ndt-users list created

[Clarke Morledge <>]

Rich,

Thanks for clarifying this for me.  It will keep me busy for awhile :-)

In our experience, we have been able to use NDT to pinpoint physical layer 
problems fairly easily based on two types of results:

a.  both C2S and S2C tests show much lower than expected; e.g. less than 2 
Mbps in both directions for a 100 Mbps path in a LAN environment.

b.  radical asymmetry in the results; e.g. C2S shows 80 Mbps or more but 
the S2C test shows less than 5 Mbps.

In many cases, NDT will correctly predict that there is a duplex mismatch. 
In other cases, we can identify a NIC problem in either a switch or a 
workstation --- or in the case of wireless, we can identify the presence 
of RF interference.

However, in the case where we have "less than radical" asymmetric results; 
e.g. 80+ Mbps one way with ~30 Mbps the other, we always will see the 
packets-out-of-order message.  These problems have been really difficult 
to pin down because of the number of variables involved, but we have yet 
to identify any physical layer issues.  The results are not necessarily 
bad, but it always leaves me (and our users) scratching our heads.

Any word on when web100 will make it into a stock Linux kernel? ;-)

Clarke Morledge
College of William and Mary
Information Technology - Network Engineering
Jones Hall (Room 18)
Williamsburg VA 23187