CogNet Nodes
The CogNet Nodes are a component of the CogNet 802.11 Testbed that capture all of the data that are stored in the CogNet Repository. They run under VoyageOS, a stripped-down version of Debian Linux. The Nodes sample packets on all 802.11b/g channels, 1-11, by switching the attached wireless card’s channel setting using the MadWifi wlanconfig command. The wireless cards are in Monitor mode for the capture process, which allows capture of all packets encountered, regardless of the access point with which the packets might be associated. A standard, unmodified version of tcpdump is used to capture the packets to disk. This file is then transferred via FTP to the CogNet Controller.
Various shell scripts perform the tasks necessary to keep the Nodes interfaced with the CogNet Controller. These scripts are described in the following list.
setup_interfaces.sh: Uses the MadWifi wlanconfig command to set up the wireless interface for capturing in monitor mode.
loop_scan.sh: Calls tcpdump (unmodified version) using the –w option to capture the data to a file. Snap length can be configured so that only the first n bytes of a packet are saved to disk; the length of the entire packet is, however, still saved. The script also uploads the capture file after a specified interval to the CogNet Controller, and begins saving to a new capture file. The capture file name has the following format:
<node id>_<random num.>_<start month>_<start day>_<start hour>_<start minute>__<end month>_<end day>_<end hour>_<end minute> Example: node001_15703_03_21_15_57__03_21_16_03
round_robin.sh: Rotates the wireless interface among 802.11b/g channels, 1-11, switching every second.
timesync.sh: Synchronizes the Node’s system time with the server’s, using ntp. This is currently performed every 15 seconds.
auto_ftp.sh: Uploads a capture file to the server automatically and sends an HTTP request to the Data Processing Module of the CogNet Controller to begin processing of the captured data.
The following diagram shows the operation of a CogNet Node:
