I am new to mt and have used ospf but not ion mt. my ospf knowledge is rusty and little. I am reworking my network and hope to use ospf to get some redundancy and load balancing.
All thew examples use privat networks for the router IP’s. I would like to know if there is a good reason to use private IP’s for all the router to router connections. They are all on my LAN so I could. Each router will be advertising (hopefully) a pool of public IP’s. Will this affect anything but traceroute? I have plenty of IP’s so that isn’t an issue.
Basically, all of these routers have an area that is a wireless cluster and I hope OSPF will be the answer to make my ever growing network more manageable and reliable. I also plan on using mt’s NAT to hand out those public IP’s in each router’s area.
I would like a traceroute come back clean to avoid any questions during troubleshooting. So, what do you think, pub or priv?
depends on the network design. if it is a standard routing design, public ip-adresses could be used.
if there were a lot of point-to-point links, proper subnetting is required, but often private IP addressed transfer-nets will be used.
sometimes large bridged networks are a choice too, but they are more difficult to handle.
i would prefer using public ip-addresses, locally extending by NAT, if necessary.
another way would be using tunnels, but this would go to far to explain here.
I already have a huge bridged network. That is what I am trying to do away with. It is easy to make something work from end to end but leads to a lot on excess traffic on the links as junk is broadcast across it.
I have a class C set apart just for doing the 252 subnets for the publics.
I plan on locally extending by public IP subnets only because many of my customers use VPN’s which don’t always behave well punching through a NAT. The network design looks something like this.
All of the peer points has a wireless point to mutipoint cluster behind them that I am going to use dhcp or PPPOE to hand out public IP’s to. All the links are licensed or unlicensed microwave hops. The reason I want to use OSPF is for redundancy as the GW ↔ peer_1 ↔ peer_2 have redundant links. Even if they didn’t I think that the benefit of having the router there segmenting traffic is worth the trouble and expense. Besides, it seems that after this major network layout change of inserting routers all over the place it makes it easier if every peer_? has the exact same topology.
I don’t want to use peer_1 or peer_2 for the GW as at some point more peer_? might be added in a location that makes that topology seem odd.