Delay Tolerant Streaming Services, Thomas Plagemann, UiO
1 like754 views
This document discusses delay tolerant streaming services for transmitting live video from mobile devices over unstable mobile ad-hoc networks. It motivates the need for such services when conventional network infrastructure is unavailable. The approach involves building an adaptive overlay network on top of the mobile ad-hoc network to enable delayed and disrupted video streaming. Several technical challenges are outlined and initial results are highlighted from experiments and simulations evaluating the feasibility of video streaming over mobile ad-hoc networks formed by mobile phones. Future work is discussed around developing a prototype system and exploring fundamental changes needed to support emerging applications and technologies.
Delay Tolerant Streaming Services, Thomas Plagemann, UiO
- 1. Delay Tolerant Streaming Services Thomas Plagemann on behalf of the DT-Stream Team
- 2. Motivation Using head mounted Streaming live video camera Mobile Ad-Hoc Network 3
- 3. Motivation (cont.) • Communication to CCC using head mounted cameras • Networking infrastructure might not be available è use mobile phones to establish wireless ad-hoc network • Problem: mobile, unstable, partitioned network • Opportunity: video can also be useful when it arrives late 4
- 4. Motivation (cont.) This is a mock-up 5
- 5. History and Funding • Pre-project in 2008 at UiO – 4 Master students – Collaboration with University of Oviedo (Spain) • Verdikt funding: 3 PhDs & 1 PostDoc (2008 – 2012) • University of Oviedo funding: 1 PhD 6
- 6. Technical Challenges • Mobile wireless space – how to be always best connected? • Mobile phones have resource restrictions and therefore the network they create – Network topology awareness – Cross-layer optimization – Simulation tools not appropriate 7
- 7. The Mobile Wireless Space High “Relative Mobility” Space/Time Paths Hybrid Environments No (Space/Time) Paths Space Paths Low High Node Density Low 8
- 8. The Overall Approach • Adaptive Overlay for Delay Tolerant Streaming DT-S DT-S DT-S DT-S overlay 3 7 5 1 8 Mobile Ad-hoc 2 4 6 Network
- 9. A few Results Highlighted • Early results used to focus the research: – Survey of video streaming over MANETs – Real world experiments • Recent results: – Modeling mobile nodes in network simulators – Systematic cross-layer optimization – At home in heterogeneous networks – Non-intrusive network clustering 10
- 10. Video Streaming over MANETs • M. Lindeberg, S. Kristiansen, T. Plagemann, V. Goebel: “Challenges and techniques for video streaming over mobile ad hoc networks”, Multimedia Systems Journal, 2010 – Over 100 papers analyzed 11
- 11. Real World Experiments • Kristiansen, S., Lindeberg, M., Rodríguez-Fernández, D., Plagemann, T.: “On the Forwarding Capability of Mobile Handhelds for Video Streaming over MANETs”, ACM MobiHeld 2010 at ACM SIGCOMM 2010, August 2010 Monitor 2.2 GHz Intel Centrino Duo Core Bomb shelter 2 GB RAM M S R Sender Receiver 2.6 GHz Intel Centrino Duo Core 2.6 GHz Intel Centrino Duo core 3 GB RAM F 3 GB RAM Forwarder Nokia N900 Take away points: Mobile phones are a bottleneck Introduce non-neglect able delay 12 Severe at saturation point
- 12. Real World vs. Simulation Nokia N900 13
- 13. Towards realistic simulation Network Simulator Execution Model Traffi 3 c SrvA SrvB Execute Progress Processing Stages Scheduler Threads Update Simulator Program Update Schedule Service Mapping Shared 2 Model resources Payload in Scheduler Request … Simulator + Execution Obtain Execution Time Distribution from SEM and Resource Utilization State SEMA SEMB SEMC 1 1. Extract protocol models from existing devices 2. Map onto protocols in existing network simulators 3. Synchronize execution with threads in a scheduler simulator Traffic Generation and Tracing 14
- 14. Initial Results Model Accuracy (10 pps, ICMP Echo) 2 Intra-Node Delay (ms) 1.5 1 Real World Node Model Vanilla Ns-3 0.5 0 0 200 400 600 800 1000 1200 1400 15 Packet Size (Bytes)
- 15. Cross-layer Adaptation • Lindeberg, M., Kristiansen, S., Goebel, V., Plagemann, T.: “MAC Layer Support for Delay Tolerant Video Transport in Disruptive MANETs”, IFIP Networking 2011, Valencia, Spain, May 2011 <OverlayMessage> 500 m Dts-Overlay Dts-Overlay Sr - Source node Ix - Intermediate node(s) Sr I1 Crx- Carrier node(s) Route CCC - Command Control Center Rejected UDP availability UDP I2 packets + CCC Cr1 Cr2 + MAC address I3 500 m Retransmission IP / OLSR /ARP status IP / OLSR queue status I4 I5 MAC MAC Command and control center (IEEE 802.11 a/b) (IEEE 802.11 a/b) Location of the accident PHY PHY (IEEE 802.11 a/b) (IEEE 802.11 a/b) Distance=1750 m Node 1 Node 2 • MAC Support w/Cross-layer Interaction • Check ARP if IP address is known (ARP Adapt) • Check MAC transmission queue: if filling, link is down, stop using it (Link Adapt) • Return packets from MAC layer instead of dropping (MAC Return) 16 • We can reduce retransmission limit, and avoid most packet losses!
