ether1 192.168.1.200/24 wlan1 10.0.10.100/32 | | | wlan1 10.0.10.101/32 ether1 192.168.2.200/24 wlan2 10.0.10.102/32 | | | wlan1 10.0.10.103/32 ether1 192.168.3.200/24 wlan2 10.0.10.104/32 | | | wlan1 10.0.10.105/32 ether1 192.168.4.200/24
[admin@WHG] routing ospf> print router-id: 0.0.0.0 distribute-default: never redistribute-connected: as-type-2 redistribute-static: as-type-2 redistribute-rip: no redistribute-bgp: no metric-default: 1 metric-connected: 20 metric-static: 20 metric-rip: 20 metric-bgp: 20
[admin@WHG] routing ospf network> print Flags: X - disabled, I - invalid # NETWORK AREA 0 192.168.x.0/24 backbone
Did you try to run this? I'd be suprised if it worked with those /32 netmasks on both sides of the wireless link. After all, 802.11 is a broadcast medium with ethernet-like layer 2 behaviour, and this implies that the routers on both sides of a wireless link will want to ARP for each others addresses and they'd better be in the same broadcast domain for that to work. Though physically such a link might appear to be point-to-point, in fact from the point of view of the transmission layer it is not. It's just a degenerate case of a broadcast network that happens to have only two nodes on it. This is not PPP over a serial link or something like that...Maybe setup the OSPF to automatically take care of the routing and the subnets like this:
ether1 192.168.1.200/24 wlan1 10.0.10.100/32 | | | wlan1 10.0.10.101/32
Ok then, I'm surprised and curious to the point that I have to build such a setup in the lab some day and sniff the wireless link just to see what exactly is going on.Yes it does work
Yep, much better, IMHO. Though, if you think about it, just being able to add a third device proves my point that the link is not exactly point-to-point by nature, doesn't it?I do tend to use /29 for point to point most of the time as it can be handy to add another device to the subnet sometimes.