Can network latency affect Terminal Server connections?

  • Thread starter Thread starter Vasoo Veerapen
  • Start date Start date
V

Vasoo Veerapen

Good day group,

I previously posted a question regarding TCP resets being sent to TS
clients. To recap, I have 5x Windows 2000 servers running NLB. We have
approximately 200 TS clients connected at any time. These clients
connect to the server via the WAN which consists of Wireless and Frame
Relay links. WAN speeds differ from site to site from 64Kbits through
to 2Mbits.

A while ago, we upgraded the wireless hardware and it seems that all
our troubles started then. All our TS clients started getting broken
connections. I sorted out the problem by modifying the following
registry settings on the TS servers, but my question is:

Why were all our clients affected and not only those on the wireless?

[HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Terminal Server]
"KeepAliveEnable"=dword:00000001
"KeepAliveInterval"=dword:00000001

[HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\Tcpip\Parameters]
"KeepAliveInterval"=dword:000003e8
"KeepAliveTime"=dword:0001d4c0
"TCPMaxDataRetransmissions"=dword:0000000a

Ideas anyone?
 
It could be a number of things. You might have an
upstream router that is having trouble with the new amount
of traffic. If one of your upstream routers is a wireless
router, it may not be handling the increased traffic load
properly. Does any of this sound feasible?

-M
-----Original Message-----
Good day group,

I previously posted a question regarding TCP resets being sent to TS
clients. To recap, I have 5x Windows 2000 servers running NLB. We have
approximately 200 TS clients connected at any time. These clients
connect to the server via the WAN which consists of Wireless and Frame
Relay links. WAN speeds differ from site to site from 64Kbits through
to 2Mbits.

A while ago, we upgraded the wireless hardware and it seems that all
our troubles started then. All our TS clients started getting broken
connections. I sorted out the problem by modifying the following
registry settings on the TS servers, but my question is:

Why were all our clients affected and not only those on the wireless?
Click to expand...
[HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Termin
al Server]
"KeepAliveEnable"=dword:00000001
"KeepAliveInterval"=dword:00000001

[HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\Tcpip
\Parameters]
"KeepAliveInterval"=dword:000003e8
"KeepAliveTime"=dword:0001d4c0
"TCPMaxDataRetransmissions"=dword:0000000a

Ideas anyone?
.
Click to expand...
 
Well, we have a VLANed Cisco 4006 switch which connects direct to the
wireless system. Users connected to Frame Relay also pass through this
switch. The handling capacity of the switch swtich should be able to
cater for the additional traffic as the max throughput of the wireless
is 100 Mbits/s. However, how could an upgrade on the wireless affect
users on Frame Relay? Is there a problem with the Windows 2000 TCP/IP
stack WRT different networks with different latencies connecting to
the same server?
 
Well, as soon as those packets hit the Cisco switch they
should all be treated as the same, so I'm a little
doubtful that Windows can even differentiate between the
physical packets coming from the wireless systems as
opposed to the frame relay systems.

This is a little odd though...could it be possible that
the new wireless hardware is causing some funny
reflections or packet loss within your network? Do you
have the latest and greatest firmware installed on the
switches?

-M
 
The Cisco version details area as follows.

boot system bootflash:cat4000-is-mz.121-13.EW2.bin
 
Back
Top