6to4; ISATAP interfaces configured over same IPv4 address (was: Re: FYI Isatap implementations info)

[Fred Templin <ftemplin@iprg.nokia.com>]

Hi Karen,

I'll focus on just the 6to4/isatap stuff for now if you don't mind:

Karen E. Nielsen (AH/TED) wrote:

> Fred Templin wrote:
>  
>
> > If a 6to4 interface and ISATAP interface are configured over the same
> > IPv4 address, it would have to be a global IPv4 address assigned to an
> > interface connected to the global Internet due to 6to4 
> > requirements. In
> > that case, I believe the correct behavior should be for the 
> > 6to4 interface
> > to take precedence for packets received with an on-link 6to4 prefix in
> > the IPv6 destination, and for the ISATAP interface to take precedence
> > for all others. Comments?
> >    
>
> When I say takes precedence over I (still) mean the following:
> First Priority:	If the prefix of the source address in the inner 
>     IPv6 header of the encapsulated packet matches a prefix of the 
>     ISATAP interface, the packet is processed by this tunnel interface.

This would seem like the correct natural choice, since it seems consistent
with the notion of "intra-site". Again, I assume by this you are intending
to include link-locals? (Perhaps that's *all* you are intending to include?)

> Second Priority: If the prefix of either the source or destination address 
>      in the inner IPv6 header is a 6to4 prefix, the packet 
>      is processed by the 6to4 tunnel interface.

Let's take a closer look at the cases; please tell me if I mis-represent 
anything:

 1) If both the source and destination are 6to4, then the packet is 
originating
    and terminating within the 6to4 domain. So, use of the 6to4 
interface would
    seem natural and appropriate. (Note that the destination might not 
belong to
    *this* 6to4 router in which case L3 might forward (redirect?) the 
packet to
    the correct 6to4 router - but, this is irrelevant for L2.5 
decapsulation.)

 2) If the destination is 6to4, but the source is not 6to4, then the 
packet is
    originating from behind a 6to4 relay router and terminating within the
    6to4 domain. Again, the 6to4 interface seems the natural choice.

 3) If the source is 6to4, but the destination is not 6to4, then the packet
    is originating from within the 6to4 domain and will eventually
    terminate within the native IPv6 Internet. To do this, it will have
    to transition from the 6to4 domain to the native IPv6 Internet via
    a relay router which may/may not occur on *this* 6to4 router.

    Is the 6to4 interface the correct interface to decapsulate for this 
case?
    It seems somewhat less obvious to me, but still a better choice than
    the ISATAP interface when the source prefix is not on-link.

> **Here again it should be emphasized that these rules are not intended to
> apply to the Isatap router-to-router tunnelling situation.**

I am fine with leaving this aspect out from the current discussion.

> I am not 100 % sure what you mean by an on-link 6to4 prefix.

I think I may have been stuck with a mis-conception that 6to4 interfaces
should only decapsulate packets with an IPv6 destination prefix of
2002:V4ADDR::/48, where "V4ADDR" is assigned to one of the
node's IPv4 interfaces.

> I can imagine that you refer to one of the following 2 situations:
> (I)
> _ _ _ _                         - - - - - - - - - - - - - - 
> |       |     ___               |  Global Internet/exterior  |         
> |  6to4 |-v6-/ R \              | _ _ _ _ _                  |
> |  Site |    \___/-globalV4ADDR-|   Intra  |                 |
> |_ _ _ _|                       |   -site  |                 | 
>                                 - - - - - - - - - - - - - - 
>
> (II)                     - - - - - - - - - - 
>     ___               |  Global Internet/   |         
>    / R \              | _ _ _ _ _   exterior|
>    \___/-globalV4ADDR-|   Intra   |         |                   
>                       |   -site   |         | 
>                       |   of 6to4 |         |   
>                       |  prefix   |         | 
>                        - - - - - - - - - -
>
> In both situations a packet with the mentioned inner source address
> should be from within the intra-site and should be processed by the Isatap
> interface and not the 6to4 interface turned towards the exterior.
>
> In (I), then in our 6to4 implementation, the packet with the mentioned source address,
> will be discarded by the 6to4 tunnel interfaces as we rely on trusted relay routers.
> It wil not be accepted as a packet from a 6to4 border router since the v6 and the v4 sources
> don't add up.
>
> In (II)
> The packet is a packet coming from and destined for the Intra-site and should
> be processed by the intra-site router interface (i.e. the Isatap
> interface).  Again
> the 6to4 interface would discard the packet,
> further It should be redirected but thats another matter.

I'm not prepared to comment on this in detail since I'm not entirely sure
what you are referring to by "the mentioned inner source address", and
your examples might have been motivated by the misconception I mentioned
above. Perhaps this discussion is obviated by the examination of the cases
I provided earlier?

Fred
ftemplin@iprg.nokia.com