Good idea, however, I don’t see how I could use MAC addresses in this case because the modems are not in the same LAN as a sender, who is on the main LAN, and so that sender cannot successfully ARP to get a modem’s MAC address in order to send stuff to it.
I’m thinking that the Firebrick is perhaps less general because it is more pragmatic and less confusing. The xml elements are very much associated with practical nouns, verbs and adjectives and you don’t have to tell it how to do things just to do them. One example is failover, there’s nothing to work out, you just mention the usb ‘dongle’ NIC and it does the right thing. Which is nice.
I’m thinking that what I want to do is actually illogical, self-contradictory anyway. I want the LAN ranges of the 2-node lans (ie miniature broadcast domains) that the modems live on to all be different, from the Brick’s point of view, for routing, and the same from all the modems’ points of view so that they each see something that fits their expectations from their config. The ultimate would be for the Firebrick to DHCP-configure the modems so that they wouldn’t need IP configuration specified in their config files. But I don’t know if the modems can res upon to DHCP on the LAN side.
That is a non starter so we’re back to the question of what would happen with a lot of 2-node LANs that all have the same IP address range and thus clash inside the Firebrick for routing purposes. And there’s still also the problem of language of expressing destinations for addressing.
So which ever way, it seems truly impossible and so AA’s design and my own might well be the only options.