Re: [nvrg-bof] First steps towards a charter

["Martin Stiemerling" <Stiemerling at nw.neclab.eu>]

Hi Roland, all,

> -----Original Message-----
> From: Roland Bless [mailto:bless at tm.uka.de]
> Sent: Wednesday, February 18, 2009 4:07 PM
> To: Martin Stiemerling
> Cc: nvrg at listserv.gwdg.de
> Subject: Re: [nvrg-bof] First steps towards a charter
> 
> Hi Martin,
> 
> Martin Stiemerling wrote:
> > It is time to get to a charter :)
> 
> Great to see some action here...
> 
> > Some issues I want to clarify before writing a charter proposal:
> > I. scope of the RG:
> >   1.  the RG focus solely on IP-based virtualization
> 
> Question: What do you mean by IP-based virtualization?
> Do you mean that the substrate is IP-based (i.e., the physical
> network on top of which the virtual networks are running)
> or that the network architecture running inside a virtual node is
> IP-based?

Solely IP-based is meant as:
- IP is the ultimate base
- which in turn means running everything over/within IP (e.g., IP-in-IP encap)
- only IP-based things in hosts and routers

I doubt that we need to be that narrow in the scope as there is also a lot of non-IP in the IETF (e.g., MPLS or better  MPLS-TP :)

> 
> >   2.  the RG does virtualization
> 
> ?? Not very clear to me. Do you mean: does work on
> virtualization enabling techniques in IP-based routers?

Router virtualization is one topic, but also on hosts and on the "wire".

> 
> >   3.  the focuses first on IP-based virtualization but also
> >       follows-up on (possibly) non-IP based  virtualization
> 
> See above.
> 
> > II. First things to tackle:
> > Joe proposed to:
> > - Virtual host architecture description
> >   Update RFC1122 / RFC1123
> >  - Virtual router architecture description
> >    Update RFC1812
> 
> That's a nice idea and I think one should first
> try to figure out the control functions and interfaces/protocols
> that we need to get things working in an interoperable fashion.
> Especially, if I think about setting up virtual node resources
> and setting up virtual links that cross domain boundaries.
> So out of scope would usually be the thing (OS, network stack)
> that runs inside the virtual node...

This does not exclude the work under II. but can be a parallel track. 


  Martin




stiemerling at nw.neclab.eu

NEC Laboratories Europe - Network Research Division
NEC Europe Limited | Registered Office: NEC House, 1 Victoria Road, London W3 6BL | Registered in England 2832014