Hi David,

> -----Original Message-----
> From: Int-area [mailto:[email protected]] On Behalf Of Black, David
> Sent: Friday, May 20, 2016 10:31 PM
> To: Brian E Carpenter; Carlos Pignataro (cpignata)
> Cc: [email protected]
> Subject: Re: [Int-area] Call for adoption of draft-xu-intarea-ip-in-udp-03
> 
> -- As GRE/UDP draft shepherd
> 
> > > Do equivalent arguments apply also to
> > > <https://datatracker.ietf.org/doc/draft-
> > ietf-tsvwg-gre-in-udp-encap/>?
> 
> Not exactly ;-).
> 
> The motivating GRE in UDP encap use cases involves operator use of GRE for
> traffic management, as summarized in the draft introduction (Section 1 - the
> word "transit" is crucial):
> 
>    ...  the source port field in the UDP header may
>    be used to provide additional entropy that may be used for load
>    balancing GRE traffic in transit networks using existing Equal-Cost
>    Multi-Path (ECMP) mechanism.
> 
> An interesting consequence is that the GRE in UDP encap draft specifies that 
> the
> GRE and UDP tunnels MUST share endpoints - the GRE and UDP headers are
> applied at the same network node and are removed at the same network node.
> 
> So, yes, we did pay attention to use cases over in TSVWG ...
> 
> -- As an individual
> 
> Like, Brian, I don't understand the softwire motivation for IP-in-UDP tunnels,
> and hence oppose intarea WG adoption of this draft.
> 
> Xiaohu's response to Brian is unhelpful (IMHO):
> 
> > I don't understand what you meant. You don't know what the Softwire
> > service traffic is? If so, please refer to RFC5565.
> 
> That suggests to me that the draft is missing a brief overview of Softwires, 
> and
> probably also lacks an applicability statement  limiting the scope of this 
> draft to
> Softwires.

I had thought the Softwire use cases are well-known to the IETF folks and 
therefore there is no need for a brief overview of that technology.

> Section 8 of RFC 5565 on "Selecting a Tunneling Technology" is not helpful in
> explaining why IP-in-IP tunnels are wanted (which would then set up the

IP-in-IP tunnels are used for carrying IPv6 traffic over IPv4 networks (see the 
6rd usage as described in RFC5596 and RFC5969 and IPv6-in-IPv4 encapsulation 
format as described in https://tools.ietf.org/html/rfc4213#page-7) and IPv4 
traffic over IPv6 networks.

BTW, the IPv4-in-IPv6 and IPv6-in-IPv4 scenarios have been described in section 
3 of RFC5565. Section 8 of the same RFC just lays out the rationale for RFC5512 
which describes how to advertise tunnel capabilities via BGP. 

Best regards,
Xiaohu

> discussion in this draft about why IP-in-UDP-in-IP tunnels are wanted courtesy
> of deployed equipment ECMP limitations on IP-in-IP tunnels).
> 
> Thanks, --David
> 
> 
> > -----Original Message-----
> > From: Int-area [mailto:[email protected]] On Behalf Of Brian E
> > Carpenter
> > Sent: Friday, May 20, 2016 12:46 AM
> > To: Carlos Pignataro (cpignata)
> > Cc: [email protected]
> > Subject: Re: [Int-area] Call for adoption of
> > draft-xu-intarea-ip-in-udp-03
> >
> > On 20/05/2016 15:39, Carlos Pignataro (cpignata) wrote:
> > > Do equivalent arguments apply also to
> > > <https://datatracker.ietf.org/doc/draft-
> > ietf-tsvwg-gre-in-udp-encap/>?
> >
> > I don't know. I've never found GRE a very interesting topic so I've
> > never looked at that draft.
> >
> > > Was this one reviewed by Int-area?
> >
> > Not as far as I know, but it's still technically in TSVWG last call if
> > you have comments.
> >
> >     Brian
> >
> > >
> > > Thumb typed by Carlos Pignataro.
> > > Excuze typofraphicak errows
> > >
> > >> On May 19, 2016, at 21:17, Brian E Carpenter
> > >> <[email protected]>
> > wrote:
> > >>
> > >> As a famous physicist once said when the discovery of the muon was
> > announced, "Who ordered that?"
> > >>
> > >> In other words, I don't understand the use case that motivated this.
> > >> In the Introduction, I find:
> > >>
> > >>  "By encapsulating the Softwire
> > >>   service traffic into an UDP tunnel and using the source port of the
> > >>   UDP header as an entropy field, the existing load-balancing
> > >>   capability as mentioned above can be leveraged to provide fine-
> > >>   grained load-balancing of Softwire service traffic traffic over IP
> > >>   networks."
> > >>
> > >> What is meant by "the Softwire service traffic"? Does it just mean
> > >> a bunch of traffic from a single customer? Why does it all need a
> > >> common header for load balancing? The individual microflows in the
> > >> traffic will all get load-balanced anyway.
> > >>
> > >> I trust we are only talking about IPv6. If we're talking about
> > >> IPv4, we certainly shouldn't be developing new methods. If we're
> > >> talking about IPv6, we have the flow label for this (and it avoids
> > >> tricky problems like finding the transport header in the presence of
> extension headers).
> > >> That's already covered in RFC 6438 for any kind of tunnel, although
> > >> IP in IP seems a lot simpler than IP in UDP in IP.
> > >>
> > >> (If you really must do IPv4, IPv4 in IPv6 could still be
> > >> load-balanced as per RFC 6438, I guess.)
> > >>
> > >>   Brian
> > >>
> > >>
> > >>> On 20/05/2016 05:03, Wassim Haddad wrote:
> > >>> Dear all,
> > >>>
> > >>> The authors of draft-xu-intarea-ip-in-udp-03 (“Encapsulating IP in
> > >>> UDP”)
> > have requested that the working group adopt this work as a WG work item.
> > >>> So far, WG chairs have not seen widespread support and considering
> > >>> that
> > lack of opposition does not qualify as support, we’re starting a
> > working group adoption call until June 3rd.
> > >>>
> > >>> If you consider that the draft should be adopted as a WG work
> > >>> item, please
> > indicate the reason.
> > >>>
> > >>>
> > >>> Regards,
> > >>>
> > >>> Wassim & Juan Carlos
> > >>>
> > >>>
> > >>>
> > >>>
> > >>>
> > >>>
> > >>>
> > >>>
> > >>> _______________________________________________
> > >>> Int-area mailing list
> > >>> [email protected]
> > >>> https://www.ietf.org/mailman/listinfo/int-area
> > >>
> > >> _______________________________________________
> > >> Int-area mailing list
> > >> [email protected]
> > >> https://www.ietf.org/mailman/listinfo/int-area
> > >
> >
> > _______________________________________________
> > Int-area mailing list
> > [email protected]
> > https://www.ietf.org/mailman/listinfo/int-area
> _______________________________________________
> Int-area mailing list
> [email protected]
> https://www.ietf.org/mailman/listinfo/int-area
_______________________________________________
Int-area mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/int-area

Reply via email to