Re: Re: question about capsule's lifetime

[Timothy Stack <stack@cs.utah.edu>]

> Hi,tim

hi,

> The following is output of tcpdump(ANEP-enabled).
> I use multi hosts in my AA ,and the topology is 
> 
> 202.198.26.9<--->202.198.26.6<--->... 
> 
> I use "tcpdump -T anep src host 202.198.26.9 and dst host 202.198.26.6
> and dst port 8004" to see what happened.

Whats the sequence of events here?  Do you send a request from ...26.9 to
26.6 which then polls the file?  And whats the expected response?  Can you
provide a trace of packets going in the other direction 26.6 -> 26.9?

> The output when <10 seconds are:
> 
> 17:41:13.774498 ws26-9.jlu.edu.cn.32802 > 202.198.26.6.8004: v1 Ants (2817) F[I:dstid:IPv4][I:srcid:UNK]
>  (frag 1559:1480@0+)

If this is a .o file...

> 17:41:13.804498 ws26-9.jlu.edu.cn.32802 > 202.198.26.6.8004: v1 Ants (125) F[I:dstid:IPv4][I:srcid:UNK]
>  (DF)
> 17:41:13.824498 ws26-9.jlu.edu.cn.32802 > 202.198.26.6.8004: v1 Ants (1189) F[I:dstid:IPv4][I:srcid:UNK]
>  (DF)

Then what are these capsules?

> The outputs when >10 second are:
> 
> 17:44:19.314498 ws26-9.jlu.edu.cn.32803 > 202.198.26.6.8004: v1 Ants (2817) F[I:dstid:IPv4][I:srcid:UNK]
>  (frag 1560:1480@0+)
> 17:44:19.344498 ws26-9.jlu.edu.cn.32803 > 202.198.26.6.8004: v1 Ants (84) F[I:dstid:IPv4][I:srcid:UNK]
>  (DF)

Again, whats this 84 byte capsule?  This doesn't track with the previous
trace where they were 126 bytes.  The difference in time between the first
and second capsules (774498/804498 and 314498/344498, a difference of
30000).

> 17:44:21.354498 ws26-9.jlu.edu.cn.32803 > 202.198.26.6.8004: v1 Ants (84) F[I:dstid:IPv4][I:srcid:UNK]
>  (DF)
> 17:44:23.364498 ws26-9.jlu.edu.cn.32803 > 202.198.26.6.8004: v1 Ants (84) F[I:dstid:IPv4][I:srcid:UNK]
>  (DF)
>  
> The capsule carries a  .o file which is 2728 bytes long.so I think the
> first line in both scenario corresponds to that capsule.That is to say,
> in both scenarios,the capsule are sent out to 202.198.26.6, so there
> may be something wrong at the node 202.198.26.6 when >10 seconds.

Can i take a peek at the code?

> 	thanks.
>  
>  
>  woodhead.
>  dcsdcs1@mail.jlu.edu.cn
>  2002-06-25

tim stack