On Tue, 31 Jul 2001, Ryan O'Connell wrote:
> Because it takes time for the BSD boxes to process the packets at each
> end then send the packets to the network card for transmission. The
> RTT measured by ping includes time spent in the local and remote IP
> stacks.
FreeBSD isn't that sloppy. Here's one FreeBSD host pinging another on a
full duplex 100Mb network. Both boxes use the fine Intel Pro100 chipset,
a.k.a. fxp. Not a Cisco switch, but an HP2424 (which oddly enough speaks
CDP):
fxp0: flags=8843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> mtu 1500
media: 100baseTX <full-duplex> status: active
--- 216.223.x.x ping statistics ---
11 packets transmitted, 11 packets received, 0% packet loss
round-trip min/avg/max/stddev = 0.185/0.208/0.312/0.034 ms
The original poster was getting 1ms+ on this sized packet! Something
seems wrong.
If I bump this to 1500 bytes, here's what I see:
--- 216.223.x.x ping statistics ---
11 packets transmitted, 11 packets received, 0% packet loss
round-trip min/avg/max/stddev = 0.768/0.775/0.790/0.006 ms
Still sub 1ms...
One difference I see is that I've fixed both sides of each connection at
100Mb FD, because I don't trust autodetect.
> Because some of the time taken in the IP stack is fixed (Amount of time it
> takes to process the IP header, make routing decisions etc) regardless of the
> packet size whereas other time is variable. (Such as the amount of time it
> takes to copy the packet between different memory locations)
Agreed. Here's a benchmark, sort of:
PING 127.0.0.1 (127.0.0.1): 56 data bytes
--- 127.0.0.1 ping statistics ---
11 packets transmitted, 11 packets received, 0% packet loss
round-trip min/avg/max/stddev = 0.107/0.123/0.137/0.010 ms
PING 127.0.0.1 (127.0.0.1): 1500 data bytes
--- 127.0.0.1 ping statistics ---
11 packets transmitted, 11 packets received, 0% packet loss
round-trip min/avg/max/stddev = 0.137/0.146/0.159/0.007 ms
It's a fairly busy machine with 254 aliases hanging off it, I assume that
adds a few microseconds somewhere...
Charles
> --
> Ryan O'Connell - <ryan@complicity.co.uk> - http://www.complicity.co.uk
>
> I'm not losing my mind, no I'm not changing my lines,
> I'm just learning new things with the passage of time
>
This archive was generated by hypermail 2b29 : Sun Aug 04 2002 - 04:12:47 EDT