Curtis Kelsey University of Missouri
Introduction Method Experiment Results Conclusion Summary
Aeronautical Networks are unique Mixture of static & dynamic nodes Extremely high speed nodes Custom network stack is necessary Dynamic airborne environment
ANTP AeroTP (TCP) AeroNP (IP) AeroRP (Routing) AeroGW* AeroGW Converts TCP AeroTP IP AeroNP Link/MAC iNET MAC PHY iNET PHY
Conversions Occur: Ground Stations Aeronautical Nodes Possible Overhead Implications Less data transferred Communication windows lost Most Significant Delay Egress conversion from MAC to IP (Similar to ARP) Egress is not constrained by time due to node movement
Does delay caused by the conversion process result in excessive data loss? Implementation of entire suite beyond the scope of one semester Implement a network simulation Use additional delay as control variable Analyze data delivery
Virtualbox or Hyper-V Requirements Gcc/g++ > 3.4 Python Mercurial Bazaar Etc… Downloading clone wget
Build./build.py –enable-examples –enable-tests Configure./waf -d debug --enable-examples --enable-tests configure Test./test.py –c core Run a Project./waf –run
10 Airborne Nodes/Routing Nodes (Wireless) Random Walk Random Speed 5 Ground Stations (Access Point) Random Location GS to Internet Direct Link 100Mbps 2ms delay
1 Destination Internet Node (Wired) 100Mbps 1/10/100/1000ms delay Traffic 100-1kb packets/10 seconds UDP Zone 1000 x 1000 area
PointToPointHelper Handles Wired/Wireless Bridge CsmaHelper Handles wired nodes WifiHelper Handles wireless nodes MobilityHelper Handles AN and RN Mobility
Packet capture enabled AP Csma (Wired) Wireless Nodes
Simulation ran for 1ms additional delay 10ms additional delay 100ms additional delay 1000ms additional delay At Wireless Network Edge
Packets captured at Wireless AP (Ground Station) Wired Node Pcap file processed with Tcpdump & sent to log files Tcpdump –nn –tt –r (pcap file) > (log file)
How many of the 100 packets got delivered? Wired Node Wireless Nodes
1ms 100% packet delivery No delay between transmit/receive 10ms 100% packet delivery No delay between transmit/receive 100ms 100% packet delivery No delay between transmit/receive 1000ms 100% packet delivery No delay between transmit/receive
Delay implemented on wired node does not affect traffic across point to point link Move delay variable to p2p link Random walk & speed for wireless nodes is not causing dropped packets Expand zone & define a high velocity Amount of data transferred needs to be increased Illustrates dropped connections
(Primary Paper) E. K. ¸Cetinkaya and J. P. G. Sterbenz. Aeronautical Gateways: Supporting TCP/IP- based Devices and Applications over Modern Telemetry Networks. In Proceedings of the International Telemetering Conference (ITC), Las Vegas, NV, October Cetinkaya, E., & Rohrer, J. (2012). Protocols for highly-dynamic airborne networks. Proceedings of the 18th annual international conference on Mobile computing and networking, 411–413. Retrieved from Narra, H., Cetinkaya, E., & Sterbenz, J. (2012). Performance analysis of AeroRP with ground station advertisements. Proceedings of the first ACM …, 43–47. Retrieved from KEN= KEN= Sterbenz, J., Pathapati, K., Nguyen, T., & Rohrer, J. (2011). Performance Analysis of the AeroTP Transport Protocol for Highly-Dynamic Airborne Telemetry Networks. Retrieved from J. P. Rohrer, E. Perrins, and J. P. G. Sterbenz. End-to-end disruption-tolerant transport protocol issues and design for airborne telemetry networks. In Proceedings of the International Telemetering Conference (ITC), San Diego, CA, October 2008 A. Jabbar, E. Perrins, and J. P. G. Sterbenz. A cross-layered protocol architecture for highly-dynamic multihop airborne telemetry networks. In Proceedings of the International Telemetering Conference (ITC), San Diego, CA, October 2008.
Introduction ns3 setup Experiment Construction Results Conclusion Summary