Score:0

Choice of NAT64 prefix

br flag

I'm about to deploy NAT64 and I noticed that it's possible to use your own prefix instead of the "well-known prefix". Now I can imagine complex scenarios with multiple NAT64 gateways serving different prefixes where that would be necessary. But for a rather simple configuration with 2 VRRP routers, is there any advantage to using your own prefix?

Score:2
kr flag

The advantage of using your own prefix is that you can route it over the internet. That way users outside your own network can use it. Whether that is a feature depends on what you plan to do

OttoEisen avatar
br flag
Good point, hadn't thought about that. _Right now_ I don't need it, but to keep my options open, I'll probably use my own prefix. With v6 there are enough, after all :-) Thanks!
Score:1
in flag

One advantage of using a different prefix for each NAT64 is it's possible to gracefully "drain" a NAT64 for maintenance.

NAT64 (like NAT44) is a stateful process, so if you have two separate NAT64s serving the same prefix and you re-route client traffic from one to the other you will immediately break any active connections.

OTOH if your NAT64s use different prefixes and you use DNS to direct client traffic you can move new connections to a different NAT64 while allowing existing connections to complete on the existing NAT64, by waiting a while between switching the DNS and taking the NAT64 down for maintenance you can reduce the number of active connections you break.

OttoEisen avatar
br flag
You're right: right now if a router goes down SSH / RDP / etc. connections drop. But since I use VRRP for redundancy, there would have to be some kind of automatism. It's on my list and I know that there's conntrackd who seems to synchronize the NAT tables for v4. Maybe something similar is available for Jool (Linux NAT64 module).
mangohost

Post an answer

Most people don’t grasp that asking a lot of questions unlocks learning and improves interpersonal bonding. In Alison’s studies, for example, though people could accurately recall how many questions had been asked in their conversations, they didn’t intuit the link between questions and liking. Across four studies, in which participants were engaged in conversations themselves or read transcripts of others’ conversations, people tended not to realize that question asking would influence—or had influenced—the level of amity between the conversationalists.