Which IP will be picked when IAD register in N-Command

[Anonymous]

I have question.   When an IAD register to N-Command and it has multiple IPs (subinterface), which IP will be picked by MSP.   In other words, what is the algorithm to pick the IP  (by vlan#? numeric order of IP?)...Thanks.

[Anonymous]

The logically closest interface.

For example, you have an eth1 and an eth2, eth1 is your LAN with an address of 192.168.123.1 and eth2 is WAN Internet of 123.123.123.31. if nCommand is on the LAN that the source IP will be 192.168.123.1 and if it's on the internet the source IP would be 123.123.123.31.

If the source interface had a secondary address it would not be used as the source for auto-link unless you specifically applied source translation to the outbound traffic.

One of the nice things about nCommand is that it tracks devices based on serial number not source IP address so you can have multiple devices behind NAT all reporting the same source IP. Since all communication is initiated from the device there are no security or connectivity issues. The exception would be the link on the device page in nCommand pointing to the devices web interface. Obviously that would need to be one-to-one and allowed.

[Anonymous]

What about multiple subinterfaces on the same physical interface? 

For example, IAD's eth 0/1  (WAN side, used to reach MSP ), there are subinterface eth 0/1.103 (ip address 165.23.31.6/30) and eth 0/1.312 (ip address 62.51.123.6/30). 

Which logical subinterface is closest?

[Anonymous · Accepted solution]

logically closest would be determined by the routing.

If I assume that one of these logical interfaces is on the same subnet as the default gateway and nCommand is not on any local subnet and there are no more specific routes that would affect traffic to nCommand. Then traffic to nCommand would be would be sent to the default gateway, the interface on the same subnet as the default gateway would be logically closest.

It helps to remember that Interfaces are all logical and Ports are physical. They may be used interchangeably now but this wasn't always the case, and this kind of fundamental view can be very helpful when trying to figure all this out.

[Anonymous]

You are right.   I just verified couple of them. Thanks.