Re: [v6ops] I-D Action: draft-colitti-v6ops-host-addr-availability-01.txt

Hi,

> -----Original Message-----
> From: v6ops [mailto:v6ops-bounces@ietf.org] On Behalf Of Alexandru Petrescu
> Sent: Friday, July 31, 2015 6:10 AM
> To: v6ops@ietf.org
> Subject: Re: [v6ops] I-D Action: draft-colitti-v6ops-host-addr-availability-01.txt
> 
> 
> 
> Le 30/07/2015 19:16, Tore Anderson a écrit :
> > * Templin, Fred L
> >
> >> You can't give a device a /64 using SLAAC; SLAAC only allows for
> >> autoconfiguration of singleton addresses taken from an on-link /64
> >> prefix.
> >
> > Hi Fred,
> >
> > Actually, SLAAC works equally well with off-link prefixes. There is
> > no inherent requirement in SLAAC that prefixes from which the
> > addresses are auto-configured must be on-link; the L and A bits in an
> > RA's PIO are completely orthogonal.

RFC4861 says:

      "L              1-bit on-link flag.  When set, indicates that this
                     prefix can be used for on-link determination.  When
                     not set the advertisement makes no statement about
                     on-link or off-link properties of the prefix.  In
                     other words, if the L flag is not set a host MUST
                     NOT conclude that an address derived from the
                     prefix is off-link.  That is, it MUST NOT update a
                     previous indication that the address is on-link."

Meaning that "L=0" conveys no useful on/off link information about
the IPv6 prefix in the PIO. The node receiving the PIO therefore has
no way of knowing whether the prefix is on/off link and much less
whether the prefix is being "delegated" to the node. At least that
is per standard IPv6ND.

> I can agree.
> 
> >> If you want to give a device a /64, the only choices I am aware of
> >> are administrative configuration or DHCPv6 PD.
> >
> > True, although 3GPP networks is an exception. There a /64 prefix
> > advertised to the UE via standard RA/PIO with SLAAC can be treated
> > as if it was a delegated prefix.  RFC7278 depends on this.
> 
> IMHO it is wrong to treat it as a delegated prefix.  RFC7278 states
> clearly that the right way to get a delegated prefix is DHCPv6-PD.

Agree.

> The treatment 'as if' is what makes operators think that one /64 in RA
> is enough space for enough devices behind the UE.

Prefix delegation for any length /64 or shorter should be possible, yes.

Thanks - Fred
fred.l.templin@boeing.com

> A /64 on a cellular link can only accommodate one device.  Something may
> make think that the /64 is a delegated prefix that _could_ accommodate
> more than one device:
> - the cellular link is a ptp link.  As such the first-hop router in the
>    cellular network has a specific routing table entry containing the IP
>    address of the UE as next-hop (instead of a '*' if it were a shared
>    link) for that /64.  This makes that any address in that /64 is
>    reachable through UE's IP address.
> 
> However, cellular links continue to advance away from being ptp links
> and more be 'shared' links.  Recently 4G ptp links have got 48bit MAC
> addresses.  In the future there will be true Neighbor Discovery on these
> links, with '*' routes instead of next-hop routes.  At that point, the
> assumption that a /64 can accommodate 2^64 devices will no longer hold,
> because the first-hop router will issue NSs for the dst address, instead
> of issuing RSs for the IP address of the UE.
> 
> Alex
> 
> > Tore
> >
> > _______________________________________________ v6ops mailing list
> > v6ops@ietf.org https://www.ietf.org/mailman/listinfo/v6ops
> >
> 
> _______________________________________________
> v6ops mailing list
> v6ops@ietf.org
> https://www.ietf.org/mailman/listinfo/v6ops