S CALABLE V IDEO C ODING IN C ONTENT -A WARE N ETWORKS Michael Grafl - PowerPoint PPT Presentation

S CALABLE V IDEO C ODING IN C ONTENT -A WARE N ETWORKS Michael Grafl Institute of Information Technology Alpen-Adria Universität Klagenfurt, Austria COMET-ENVISION workshop on Future Media Distribution Networks Nov. 10-11 2011, Slough, UK Michael Grafl Scalable Video Coding in Content-Aware Networks 1

O UTLINE  Introduction  ALICANTE  Project Overview  Conceptual Architecture  Comparison to ICN  Use Cases for SVC in CAN (Analysis wrt. ICN research challenges)  Unicast  Multicast  P2P Streaming  Web/HTTP Streaming  Step-by-Step Walkthrough  Conclusions Michael Grafl Scalable Video Coding in Content-Aware Networks 2

I NTRODUCTION  Information-Centric Networking (ICN)  Revolutionary approach  Content-Aware Networking (CAN)  Evolutionary approach  ALICANTE project  Scalable Video Coding (SVC)  Extension of H.264/MPEG-4 AVC  Spatial, temporal and quality (SNR) scalability  Base layer + multiple enhancement layers  Coding overhead: ~ 10% wrt. H.264 Michael Grafl Scalable Video Coding in Content-Aware Networks 3

ALICANTE  Project Info:  EU FP7-ICT project  Duration: March 2010 – Aug 2013  20 partners  "Medi a Ecosystem Dep l oyment through Ubiqu i tous C ontent- A ware N e t work E nvironments"  Goal: New Home-Box layer and CAN layer with distributed cross-layer adaptation and universal multimedia access enabling cooperation between providers, operators, and end-users http://ict-alicante.eu Michael Grafl Scalable Video Coding in Content-Aware Networks 4

ALICANTE C ONCEPTUAL A RCHITECTURE User Environment: devices and end user User Environment Context- aware Service Environment: content and services Service Environment HB HB HB HB HB Layer: networked Home-Box Layer Content- components as overlay HB Network- aware aware MANE MANE CAN Layer: in-network CAN CAN CAN components as overlay MANE MANE CAN CAN CAN Physical Layer: AS AS AS autonomous systems Network Environment Michael Grafl Scalable Video Coding in Content-Aware Networks 5

ALICANTE VS . F ULL ICN A PPROACH  Evolutionary approach for FI (Mid-way to full ICN) Approaches:  Caching and storage  In Home-Boxes (network edge)  Best effort and Content Servers  QoS-based virtual splitting  Scalable and Cost-Efficient  Content-aware networks Content Distribution  Content-type awareness ALICANTE  Name/location resolution –  Service-aware networking approach at Service level (not in routers)  Full ICN  Content-awareness  Content/object awareness  Name/location resolution,  aggregated CA and associated routing of requests, processing at network level caching at network nodes  Deployment Degree of awareness on upper  Seamless/incremental deployment layer information at network level Michael Grafl Scalable Video Coding in Content-Aware Networks 6

U SE C ASES FOR SVC IN CAN  Role of scalable media formats for enabling content-aware networking  Unicast, Multicast, Peer-to-Peer Streaming, Web/HTTP Streaming  Analysis wrt. ICN research challenges:  Routing & Forwarding  Caching & Buffering  Quality of Service/Experience (QoS/QoE) Michael Grafl Scalable Video Coding in Content-Aware Networks 7

S YSTEM O VERVIEW FOR U SE C ASES S1 R1 Mobile U1 ICNN2 ICNN1 R2 U2 HD-Ready TV Buffer Buffer S2 R3 SVC-Base Layer U3 Full-HD TV Enhancement Layer 1 Enhancement Layer 2 Michael Grafl Scalable Video Coding in Content-Aware Networks 8

U SE C ASES : U NICAST  Example: Video on Demand (VoD)  RTP (with SST of SVC) and RTSP  Routing & Forwarding:  ICN node can react to network fluctuations • In-network adaptation of SVC at ICN node (for short-term fluctuations) • Signal to sender for dropping SVC layers (for long-term fluctuations)  Caching & Buffering:  ICN node can perform prefix caching • Reduce start-up delay • Selective caching of SVC layers  QoS/QoE: (applies to all use cases)  Consider terminal capabilities when requesting SVC layers  Monitor network conditions at ICN nodes (cf. ALICANTE)  Smooth, undistorted playout Michael Grafl Scalable Video Coding in Content-Aware Networks 9

