On 07/11/2016 08:52 AM, Brent Smith wrote:
> What is common practice for firewalling with IPv6?
>

The approach I have taken is:

1. Define IPv6 address blocks to parallel IPv4 address space, including 
private address space.
2. Apply the same TCP/UDP port rules in IPv6 as I currently have in IPv4.
3. Where NAT is being used for IPv4, and it is expected to be providing 
security to the hosts behind it, I set up a stateful firewall, reject or 
deny incoming connections, and allow outgoing connections. For 
Mikrotik/IPtables terminology, IIRC, that would be rejecting/dropping 
packets for connections in the invalid and new states from the outside, 
and allowing packets for connections in the related and established states.
4. I generally do not block ping even with IPv4. If it becomes a 
problem, I rate limit it. Remember that for IPv6, ICMP is used for 
discovering path size. If that type of ICMP packet is blocked, you can 
have weird problems where you can reach the host over IPv6, but trying 
to transfer data over that path doesn't work. (Especially if you have 
something with a smaller MTU in the path, like PPPoE.)


I really have not tackled the issue of tunnels. But if a host wants to 
set up some kind of a tunnel, then I figure it's up to them to decide 
what incoming traffic they may want to block.

--
Erik Andersen
_______________________________________________
Mikrotik-users mailing list
[email protected]
http://lists.wispa.org/mailman/listinfo/mikrotik-users

Reply via email to