[Planetlab-users] A question about your hosting PlanetLab nodes

[ltang]

Hi Everyone,

   I'm writing to ask some help and suggestion to explain the following 
strange experiment result.

   Basically, I used this node as a PlanetLab node 
'planetlab2.win.trlabs.ca' to 'ping' a large number of destination IP 
addresses, which were extracted from each IP prefix in the BGP routing 
table. I did the 'ping' experiment twice, respectively on January 6~10 
and on February 6~13, towards the same destination IP list. The results 
of both experiments are generally consistent with each other, showing 
nearly 30% of the destination IP addresses cannot be directly reached 
from 'planetlab2.win.trlabs.ca' through 'ping', but can if 
'planetlab2.win.trlabs.ca' delegate the probe to another PlanetLab node. 
However, on February 15, when I restarted this experiment again, most of 
the unreachable destination IP addresses in both last two experiments 
become reachable by  'planetlab2.win.trlabs.ca'. Therefore, I wonder if 
this is due to some change of the network condition of 
'planetlab2.win.trlabs.ca'.
   To provide further information that may be helpful, I found in both
two previous experiments, there was a clear time point before which more
than 95% of the destination IP addresses were unreachable, but after
which more than 95% are reachable. Do you think it is possibly due to
any security reaction triggered by the frequent and flooding 'ping'?

   I'll be very grateful to those that give any information and 
suggestion to this question. Thanks in advance!


Best Regards,
- Li