U SE C ASES : M ULTICAST  Receiver-Driven Layered Multicast (RDLM) of SVC  RTP in MST mode (each SVC layer in own session)  Routing & Forwarding:  ICN nodes adapt to network conditions through subscription to SVC layers  ICN nodes as bridges between native and overlay multicast (ALICANTE: virtual content-aware network of ICN nodes)  Selective treatment of SVC layers (MPLS, DiffServ)  Caching & Buffering:  Prefix caching to reduce start-up delay in non-live scenarios Michael Grafl Scalable Video Coding in Content-Aware Networks 10

U SE C ASES : P2P S TREAMING  Receivers request pieces from multiple senders  P2P network as overlay  Receiver only requests SVC layers supported by end-user terminal  Routing & Forwarding:  ICN nodes can act as peers, forming an in-network overlay  Caching & Buffering:  Aggregate requests and perform information-centric buffering (during sliding window) at ICN nodes Michael Grafl Scalable Video Coding in Content-Aware Networks 11

U SE C ASES : W EB /HTTP S TREAMING  Download via HTTP (partial) GET requests  Content fragmented into segments (e.g., per SVC layer and GOP)  Manifest file describes structure of segments and available representations  Standard: Dynamic Adaptive Streaming over HTTP (DASH)  Overcome NAT traversal & firewall issues  Stateless sender  Unicast, multicast, and multisource (P2P-like) scenarios  Routing & Forwarding:  ICN node signals network condition to receiver (  implicit adaptation)  Caching & Buffering:  SVC-based prefix caching using HTTP-based CDN infrastructure  Buffering during sliding window creates multicast tree  Information-centric buffering in multisource scenario Michael Grafl Scalable Video Coding in Content-Aware Networks 12

S TEP - BY -S TEP W ALKTHROUGH 8 AVC decoding AVC decoding 6 Source 1 7 9 5 HB SVC to AVC transcoding HB HB SVC 2 encoding SVC to AVC 3 4 transcoding CAN MANE MANE 2 nd SVC Adaptation 1 st SVC Adaptation at MANE at MANE HB Source Stream Base Layer (AVC) Enhancement Layer 1 Enhancement Layer 2 Michael Grafl Scalable Video Coding in Content-Aware Networks 13

C ONCLUSIONS  Towards ICN: Scalable media coding formats (e.g., SVC) in combination with in-network adaptation  Routing & Forwarding  Caching & Buffering  QoS/QoE  Enabling content-awareness within the (core) network  Context-awareness at receiver & sender (& ICN node)  ALICANTE  Towards deployment of a networked "Media Ecosystem"  Collaboration of CAN layer and Home-Box layer Michael Grafl Scalable Video Coding in Content-Aware Networks 14

L ITERATURE [1] J . Pan, S. Paul, R. Jain, “A survey of the research on future internet architectures”, IEEE Communications Magazine , vol.49, no.7, pp.26-36, July 2011. [2] V. Jacobson, D. Smetters, J. Thornton, M. Plass, N. Briggs, R. Braynard, “Networking named content”, Proc. of ACM CoNEXT 2009 , Rome, Italy, December 2009. [3] H. Koumaras et al., “Media Ecosystems: A Novel Approach for Content - Awareness in Future Networks,” Future Internet: Achievements and Promising Technology , Springer Verlag, pp. 369-380, May 2011. [4] ALICANTE Web site, http://ict-alicante.eu/. [5] M . Wien et al., “Performance Analysis of SVC,” Circuits and Systems for Video Technology, IEEE Transactions on , vol. 17, no. 9, pp. 1194-1203, 2007. [6] T. Stockhammer , “Dynamic adaptive streaming over HTTP – standards and design principles,” in Proceedings of the Second Annual ACM Conference on Multimedia Systems , New York, NY, USA, pp. 133 – 144, February 2011. [7] M . Grafl, et al., “Scalable Video Coding in Content -Aware Networks: Research Challenges and Open Issues,” in: N. Blefari-Melazzi, G. Bianchi, and L. Salgarelli (eds.), Trustworthy Internet , Springer, pp. 349-358, June 2011. Michael Grafl Scalable Video Coding in Content-Aware Networks 15

T HANK YOU FOR YOUR A TTENTION ! Questions? Michael Grafl Scalable Video Coding in Content-Aware Networks 16