethods
Event Generator
AG
Media Interfaces
RateControl
ErrorControl
Rematching Media Type
DEPT. OF INFO. & COMM., GIST
Page 10
ImplementationImplementation
• Service-capability negotiation– Session description by SDP (session description protocol)
– Session announcement by SAP (session announcement protocol)
• Versatile video support– DV and its application
• 720x480 DV-encoded 30Mbps digital video by using DVTS
– HDV and its application• 1280x720 MPEG2-encoded 19.2Mbps high-definition digital video by
using VLC
DEPT. OF INFO. & COMM., GIST
Page 11
Implementation (cont’)Implementation (cont’)
• Application-layer QoS control– Principle of one-to-one performance evaluation
• If packet loss rate of each receiver is greater than a pre-defined loss threshold, the receiver suffers quality deterioration. By this simple principle, every receiver is classified into two nodes having overload or proper load.
– Guide to rate control considering one-to-many video distribution• If the proportion overloaded nodes to total nodes is not less then the pre-de
fined threshold THmax, frame rate is reduces. If the proportion of loaded nodes to total nodes is not less than the pre-defined threshold THmin, the frame rate is increased.
– Execution of adaptive transmission• The adaptation manager allows adapative video applications to regulate dr
op rate – 0%, 33%, and 50%, and the adaptive video applications control frame rate by frame dropping.
DEPT. OF INFO. & COMM., GIST
Page 12
Sequence diagram of Sequence diagram of network adaptationnetwork adaptation
AG MediaInterface
video producer
AG MediaArbitrator
Shared application
AG MediaArbitrator
Shared application
AG MediaInterface
Video consumer
AG MediaArbitrator
Shared application
AG MediaInterface
Video consumer
Register Register Register
Session announcement Session announcement
Tx command
Rx command Rx commandEnd-to-end multicast network measurement
Measurement report
Detecting a node suffering the worst loss rate
Sender-size network adaptation considering the aspect of one-to-many video distribution
Drop rate command
AG Node #1 (sender-side) AG Node #2 (receiver-side) AG Node #3 (receiver-side)
Venue Server
Session announcement
Session description data
DEPT. OF INFO. & COMM., GIST
Page 13
Implementation resultsImplementation results
AG node enabling HDV serviceAG with AG Media
DEPT. OF INFO. & COMM., GIST
Page 14
Test bed for network-Test bed for network-adaptation experimentsadaptation experiments
DV sender
DV receiver 1
DV receiver 2
DEPT. OF INFO. & COMM., GIST
Page 15
Experiment Experiment environmentsenvironments
• Experiment scenario– No condition
– Network load
• Experiment parameters– loss threshold = 15%
– THmax = 0 and THmin = 100
• Test machine specifications– DV sender
• Desktop equipped with Intel 2.8GHz CPU and 512MB RAM
– DV receiver 1, 2• DELL D800 laptop equipped with Intel 1.7GHz mobile CPU and 512MB
RAM
DEPT. OF INFO. & COMM., GIST
Page 16
Experiment resultExperiment result
Network loadNo condition No condition
DEPT. OF INFO. & COMM., GIST
Page 17
ConclusionConclusion
• Conclusion– Designing AG Media Architecture to enabled DV and HDV support wi
th application QoS.
– Implementing the prototype associated with Access Grid.
– Verifying better users’ quality of experience by demonstration on the test bed.
– One-to-many network adaptation scheme conceals quality variation resulted from network problems.
DEPT. OF INFO. & COMM., GIST
Page 18
ReferencesReferences1. R. Stevens, M. E. Papka, and T. Disz, “Prototyping the workspaces of the future,” IEEE Internet Computing, pp. 51.58, 2003.2. L. Childers, T. Disz, R. Olson, M. E. Papka, R. Stevens, and T. Udeshi, “Access Grid: Immersive group-to-group collaborativ
e visualization,” Proc. of Immersive Projection Technology Workshop, 2000.3. B. Corri, S. Marsh, and S. Noel, “Towards quality of experience in advanced collaborative environments,” Proc. of the 3rd An
nual Workshop on Advanced Collaborative Environments, 2003.4. M. Handley, C. Perkins, and E. Whelan, “Session announcement protocol,” IETF RFC 2974, 2000.5. M. .Handley and V. Jacobson, “SDP: Session description protocol,” IETF RFC 2327, 2003.6. W. Zhu and N. Georganas, “JQOS: a QoS-based Internet videoconferencing system using the Java media framework (JMF),”
Proc. of Canadian Conference on Electrical and Computer Engineering, 2001.7. Z. Chen, S.-M. Tan, R. H. Campbell, and Y. Li, “Real time video and audio in the world wide web,” Proc. of 4th International
World Wide Web Conference, 1995.8. D. Kutscher, J. Ott, and C. Bormann, “Session Description and Capability Negotiation,” IETF MMUSIC Internet-Draft, 2003.9. X. Wang and H. Schulzrinne, “Comparison of adaptive internet multimedia applications,” IEICE Transactions on Communica
tions , pp. 806.818, 1999.10. A. Ogawa, K. Kobayashi, K. Sugiura, O. Nakamura, and J. Murai, “Design and implementation of DV based video over RT
P,” Proc. of Packet Video Workshop, 2000.11. J.-W. Park and S. Han and J.W. Kim, “End-to-end monitoring service for multicast-based high-quality real-time media deliver
y,” Proc. of 3rd IEEE/IFIP Workshop on End-to-End Monitoring Techniques and Services, pp. 142.151, 2005.12. I. Busse, B. Deffner, and H. Schulzrinne, “Dynamic QoS control of multimedia applications based on RTP,” Computer Com
munications , pp. 49.58, 1996.13. K. Ueda, H. Ohsaki, S. Shimojo, and H. Miyahara, “Design and implementation of real-time digital video streaming system o
ver IPv6 network using feedback control,” Proc. of Symposium on Applications and the Internet, pp. 111.119, 2003.14. S. McCanne, V. Jacobson, and M. Vetterli, “Receiver-driven layered multicast,” Proc. of ACMSIGCOMM, pp. 117.130, 199
6.
DEPT. OF INFO. & COMM., GIST
Page 19
Q & AQ & A
High-quality video service on AG is available from http://nm.gist.ac.kr/agdv
Related Documents
