[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [shim6] about draft-ietf-v6ops-cpe-simple-security-07



On 2009-07-29 02:01, Iljitsch van Beijnum wrote:
> [to v6ops and shim6, prune as necessary]

Pruned to v6ops

...
> 3.2.4.  6to4 Tunnels
> 
>    Typical dual-stack IPv4/IPv6 residential gateways use private IPv4
>    address ranges and network address/port translation on a single IPv4
>    address assigned by the service provider.  The use of private
>    addresses prevents interior hosts from using 6to4 [RFC3068] tunnels.

> 
> Usually the private addresses are on the inside and a public on the
> outside (which will presumably start to change as IPv4 depletes but will
> never completely go away) so 6to4 can work just fine in those cases.

Well, the text is unclear. If the hosts behind the CPE try to use
6to4 [RFC3056] tunnels (with or without the anycast extension [RFC3068]),
they will fail if the LAN is using RFC1918 and NAT. But if the CPE
itself acts as a 6to4 gateway as defined and intended in [RFC3056],
the hosts will see *native* IPv6 and everything will work fine.
Except for one case:

     The use of private addresses by an ISP "outside" the CPE
     would mske it impossible for the CPE to operate
     as a 6to4 [RFC3056] gateway.

IMHO every IPv6 CPE SHOULD contain a 6to4 gateway, but that belongs
in another draft.

    Brian


> 
>    The use of Level 3 Multihoming Shim
>    Protocol for IPv6 (SHIM6) [I-D.ietf-shim6-proto] as a site multi-
>    homing solution is not generally compatible with IPv6 simple
>    security.
> 
> (Now RFC 5533.) Why? Assuming that there are two CPEs with each an ISP
> link and advertising separate address ranges, then if the host sends
> normal packets over CPE A and then when the ISP connected to CPE A
> fails, the host can send protocol 140 signaling and encapsulated data
> packets to CPE B. This should work. An optimization would be to allow
> incoming protocol 140 whenever there is communication between an
> internal and external host already, whatever the type. (Without that,
> the shim6 signaling fails when the first host initiates it and only
> succeeds if the second host initiates it soon enough after.)
> _______________________________________________
> shim6 mailing list
> shim6@ietf.org
> https://www.ietf.org/mailman/listinfo/shim6
>