[ietf81-tech] Curious Missing Associations on bora Graphs

[Bill Jensen]

Folks,

While sitting in the plenary, I was curious why the (bora) graph 
association count seemed not to reflect the likely associations for the 
ballroom.  Sure enough, there was a minor range error in the bora script 
that the produces the bora associations graph.  I have fixed the script 
and things look more plausible.  It is unlikely the results will 
significantly change the overall peak association count but it certainly 
affects the count for the Monday evening plenary counts.

Bear in mind this script is a non-elegant way to get the counts but it 
yields easy graph results.  Another means of producing the counts is 
poll the databases, however this has been problematic in the past. 
Fortunately, rrd-based SNMP gathered results exist.  Its simply that 
they are the most difficult to reflect in graphs due to the requirement 
for hand editing.  This problem will (hopefully) get resolved through 
the use of some of the newer plugins and/or careful graph reuse 
techniques.

Thanks,
-wej