I have tried to google the topic, but I can’t somehow decide, if having two IP addresses from the same subnet on two different physical interfaces, is a problem or not.
The situation is quite simple - imagine PC / notebook connected by both the ethernet cable and wi-fi to one MT router via a switch, being offered IP addresses from the same subnet. E.g. eth: 10.0.1.1/24 and wi-fi: 10.0.1.2/24.
I just wonder, if it could posses some problems in communication? Or even some loops? I have found we’ve got such setup on multiple locations. I reconfigured one of them to use the VLANs to separate the traffic, but was told by one of the suppliers, that there should not be any problem to such setup.
So - is that just me having wrong assumption of eventual problems? E.g. some notebook producers allow you tu disable wi-fi, if eth is connected, via a bios setting and I think there might be a good reason for that …
If none of LAN infrastructure devices “play smart”, then the most likely outcome would be asymmetrical packet flow: if some connection uses laptop’s wireless IP address (ingress will flow through wireless) while routing priorities (e.g. due to lower metric) will use ethernet for egress.
If addresses for different interfaces are not in same subnet, then connections might break due to reasons explained in previous paragraph, only in this case some router/firewall might discard packets with “alien” src-address.
Which means it’s the matter of sanity not to use both wired and wireless interface at the same tine … unless one knows better.
If the ethernet interface and the wireless interface on the router are on the same bridge, then it may not be a big problem, the laptop would just select which interface the traffic would go out of.
If the wlan1 and etherx were on separate interfaces that require routing, you’d have problems, not least of which the DHCP config handing out those IP addresses on different subnets would be wrong.
Thats up to the individual running the laptop. Connectivity is required and provided.
Two separate mac addresses two connections, only one will be used at a time.
No need to be afraid of any loops, as the two interfaces are not bridged on the Windows PC. The only thing I can imagine to happen could be that the PC would respond to the ARP request for one of the addresses also with the MAC address of the “wrong” interface, but /tool sniffer on the Mikrotik or Wireshark on the PC should quickly tell this, and it may not cause a problem either if the PC accepts IP packets even if they come in via a “wrong” interface, which it should. So the biggest issue should be that you would be unable to predict which of the two addresses the Windows will choose for outgoing traffic.
The disabling of WiFi if Ethernet port is connected has more to do with saving power, WiFi bandwidth, and IP address pool if wired connection is available. And it may become a PITA if you need to use each interface in another network, e.g. one of them for internet connection and access via TeamViewer/Anydesk/Remote Desktop and the other one to connect to a device in LAN, and you have to reboot to disable the feature in BIOS.
The router is CCR1009, no switch, no bridge there. The rest of the LAN, including wi-fi APs, come from the separate switch, connected to the eth2 port of the MT router, via just one cable. No VLAN there.
So basically - MT gives DHCP addresses to whatever is on the switch, which means - when you connect your notebook by both the eth and wi-fi, you obtain 2 addresses from identical subnet, assigned to two different physical interfaces.
And that’s where I wondered, if it’s correct, or not …
Ok, I should have more correctly said that the AP/WiFi connection and the wired connection were in the same broadcast domain, which having them on the same bridge in the router would also accomplish.
As others have said, the OS should be smart enough to use one or the other, and as long as they are not bonded on the laptop (not normally done) then you should not have to worry about loops.
It is possible, on the WIFI-AP just need to use “bridge to vlan” or “bridge to LAN” - then “the WIFI clients will look like as a wired LAN clients”.
The other question is also possible but not with mikrotik router - that technology is called VRF ( virtual routing and forwarding ) - because mikrotik’s VRF implementation is not the “true VRF” they using the “routing mark” hacks. The VRF is like a L3 domain, when you assign an interface to specific VRF - domain, it has its unique RIB and FIB, ( route table ), when you want exchange traffic between the VRF-Domains thats called “VRF route leaking” in this case you need to use some dynamic-routing protocol to exchange the route-tables between the VRF-Domains, that protocol is MP-BGP ( Multiprotocol - BGP ).
When you using VRF without BGP / MP-BGP - thats called “VRF-lite” in cisco terminology.
Vendors: Cisco, Juniper, HPE Comware
I introduced the basics, now you can use google for the rest.
I am sorry but I am not that skilled to have an experience with VRF, etc. As for bridging the wi-fi to LAN, I am not sure. It is an Aruba virtual controller + several APs setup, so don’t want to mess with that.
Maybe the easiest it is to split the network using VLANs, settting-up a DHCP for a wi-fi part, so that I can distinguish the traffic. But hey - thanks anyway for that VRF stuff - there’s always something new to learn …