RE: [PWE3] Re: I-D ACTION:draft-ash-avt-ecrtp-over-mpls-protocol- 00.txt

[Sasha Vainshtein <Sasha at AXERRA.com>]

Curtis and Stewart,
Thank you for the detailed explanation.

To handle the new registry in a "safe" way, I would suggest to specify
that only values 0 to 3 can be used, with 0 assigned to unindentified 
payload type, 1 - to identified payload type and leave 2 and 3 unassigned.

In parallel IANA should declare values 1 to 3 (currently unassigned)
as reserved in the version numbers registry to prevent any attempt to 
assign these values in future.

Does it make sense?

With best regards,
                                   Sasha Vainshtein
email:     sasha@axerra.com <mailto:sasha@axerra.com> 
tel:       +972-3-7659993 (office)
           +972-8-9254948 (res.)
           +972-58-674833 (cell.)
 


> -----Original Message-----
> From: Curtis Villamizar [mailto:curtis@workhorse.fictitious.org]
> Sent: Monday, February 16, 2004 5:08 PM
> To: stbryant@cisco.com
> Cc: Sasha Vainshtein; avt@ietf.org; pwe3@ietf.org; Curtis Villamizar
> Subject: Re: [PWE3] Re: I-D 
> ACTION:draft-ash-avt-ecrtp-over-mpls-protocol- 00.txt 
> 
> 
> 
> In message <4030C043.3030509@cisco.com>, Stewart Bryant writes:
> > 
> > 
> > 4 and 6 will be preallocated, so that they do not get trodden on.
> > 
> > We need to find out if there is any use of 5, 7, 8 and 9. I.e.
> > are there any routers in the Internet that carry IP versions
> > 5, 7, 8 or 9 over MPLS. If there are, we need to include them
> > in the registry as pre-allocated values. If not, then a
> > new implementation would need to operate under the rules
> > proposed below (i.e. run under 0, 1 or their own value).
> > 
> > The IESG would advise IANA on the usage of this registry.
> > 
> > - Stewart
> 
> 
> For all pracical purposes 7, 8 and 9 are not used at all, over MPLS or
> otherwise.  These are left overs from the competing proposals to
> replace IPv4.  The decision was made to go with IPv6.  Since then none
> of them have been used.  It would be reasonable for IESG/IANA to
> declare these values as having been temporarily assigned for protocol
> experiments which are long ago over and available for reclaim, but
> apparently they are choosing not to do so.  There is (or was at some
> point) some use of 5 (ST-II).  Use of ST-II may have disappeared, but
> I'm uncertain of that.
> 
> IESG does have a very good point in that what these are being used for
> is not IP version numbers where the other 20 bytes are also an IP
> header.  That IPv4 and IPv6 values are reserved in the new registry to
> allow the largest single contributor of traffic to be placed in MPLS
> without a shim is good, and not breaking very widespread current usage
> is good.  If you reserve any number for "other IP version" then an IP
> packet with 5, 7, 8 and 9 or any new assignments of IP versions could
> be carried with a shim.
> 
> Curtis
> 

_______________________________________________
pwe3 mailing list
pwe3@ietf.org
https://www1.ietf.org/mailman/listinfo/pwe3