Evaluation of Live Streaming Multicast over Dynamic Overlay in Urban VANETs Yi-Ling Hsieh MBL, NCTU 2011/04/27 Outline Introduction Related works Model of dynamic overlay Evaluation and discussion Conclusion Future work References Introduction Emerging of vehicular communication and application Standards such as DSRC, WAVE, 802.11p Vehicular safety Infotainment Charateristics of VANETs High mobility (and result in high packet loss) Vehicles move in pre-defined route (map) Literatures for multimedia distribution in VANETs Streaming QoS in vehicular environments Multicast Streaming over P2P and overlay networks Introduction Scenarios of vehicular environments Free space Highway (or straight road) Urban (or street) (not yet further studied in literatures of multimedia distribution) Urban environment Many road intersections and obstacles Packet loss rate is much higher due to routing is easily broken and rebuild Introduction Our focus scenario Multicast overlay Overlay multicast vs. network-layer multicast: overlay multicast offers the flexibility of node capability for the protocol Interested vehicles (called group members) form a multicast tree The multicast source node is the tree’s root node The group members are organized as an overlay Since routing is easily broken due to urban environment, organize a small group of nodes by overlay is more flexible Related works Overlay multicast in MANETs Static overlay or dynamic overlay Tree or mesh Low speed (1~10m/s, 3.6~36km/hr) Need further study if applied in VANETs (5~25m/s, 18~90km/hr) Related works Existing literature of multimedia distribution in VANETs: CodeTorrent  and NCDD  They utilized the random network coding technique for data dissemination in VANETs However, it requires all nodes periodically broadcast pieces information and may consume much time to collect enough pieces to decode if the multicast group size is small Related works SMUG Every node may dynamically be selected as a forwarder, and its transmissions are scheduled by color-filtering rule However, if the SMUG-capable nodes are not dense enough, the forwarding paths are easily broken. Besides, SMUG can only be applied in TDMA-based ad hoc networks. Related works Multimedia data delivery through an overlay over VANETs N.N. Qadri et al.  proposed a series of evaluation papers impact of radio models and mobility models over overlay networking in VANETs improvement by video error resilience techniques such as MDC and FMO. the overlay topology was fixed, even though nodes are mobile. And only 7 nodes. Related works Out contributions Evaluate the improvement of using “dynamic” overlay in high speed, easily routing broken VANETs (firstly studied) The impact of considering obstacle in simulation for urban VANETs (firstly studied) Evaluate the improvement of using multiple parents mesh Model of dynamic overlay in urban VANETs Model of the overlay structure Tree 1 child has only 1 parent If the routing fail, the child and all offspring can not receive any streaming improvement: multiple parents (as mesh) Mesh 1 child has 2 parents Pa: current parent Pb: last parent Model of dynamic overlay If QoS of parent is low, or route fail between the parent, change to a new parent Probe for a better QoS and less-distant parent Model of dynamic overlay Probe procedure We use Continuouty Index, CI, as QoS value: ratio of streaming packets received Probe for less-distant nodes Model of dynamic overlay Probe procedure For mesh (1 child has 2 parents) Current parent become last parent (Pb) New parent become current parent (Pa) Evaluation and discussion Simulator: Qualnet 5.0 Transmission power:19.3 dBm (~625m) Receiver power:58.4 dBm Node movment trace generator: VanetMobiSim 1.0 Node speed: [5 m/s, 25 m/s] IDM-LC mobility (intelligent driving model - lane changing) model Map: Grid 1000mx1000m, section 100m Underlying routing protocol: Location Aware Routing (LAR) Packet generation: CBR as 64kbps, 128kbps, 256kbps Movement table exchange interval: 5s Probe interval: 10s Evaluation and discussion 0.95 d0b0 0.85 d0b1 0.75 CI d0b1_2p 0.65 d1b1_1p 0.55 d1b1_2p 0.45 5 10 15 20 Number of group nodes d: dynamic Fig 1. CI under 64kbps b: blocking obstacles simulated 2p: 2 parents mesh 1 0.8 d1b1_1p@64kbps Evaluation and discussion 0.6 0.4 d1b1_1p@128kbps d1b1_1p@256kbps 0.2 0 5 10 15 20 1 0.8 d1b1_2p@64kbps 0.6 d1b1_2@128kbps 0.4 0.2 d1b1_2p@256kbps 0 5 10 15 20 d: dynamic Fig 2. CI comparison b: blocking obstacles simulated among 64,128,256 kbps 2p: 2 parents mesh 1.8 1.6 1.4 d0b0 1.2 d0b1 Delay (s) 1 d0b1_2p 0.8 0.6 d1b1_1p 0.4 d1b1_2p 0.2 0 5 10 15 20 Number of group nodes Fig 3. Total delay under 64kbps d: dynamic b: blocking obstacles simulated 2p: 2 parents mesh Evaluation and discussion 1 0.9 0.8 64kbps 0.7 CI 128kbps 0.6 128kbps(RD) 0.5 0.4 0.3 100 200 400 Road section interval RD (density on road): Sec 100:100 nodes Sec 200:54 nodes Sec 400:21 nodes Conclusion We firstly studied the improvement of using “dynamic” overlay in high speed, easily routing broken VANETs, and the impact of considering obstacle in simulation for urban VANETs We also evaluate the improvement of using multiple parents mesh References 1. U. Lee, J.-S. Park and M. Gerla, “CodeTorrent: a content distribution using network coding in VANET,” in Proc. of First ACM Workshop on Decentralized Resource Sharing in Mobile Computing and Networking, pp. 1-5, 2006. 2. Joon-Sang Park et al. , "Emergency related video streaming in VANET using network coding," in Proc. of the 3rd International Workshop on Vehicular Ad Hoc Networks, ACM, pp. 102-103, 2006. 3. F. Soldo, C. Casetti, C.-F. Chiasserini, and P. Chaparro, “Streaming media distribution in VANETs,” in Proc. of IEEE GLOBECOM, 2008. 4. N.N. Qadri, M. Fleury, M. Altaf, M. Ghanbari, "Multi-source video streaming in a wireless vehicular ad hoc network," IET Communications, vol. 4, pp. 1300-1311, 2010.
Pages to are hidden for
"Evaluation of Live Streaming Multicast over Dynamic Overlay in "Please download to view full document