Hi Len,

[...]
> ends of the link.  Currently I have mru set to 1024, mtu is being
> determined by the system on the other end and according to netstat is
> 1002.  Any ideas as to what might be causing this would be most
> welcomed.  Thanks.

though I don't think I am really qualified to answer your questions,
I might have some guesses:

First the MTU might be different in both directions. Yes, this is possible :-)
Other factors might be that fragmentation will take it's time. Let's imagine that
the other machine is connected to the ethernet. Then incoming packages might be
bigger then the MTU and therefore the packets need to be fragmented before they
can be sent to your machine. In the other direction, no fragmentaion is needed on
the other side so packets can be forwarded without delay. I don't think
that this might cause such a big delay, though. These are just my rough guesses.
If you really want to know what's going on, get yourself the book TCP/IP
Illustrated written by Richard W. Stevens. This is an _excellent_ book about
TCP/IP. It analyses all the protocols of the TCP/IP suite and shows how to
actually see what is going on on the wire. Highly recommended :-)

  Guenther