Is anyone testing DCCP with limited bandwidth on a low latency network?
I want the system to drop packets instead of buffering them.
I'm having a hard time setting this up, and would appreciate advice from anyone
else.
I tried to use tc/netem as follows:
----------------
tc qdisc add dev eth0 root handle 1:0 netem
tc qdisc add dev eth0 parent 1:1 handle 10: tbf rate 70kbit burst 1000 latency
20ms
-------------------
I have two machines on a Fast Ethernet lan, (no bridge - which may be
my problem) and I use this command on both systems.
1) First of all, when I use ttcp or iperf, one side has twice the
throughput as the other. So netem doesn't seem symmetric.
2) Second, I used ttcp to send a small number of packets:
ttcp -n200 -l100 -t receiver
This stalls, even after waiting for an hour. I ran tcpdump during this,
and no packets were being sent, except for arp and the pings I
initiated. Wnen I removed the filter, ttcp completed. I'm not sure I understand
why.
3) When I try to use DCCP with the above filter, tfrc.tfrcrx_rtt grows
to some very large values (e.g. 44936) even when I setsockopt(fd,
SOL_SOCKET, SO_SNDTIMEO,..) to 400.
Thanks for any advice.
Bruce
-
To unsubscribe from this list: send the line "unsubscribe dccp" in
the body of a message to [EMAIL PROTECTED]
More majordomo info at http://vger.kernel.org/majordomo-info.html
