A Geographical Routing Protocol for Highly-Dynamic Aeronautical Networks Kevin Peters, Abdul Jabbar, Egemen K. Çetinkaya, James P.G. Sterbenz Department of Electrical Engineering & Computer Science Information Technology & Telecommunications Research Center The University of Kansas {kevjay jabbar ekc jpgs}@ittc.ku.edu http://wiki.ittc.ku.edu/resilinets 29 March 2011 2011 Sterbenz
AeroRP Analysis Abstract Emerging networked systems require domain-specific routing protocols to cope with the challenges faced by the aeronautical environment. We present a geographic routing protocol AeroRP for multihop routing in highly dynamic MANETs. The AeroRP algorithm uses velocity-based heuristics to deliver the packets to destinations in a multi-mach speed environment. Furthermore, we present the decision metrics used to forward the packets by the various AeroRP operational modes. The analysis of the ns-3 simulations shows AeroRP has several advantages over other MANET routing protocols in terms of PDR, accuracy, delay, and overhead. Moreover, AeroRP offers performance tradeoffs in the form of different AeroRP modes. 29 March 2011 AeroRP Analysis 2
AeroRP Analysis Outline Introduction and motivation AeroRP Simulation analysis Conclusions 29 March 2011 AeroRP Analysis 3
AeroRP Analysis Introduction and Motivation Introduction and motivation AeroRP Simulation analysis Conclusions 29 March 2011 AeroRP Analysis 4
Airborne Telemetry Networking Scenario and Environment Very high relative velocity Mach 7 10 s contact dynamic topology Communication channel limited spectrum asymmetric links Multihop data down omni C&C up directional among TAs through relay nodes 29 March 2011 AeroRP Analysis 5 GS GW TAs TA TA test article RN relay node Internet TA RN GS GW GS ground station GW gateway
Mobile Ad Hoc Networking Background Mobile Ad Hoc Network (MANET) no pre-established infrastructure self-configuring multihop wireless mobility MANET routing each node acts as a router as well as end system topology vs. geography proactive vs. reactive 29 March 2011 AeroRP Analysis 6
Geographic Routing Protocols Related Work MFR: Most forward with radius r NFP: Nearest with forward progress Compass Speeds up to 250 m/s S D B NFP Compass MFR C A 29 March 2011 AeroRP Analysis 7
AeroRP Analysis AeroRP Introduction and motivation AeroRP Simulation analysis Conclusions 29 March 2011 AeroRP Analysis 8
Neighbor discovery periodic beacons beaconless promiscuous Speed component Time to intercept (TTI) Purging neighbors stale entries predicting nodes out of range What to do when there is no route? ferry or buffer or drop AeroRP Overview 29 March 2011 AeroRP Analysis 9
AeroRP Neighbor Discovery Beacon mode send periodic HELLO beacons containing trajectory data Beaconless promiscuous mode nodes can overhear all packets from neighbors add location information to data packets 29 March 2011 AeroRP Analysis 10
n 0 = source node AeroRP Routing Metrics Variables x 0, y 0, z 0 = x, y, and z geographic coordinates v x0,v y0,v z0 = x, y, and z components of velocity n i = potential next hop x i, y i, z i = x, y, and z geographic coordinates v xi, v yi, v zi = x, y, and z components of velocity D = destination x d, y d, z d = x, y, and z geographic coordinates R = transmission range 29 March 2011 AeroRP Analysis 11
AeroRP Routing Metrics Speed Angle between x-axis of node s plane and velocity Angle between x-axis of node's plane and destination Velocity of node 180 Θ = atan2( v yi, vxi ) π Θ = atan2( yi y, xi xd ) d 180 π Relative speed of node w.r.t to destination s v + d i 2 2 = vxi vyi = v i cos( Θ Θ) 29 March 2011 AeroRP Analysis 12
AeroRP Routing Metrics Example s d 180 Θ = atan 2( v yi, vxi ) π 180 = atan 2( 14.15, 14.15) π = 135.2 180 Θ = atan 2( yi yd, xi xd ) π = atan 2(200 1200,600 1000) = 111.7 180 π 135.2 ( 111.7 ) = 23. 5 v i = 20 m/s n i Θ = 135. 2 Θ = 111. 7 v + 2 2 i = vxi vyi 2 = 14.15 + 20 m/s 14.15 2 s d = v i cos( Θ Θ) = 20 m/s cos( 135.2 ( 111.7 )) =18.4 m/s 29 March 2011 AeroRP Analysis 13
29 March 2011 AeroRP Analysis 14 AeroRP Routing Metrics Predict Neighbor out of Range Predict where neighbor will be in the future Is the neighbor going to be out of range? 2 0 2 0 2 0 p 0 1 0 1 0 1 ) ( ) ( ) ( = ) ( = ) ( = ) ( = i i i zi i i yi i i xi i i z z y y x x d t t v z z t t v y y t t v x x + + + + + < R for false R for true OutOfRange = p p d d
AeroRP Analysis Simulation Analysis Introduction and motivation AeroRP Simulation analysis Conclusions 29 March 2011 AeroRP Analysis 15
