In some email I received from Carson Gaspar, sie wrote: [...] > OK, I understand all the reasons why NAT is an abomination, in general. > What I _don't_ understand is why NAT is more evil on a bridge than on a > router. Can someone who espouses that religion enlighten me? My initial > analysis yields: > > - NAT frequently occurs between a real network and a virtual network (one > which exists nowhere, except in a routing table). The firewall having or > not having IP addresses seems to be irrelevant to the discussion. > > - NAT on a bridge makes 2 (or more) parts of a network segment have a > different view of the routing universe. I can see this being bizarre if > there are hosts on the local nets, but if the bridge has no local hosts, > this should not be an issue.
My personal preference is to only ever run one IP network "per wire". NAT on a bridge effectively means you have at least two and in getting to one from the other, the packets aren't routed, making it harder to determine where things are going. Darren
