We don’t perform any NAT operation. Let's first discuss the encapsulation: In SD-WAN there is an IP tunneling of some sort (there are several options but not relevant here). So we have an inner payload which can be v4 or v6 and goes end to end, the inner payload is encapsulated in an outer IP tunnel.
So what happens is the following. I make abstraction of inner payload SD-WAN CPE2 (v4) SD-WAN CPE1 (v4) <----------- v4 ---------> GW <-----------------v6 ------------- > SD-WAN CPE3(v6). Upstream traffic from SD-WAN CPE1 src-ip v4: A1, dst-IP v4 B1 ---------------> mapping < src-Ip v6: C1, dst-ip v6 D1 -----> Upstream traffic from SD-WAN CPE2 src-ip v4: A2, dst-IP v4 B1 ---------------> mapping < src-Ip v6: C1, dst-ip v6 D1 -----> So irrespective of which src-CPE 1 or 2 is sending to the v6 SD-WAN CPE3 will use the same src/dst v6 address With NAT64 even stateless the src-IP or src-port will be different to ensure you can map the reverse. This is also used to map private IPv4 to public Ipv4, etc and many other operations. On 06/11/2018, 14:53, "Gert Doering" <g...@space.net> wrote: Hi, On Tue, Nov 06, 2018 at 01:06:52AM +0000, Henderickx, Wim (Nokia - BE/Antwerp) wrote: > Even a NAT64 is not needed. What we do now is we go to a GW (lets call it like this for now) and the GW rewrites the outer header of the tunnel srcIP and dstIP from v4 to v6 or vice versa. It is kind of decap/encap operation on the outer header as if you terminate the tunnel and reinitiate the tunnel. No NAT64 required. > So how exactly is this different from a NAT64? "You take off the IPv4 header, put on an IPv6 header, and remember where the response packets needs to be sent to" If you preconfigure the mapping, it's still a NAT64 :-) Gert Doering -- NetMaster -- have you enabled IPv6 on something today...? SpaceNet AG Vorstand: Sebastian v. Bomhard, Michael Emmer Joseph-Dollinger-Bogen 14 Aufsichtsratsvors.: A. Grundner-Culemann D-80807 Muenchen HRB: 136055 (AG Muenchen) Tel: +49 (0)89/32356-444 USt-IdNr.: DE813185279 _______________________________________________ Int-area mailing list Int-area@ietf.org https://www.ietf.org/mailman/listinfo/int-area