Multiple NIC segregation in Windows

[gsx-rob]

I'm in a shop that requires our workstations to have two NICs. One attaches to a router and through to a DCHP internet connection. The other is a static IP via crossover to our instrumentation system. The issue is that these systems can be extremely remote and rely on satellite or 3G Network connections and are not overly reliable. Every time our net connection drops, BOTH NICs reset, causing us data loss from our instrumentation. Does anyone know of a way to separate the connections so that if one drops, the other is left unmolested? Any help would be greatly appreciated.

 

[gsx-rob]

I'm in a shop that requires our workstations to have two NICs. One attaches to a router and through to a DCHP internet connection. The other is a static IP via crossover to our instrumentation system. The issue is that these systems can be extremely remote and rely on satellite or 3G Network connections and are not overly reliable. Every time our net connection drops, BOTH NICs reset, causing us data loss from our instrumentation. Does anyone know of a way to separate the connections so that if one drops, the other is left unmolested? Any help would be greatly appreciated.

BTW> These are all XP 32-bit boxes

 

[eibgrad]

Are we talking about Windows systems here? I'm unaware of any situation where Windows resets other NICs when some given NIC drops its internet connection. They're unrelated, at least logically. Is there some other piece of information here that perhaps we're missing? Are perhaps these connections bridged? Or using ICS between them? IOW, what’s the relationship between the network connections, if any?

 

[gsx-rob]

This is on Windows XP 32bit with two physical NICs. One is an onboard Intel NIC that connects to our instrumentation system and requires continuous data flow. the second is via a USB to Ethernet adapter (no expansion slots available in our case) I've also tested using a WiFi USB adapter with the same result. As soon as the Internet side connection drops, we also lose our instrumentation side connection until it has a chance to reestablish. In the field it is due to a bad connection to the net, but I can replicate in the lab by yanking the cable internet side, or the USB WiFi with the same result.

 

[eibgrad]

Still doesn't make sense. Is the instrumentation side *using* that other network connection to access the Internet? IOW, if the internet connection drops, is that other NIC, the one connected to the instrumentation side, somehow dependent on the presence of the other connection, perhaps for providing Internet access to it? Or are these 100%, TOTALLY, UNRELATED network connections? Again, this just doesn't make sense. Windows will allow you to install dozens of network adapters and make numerous network connections, ALL of which act independently, at least by default. Whatever happens on one network connection has absolutely nothing to do w/ any others. Not unless they have some sort of dependency, such as being bridged, or one is providing internet access to the other. There has to be SOMETHING like that, or else I have no clue what's going on there. Or maybe there’s some other event that’s effecting both network connections at the same time, but not just because the internet connection is lost on one of them.

 

[gsx-rob]

The instrumentation side has no dependence or capability to connect to the net. They are completely different networks. What lead me to believe that Windows is the culprit is that when a secondary NIC is added, a disruption of that connection causes the instrumentation connection to drop, showing the obligatory "X" through the connection in the taskbar and then reconnecting after 4 seconds. Perhaps a bug in the instrumentation software itself then?