RE: New Version Notification for draft-li-6man-hbh-fwd-hdr-00.txt

["Pengshuping (Peng Shuping)" <[e-mail]>]

Hi Toreless, 

> -----Original Message-----
> From: Toerless Eckert [mailto:[e-mail]]
> Sent: Thursday, July 9, 2020 6:32 PM
> To: Pengshuping (Peng Shuping) <[e-mail]>
> Cc: Brian E Carpenter <[e-mail]>;
> [e-mail]; [e-mail]; Bob Hinden
> <[e-mail]>
> Subject: Re: New Version Notification for draft-li-6man-hbh-fwd-hdr-00.txt
> 
> On Thu, Jul 09, 2020 at 10:06:35AM +0000, Pengshuping (Peng Shuping)
> wrote:
> > We only used "forwarding plane" not "data plane" in this draft. In the
> context of this draft, from the perspective of the separation of the control
> plane and forwarding plane of a modern router, they could be taken as the
> same.
> 
> How about inband signaling such as
> 
> https://tools.ietf.org/internet-drafts/draft-martinsen-mmusic-malice-00.txt
> https://tools.ietf.org/html/draft-han-tsvwg-ip-transport-qos-03
> 
> Older experiments would punt to CPU, newer PoC would process this in the
> FPE ASIC.
> 
> Is this control plane if its punted and forwarding plane if its run in the FPE ?
> Do you think there should be a distinction between data plane and
> forwarding plane in terminology ?
> 
> I don't think we have an agreed upon definition of what forwarding plane
> means.
> Most likely it has been used as a stand-in for data plane, but that typically
> comes from times when there was no signalling processed by FPE.
> 
> To me, forwarding plane is anything running in an FPE.

In the context of this draft, we take the forwarding plane of a router as the fast path, while the control plane as the slow path.

For the fast path and slow path, we refer to the terminology defined in RFC6398. 

   o  Fast path: Hardware or Application-Specific Integrated Circuit
      (ASIC) processing path for packets.  This is the nominal
      processing path within a router for IP datagrams.

   o  Slow path: Software processing path for packets.  This is a sub-
      nominal processing path for packets that require special
      processing or differ from assumptions made in fast-path
      heuristics.

It seems that these terminologies caused some misunderstanding. We need to put some text in the draft to clarify. It will be great if we could also get some help.

Best regards, 
Shuping


> 
> Just saying.
> 
> Cheers
>     Toerless
> 
> > > > Best regards,
> > > > Shuping
> > > >
> > > >
> > > > Cheers
> > > >     Toerless
> > > >
> > > > >
> > > > > Quickstart is an Experimental RFC from 13 years ago. The RFC
> > > unfortunately doesn't appear to state the time limit for the
> > > experiment or how we would evaluate the experiment, but it does say:
> > > > >
> > > > > "As a result of these concerns,
> > > > > and as a result of the difficulties and seeming absence of
> > > > > motivation for routers, such as core routers to deploy
> > > > > Quick-Start, Quick-Start is being proposed as a mechanism that
> > > > > could be of use in controlled environments, and not as a
> > > > > mechanism that would be intended or appropriate for ubiquitous
> > > > > deployment in the global
> > > Internet."
> > > > >
> > > > > So to honest I don't think that Quickstart matters at all.
> > > > >
> > > > > Looking at the list of "Forwarding Plane" options, I see that it
> > > > > includes
> > > the RPL option. RFC6553 says:
> > > > >
> > > > > "The RPL Option provides a mechanism to include routing
> > > > > information with each datagram that a router forwards.  When
> > > > > receiving datagrams that include routing information, RPL
> > > > > routers process the routing information to help maintain the routing
> topology."
> > > > >
> > > > > I can't see how that is an action in the forwarding plane; it
> > > > > seems to
> > > describe a control plane action since it is updating the topology maps.
> > > > >
> > > > > On balance it seems to me that almost any conceivable HbH
> > > > > option,
> > > which by definition is intended for consumption by routers, is a
> > > control plane action, i.e. it distracts the router from its
> > > elementary job of looking for the longest match for the destination
> address in its forwarding table.
> > > > >
> > > > > (The Jumbo payload option is not really an exception to this,
> > > > > since as well as doing the longest match, the router must check
> > > > > that the outbound link supports the required MTU. Anyway, since
> > > > > the packet size by definition exceeds 65,575 bytes, nobody will
> > > > > really care about some overhead. Also, like most of the defined
> > > > > HbH options, it's a limited-domain option.)
> > > > >
> > > > > Regards
> > > > >     Brian
> > > > >
> > > > >
> > > > >
> > > > >
> > > > >
> > > > > ----------------------------------------------------------------
> > > > > ---- IETF IPv6 working group mailing list [e-mail]
> > > > > Administrative
> > > > > Requests: https://www.ietf.org/mailman/listinfo/ipv6
> > > > > ----------------------------------------------------------------
> > > > > ----
> > > >
> > > > --
> > > > ---
> > > > [e-mail]
> > >
> > > --
> > > ---
> > > [e-mail]
> 
> --
> ---
> [e-mail]