Terminal Serices Clients Disconnect on Minimize

[[e-mail]]

I have a Windows 2000 Advanced Server with Terminal Server running in
application mode. We have 7 users on our LAN connected to the server
via a Cisco switch. We also have 43 remote users connecting through
our T1, Cisco router, PIX 506E, and, ultimately, the same Cisco switch
that our LAN users connect through. When my users here on the LAN
minimize their TS connections, they are disconnected after a period of
about 5 minutes when they try to return to their session. I have been
up and down the newsgroups, enabled and tweaked KeepAlives and
MaxTcpDataRetransmissions, and checked the connection settings in both
AD and the Terminal Services snap-in, all to no avail. The funny thing
is that the remote users do not experience the same problem. Even when
I connect from home through my DSL I do not get disconnected. It seems
to be primarily my WinXP Pro SP2 clients (although I do not have any
SP1 clients), but I have had reports from some of my Win2k clients in
the office that they, too, have been disconnected. I put a packet
sniffer on my machine, monitoring port 3389, and sometimes I see
packets with the session minimized, and other times I do not. Even
when I do see packets, it seems as though I will get a short burst when
I activate my TS windows, but it still disconnects. Does anyone have
any idea what the problem could be? I HAVE NOT adjusted the MTU
settings, as I can't imagine it being a factor in a NIC to switch to
NIC setup, as I have here on my LAN. I'm ready to kick this thing over
and tell my boss that the server had a most tragic accident. :oP

[[e-mail]]

Update on this question: See the result of my packet sniff on my
minimized connection, as well as my arp table, which I outputted
immediately afterwards (sorry if it looks rough).

Time || Src. MAC || Dest. MAC || Src. IP || Dest. IP
23:51:06.43 || 00A0CC-308FA0 || 004045-25657D || 192.168.1.3 ||
192.168.1.101
23:51:06.62 || 004045-25657D || 00A0CC-308FA0 || 192.168.1.101 ||
192.168.1.3
23:56:53.92 || 004045-25657D || 0011BB-3DBCA3 || 192.168.1.101 ||
192.168.1.3
23:56:53.92 || 0011BB-3DBCA3 || 004045-25657D || 192.168.1.3 ||
192.168.1.101

ARP Table
Internet Address Physical Address Type
192.168.1.1 00-11-bb-3d-bc-a3 dynamic
192.168.1.3 00-a0-cc-30-8f-a0 dynamic
192.168.1.5 00-50-8b-6f-e3-3d dynamic

Now, 192.168.1.3 is my Terminal Server, however it appears as though
the interface at 192.168.1.1 being addressed after the long pause.
192.168.1.1 is the default router, and it is my PIX 506E firewall. Is
the lack of keepalives dropping the MAC translation for 192.168.1.3
from my ARP table causing my machine to forward packets to the default
router? I don't understand.

[[e-mail]]

Problem solved! After some more packet sniffing, I realized that the
PIX was responding to ALL ARP requests. After the connection went idle
and the machine dropped the entry from the ARP table, it would send out
a broadcast to see who had the 192.168.1.3 IP address, and the PIX kept
responding. Apparently, (and I don't really know that much about the
PIX) there is a setting called "Proxy-ARP" which I disabled on the
internal interface and now things are moving right along.