- 16. Cross-layer Adaptation (cont.) No “hard-wiring” to particular protocol implementations Standardized, convenient, and efficient access to information Using complex events 17
- 17. Multihoming in Heterogenous Network Paradigms • D. Rodriguez-Fernandez, I. Martinez-Yelmo, E. Munthe-Kaas, T. Plagemann: “Always Best (Dis-)Connected: Challenges to Interconnect Highly Heterogeneous Networks”, Special Issue of the Journal of Internet Engineering on Future Network Architectures, 2012 Community MANET DTN MANET Internet Internet Real World 3G MANET DTN 18 MANET
- 18. Multihoming in Heterogenous Network Paradigms (cont.) Applications Application Cross-layering Information Framework Community Socket API Community Framework Community Socket Community Overlays Monitoring Framework Community A Community B … Community Support Layer (CSL) NSAL API NSAL Internet MANET DTN … Underlays NSAL NSAL NSAL IP IP (MANET) DTN … Networking Substrata 19 3G 802.11 …
- 19. Non-intrusive Clustering of MANETS • Drugan, O., Munthe-Kaas, E., Plagemann, T.: “Detecting Communities in Sparse MANETs”, IEEE/ACM Transactions on Networking, 2011 Topology from OLSR routing table….. …. and with exact position… … and 100m range 20
- 20. Where are we now? • All PhD students scheduled to submit their thesis in 2012 • A good set of papers is published but several are in the queue • Most results are useful far beyond DT- Stream (cf. highlights) • Many of the core pieces are implemented for an integrated demo prototype (also with numerous MSc Theses) 21
- 21. Where do we go … • Look for funding to move basic research results of DT-Stream to applied research results, i.e., integrated demo prototype ++ • Head for further long term challenges … next slides 22
- 22. Future Technological Developments • Bigger, faster, higher resolution, more media, more Ds – Data centers, scientific computing, home entertainment • Smaller, everywhere, new range of I/O devices, energy concerns – Smart phones, sensors, actuators • Ever increasing heterogeneity in computing and networking • è diversity, separation, and seamless integration 23
- 23. Rethink Fundamentals • “there is a need to deeply rethink the modelling and architecting of future pervasive systems” M. Conti et al., CNR • “to fully realize the potential of CPS, the core abstractions of computing need to be rethought” A. Lee, Stanford University 24
- 24. Rethink Fundamentals of the Future Internet Assume the hotel gets old…. … where will stakeholders invest …upper floors renovations brings “immediate” turn on investment …but what if the foundations are rotten? …have we invested enough in the fundamentals of the Internet, IP, DNS, BGP? Hotel Internet 25
- 25. Technological Challenges • From cross-layer optimization to new foundations for engineering computer/network systems? – Layers simplify design and engineering – Layers simplify testing – Layers are in conflict with context aware solutions 26
- 26. Technological Challenges • Smart phones and wireless sensors and actuators: – Promise to solve many challenges society faces, e.g., sustainable environment etc., demographic change ++ – Sharing enables many new solutions, e.g., pervasive sensing vs. privacy and ownership – From application specific solutions to foundations that span several/all application domains? 27
- 27. Non-technical (non-trivial) Challenge • Acquire funding 28
- 28. Questions? 29