Re: Other use cases for header insertion (was Re: Header Insertion and TI-FA)

[Stewart Bryant <[e-mail]>]

I am glad that we accept the NP is not SR!  Although it took a long time to get there, both in terms of the disintertwining and realisation that the concept was needed. It is a long time since I read it, but I am concerned about whether it has the full generality of the opaque instruction that MPLS has. Assuming that it has the next issue and it is one we are discussing here is the overhead of adding a new instruction to a packet in flight.

- Stewart


> On 13 May 2020, at 10:48, Robert Raszuk <[e-mail]> wrote:
> 
> Stewart,
> 
> Don't we already have this indirection with IPv6 today ? And I am not talking about SRv6 ... just in general least significant bits can encode any function of the ultimate destination. 
> 
> Such function can for example be of 20 bits and be used as context demux for packets. 
> 
> I am not talking about path engineering ... just flexibility an endpoint can expose already today with vanilla v6. 
> 
> Thx,
> R.
> 
> On Wed, May 13, 2020 at 11:42 AM Stewart Bryant <[e-mail] <mailto:[e-mail]>> wrote:
> 
> 
> > On 12 May 2020, at 21:29, Brian E Carpenter <[e-mail] <mailto:[e-mail]>> wrote:
> > 
> > On 13-May-20 02:56, Stewart Bryant wrote:
> >> It would also mean a new Ethertype, and I am sure that as we think about it other issues will come up.
> >> 
> >> … but I am sure that you are not the only one wondering the same thing.
> > 
> > Yes, there would be a lot of competition if there was a *serious* proposal
> > to add a new protocol number. Maybe we should make the first provision of
> > IPv7 that there should be a sub-version number.
> > 
> > Sigh.
> > 
> >   Brian
> > 
> 
> The big questions for me is the extent to which it should be self describing vis described by reference to context, and the extent to which it should be both the UNI and the network protocol.
> 
> Like it or hate it look how powerful MPLS became through the generality that a 20bit pointer to a procedure in the NPU and an arbitrary stack provided.
> 
> The self describing nature of IP is both an advantage and a significant limitation to its future extensibility.
> 
> So I am not sure if we need a sub-version or not, but we do need to look at building generality in from the beginning.
> 
> - Stewart
> 
> 
> 
> 
>