Link Local IPv6 and NAT

[notaciscoguy]

I have a serious question which there seems to be little information, or consensus regarding. as I understand it, an IPv4 address in the 169.254/16 range is bad, and can't be used to route over the internet, even if NAT is being used by the router. there seems to be less agreement on whether this is true for IPv6 Link local addresses using the FE80 prefix. I understand there are undecillions of addresses and it is unnecessary to use NAT for an IPv6 addressing scema, however if a person used NAT64 technically it would be unnecessary to have to find a unique unicast address and prefix and only have to use locally unique IPv6 addresses. my question is that since the ULAs(as I understand the IPv6 version of private addresses in IPv4) have been long since deprecated, has the allowance been made for link locals to be used in a NAT64 environment?

I'm really not after a debate as to whether NAT is a good idea, nat kills voip, ipv6 is the devil, ipv4 is a dinosaur, etc, I am strictly looking to discuss the ability of link local addresses to be used in a NAT schema and a basic comparison between IPv4 and IPv6 link local address types. 

[Harold Ritter]

Hi,

I would first like to rectify something you mentioned. Unique Local Addresses (ULA) have not been deprecated. I think you were probably referring to the Site Local Addresses, which have indeed been deprecated via rfc3879.

Deprecating Site Local Addresses:

https://tools.ietf.org/html/rfc3879

ULA are indeed often compared to IPv4 RFC1918 addresses. The big difference is that ULA can be registered, unlike RFC1918 addresses.

Concerning your original question, I have not heard of any plans or discussions on whether NAT should be supported in conjunction with Link Local Addresses (LLA). On the other hand, there are lots of discussions and plans to support NAT in conjunction with ULAs.

Cisco Best Practices recommends to use Globally Unique Addresses (GUA), obviously without NAT.

Regards,

[notaciscoguy]

ahh, I was under the impression that site local and unique local were one and the same. It seems that I have some reading to do. thanks for the correction.