23C3 Sputnik Tracking Analysis

Jan. 30, 2007
download (version 0.3) of my code to analyse Sputnik data (under GPL):
  1. concatinate all binary log files, cat log-* > log.bin
  2. run cvtbeacon log.bin > log.txt
  3. run cut -d \ -f 1,2 log.txt | python beaconstat.py to compute histogram data for plotting
  4. run cut -d \ -f 1,2,3 log.txt | python beaconadjacency.py to compute graphviz plot showing receiver adjacency
or simply try the Makefile! :)

the actual Sputnik data is available from here (under CC-A-NC-SA 2.5)

Note: file log-2006-12-27-13 is corrupt; 8 bytes at offset 0xc4ff8 must be deleted!


Sputnik analysis results below...

(animated GIF, 2.13 MB)
tracking of a Sputnik user whose tag is easily identifyable because its reset counter starts from > 0xf000;
(using first 660 frames covering Dec. 27, 20:57 to Dec. 28, 14:19)

Sputnik receiver adjacency
receiver adjacency (note: 1.16 and 3.5 have been assigned to floor plan according to analysis);
numbers on edges indicate the trackings received simultanously by two receivers

Sputnik trackings 
per minute
trackings per minute (total)

Sputnik trackings per minute (floor A)
trackings per minute (for the individual receiver IPs on floor A)

Sputnik trackings per minute (floor B)
trackings per minute (for the individual receiver IPs on floor B)

Sputnik trackings per minute (floor C)
trackings per minute (for the individual receiver IPs on floor C)


Peter Meerwald, pmeerw@pmeerw.net