> We run all Dell hardware with Intel NICs.  And we run some very
> specialized applications on them.  None of them exhibit any
> performance problems (or ever have).  So I have a very hard time
> believing that this NIC is to blame.

Someone who knows more about these things can correct me if I am wrong, but everything 
I have read
points to problems in the low-level software of the built in Gigabit NIC. I also 
believe there is a
problem with some or most other Gigabit NICs.

Why do I say this? People who have this problem and either remove the NIC and replace 
it with a
100Base-T Nic (if the Gigabit nic is removable), or disabling it on the motherboard 
(when the nic is
not removable) see their problem go away. However, just disabling it in Windows and 
using another
nic does *not* seem to make the problem go away. If these statements are correct, 
there must be a
problem with the firmware on the nic (or Windows is actually talking to the disabled 
nic). The only
way this would not be true is if Imail was using low level software to talk directly 
to a (possibly
disabled in Windows) nic. At least one person who knows far more than I do said that 
this is not the
case; I think it boggles the mind to think that it would.

So who is to blame? First off, problems seem to exisit when Windows has disabled the 
nic. This would
seem to indicate that Microsoft's code does not completely disable the nic. If this is 
not the case,
then the firmware on the nic is actively disrupting other nics somehow.

None of this should be taken to say that Imail is perfect; of course there are bugs. 
Still, I just
can't see how Imail could be to blame unless someone can show that Imail talks 
directly to nics.
Having said that, it is certainly clear that Imail uses the network in a way that 
other applications
don't. This is largely a function of how mail communications works compared to, say, 
webpages or
databases. It is certainly possible though that Imail could be written in such a way 
that the
network isn't stressed in "the wrong way" to cause this problem. However, if the 
"real" flaw exists
in the nic, it would certainly be better to have it fixed there.

If anything here is not true, please correct me.

Paul Navarre


To Unsubscribe: http://www.ipswitch.com/support/mailing-lists.html
List Archive: http://www.mail-archive.com/imail_forum%40list.ipswitch.com/
Knowledge Base/FAQ: http://www.ipswitch.com/support/IMail/

Reply via email to