Simulation Parameters Variables Routing protocol: OLSR, DSDV, AODV, and AeroRP AeroRP modes: beaconless and beacon ferry, buffer, and drop Node density: 10, 20, 30,, 100 29 March 2011 AeroRP Analysis 16
Simulation Parameters Constants Area: 150 km 150 km 1 km Mobility model: Random waypoint 1200 m/s constant velocity and 0 s pause time Simulation time: 1000 s application time and 100 s warm-up time Physical channel model: 802.11b @ 11 Mb/s Friis propagation loss model Transmission power: 50 dbm, 27800 m range Application traffic: constant bit rate 1000 B packets with a rate of 8 kb/s 29 March 2011 AeroRP Analysis 17
Simulation Analysis Performance Metrics Packet delivery ratio (PDR) packets received divided by packets sent by application not necessarily all packets from application sent at MAC Accuracy packets received divided by packets sent by MAC measures quality of a route in highly dynamic topology Overhead excess bytes used to move data from source to destination Delay delay observed at the MAC layer from source to destination 29 March 2011 AeroRP Analysis 18
Simulation Analysis Packet Delivery Ratio (PDR) AeroRP performance increases as node density increases DSDV and AODV performing worst due to overhead 29 March 2011 AeroRP Analysis 19
Simulation Analysis Accuracy AeroRP ferry and buffer modes deliver most of the packets Beaconless mode has to rely neighboring nodes 29 March 2011 AeroRP Analysis 20
Simulation Analysis Overhead DSDV and AODV overhead reached to 1.5 Mb/s at 100 nodes AeroRP beaconless promiscuous mode performs best control packets piggybacked, no route -> drop packets 29 March 2011 AeroRP Analysis 21
Simulation Analysis Delay Higher delay for AeroRP buffer and ferry modes AeroRP drop mode has the lowest delay 29 March 2011 AeroRP Analysis 22
AeroRP Analysis Conclusions Introduction and motivation AeroRP Simulation analysis Conclusions 29 March 2011 AeroRP Analysis 23
AeroRP Analysis Conclusions Various AeroRP modes outperform other protocols AODV and DSDV did not perform well low PDR high overhead OLSR performed better compared to AODV & DSDV AeroRP has very high accuracy uses network resources efficiently Performance tradeoffs better PDR, accuracy, overhead higher delay due to buffering and ferrying 29 March 2011 AeroRP Analysis 24
References [RJCS2010] J.P. Rohrer, A. Jabbar, E.K. Çetinkaya, J.P.G. Sterbenz, Airborne Telemetry Networks: Challenges and Solutions in the ANTP Suite, Proc. of IEEE MILCOM, San Jose, CA, Nov. 2010, pp. 74 79. [PCJS2010] K. Peters, E.K. Çetinkaya, A. Jabbar, J.P.G. Sterbenz, Analysis of a Geolocation-Assisted Routing Protocol for Airborne Telemetry Networks, Proc. of ITC, San Diego, CA, Oct. 2010. [RJPS2008] J.P. Rohrer, A. Jabbar, E. Perrins, J.P.G. Sterbenz, Cross- Layer Architectural Framework for Highly-Mobile Multihop Airborne Telemetry Networks, Proc. of IEEE MILCOM, San Diego, CA, Nov. 2008, pp. 1 9. 29 March 2011 AeroRP Analysis 25
Acknowledgements DoD Test Resource Management Center Test and Evaluation / Science and Technology Program International Foundation for Telemetering (IFT) Justin P. Rohrer @ ResiliNets group at KU Kip Temple @ Edwards Air Force Base inet Working Group 29 March 2011 AeroRP Analysis 26
AeroRP Analysis Questions 29 March 2011 AeroRP Analysis 27
Appendix AeroRP Routing Flow Chart Receive data packet Is the packet for this destination? Yes Deliver packet up stack to application No Iterate through neighbors to find if they are expired or predicted to be out of range and remove from neighbor table if required Are any of the neighbors the final destination? Yes Send packet to destination (add delay if packet is queued) No Send to neighbor with the best TTI (add delay if packet is queued) No Calculate TTI of all neighbors and choose neighbor with lowest nonzero TTI. Does the local node have the best TTI? Yes Queue packet for ferrying or buffering if enabled and add queue delay 29 March 2011 AeroRP Analysis 28
Appendix Routing Metrics Time to Intercept (TTI) Δ d 2 2 2 = ( xd xi ) + ( yd yi ) + ( zd zi ) n 0 0 TTI = Δ d R sd for s n 2 d < 0 and Δd >R otherwise n 1 Choose lowest nonzero TTI 29 March 2011 AeroRP Analysis 29
Appendix Routing Metrics Example Prediction xi = xi + vxi ( t 1 t0) = 800 + 200(62 60) = 400 yi = yi + vyi ( t 1 t0) = 1000 + 200(62 60) = 1400 n 1 d p = 721.1 m n 1 z i = zi + vzi ( t 1 t0) = 0 + 0(62 60) = 0 n 0 n 0 d 2 2 2 p = ( x0 xi ) + ( y0 yi) + ( z0 zi ) 2 2 = (800 400) + (800 1400) + (0 = 721.2 m 0) 2 true OutOfRange = false = true for d p R for d <R p R = 600 m 29 March 2011 AeroRP Analysis 30