RE: NETLMM and PMIPv6 (RE: [netlmm] Consensus calltomakedraft-giaretta-netlmm-dt-protocol-02.txt a WG draft)

["Narayanan, Vidya" <[e-mail]>]

Putting the fact that George dislikes any kind of network-based mobility
aside :), I have to agree with a few of his points in the email below. 

Specifically, we should be having the discussion on what we can re-use
with a network-based mobility solution based on MIP6 vs. a total
re-design.

I will try to get a thread going on that topic shortly and hopefully, we
can make some progress in figuring this out and getting consensus. 

Vidya

> -----Original Message-----
> From: Tsirtsis, George [mailto:[e-mail]] 
> Sent: Friday, October 06, 2006 6:44 AM
> To: Julien Laganier; [e-mail]
> Cc: [e-mail]
> Subject: RE: NETLMM and PMIPv6 (RE: [netlmm] Consensus 
> calltomakedraft-giaretta-netlmm-dt-protocol-02.txt a WG draft)
> 
> It is well known that I do not like NETLMM at all, whether it 
> is done in NETLMM or MIPv6 WG. Putting my personal opinion 
> aside, however, I would like to say this:
> 
> Let's assume that the requirements for NETLMM and PMIPv6 are 
> virtually identical, which I think is a reasonable assumption.
> 
> NETLMM is currently attempting to address these requirements 
> by designing a protocol from scratch.
> 
> A PMIPv6 based design would presumably define the delta from 
> the standard MIPv6 solution required. The solution should 
> also be able to re-use at least some of the extensions 
> already defined in MIPv6 and maybe even allow for MIPv6 and 
> PMIPv6 mobiles and networks to coexist.
> Such assumptions should, however, be scrutinized if/when 
> PMIPv6 design is done, to make sure they still whole water.
> 
> Redesigning things from scratch is some times a good thing 
> but the new design has to be significantly better to 
> counteract the risk inherent with any new approach. Green 
> field designs are also best done by small groups of people 
> and they come with running code, rather than design by 
> committees (no offence to NETLMM).
> Extending an existing protocol is always an easier and more 
> palpable approach. Arguably, when the delta between the old 
> and the new approach is not that significant, extending an 
> existing protocol is highly preferable. This way, it is easy 
> for everyone to come up to speed with the changes and gives 
> everyone a warm fuzzy feeling about the whole thing (whether 
> such feeling is justified or not is another story).
> 
> I think proponents of PMIPv6 are making a mistake by pretending that
> PMIPv6 will somehow be different from NETLMM. Instead there 
> should be a discussion about the benefits of extending the 
> existing approach vs the perceived risks involved with 
> reinventing the wheel. If the majority of the NETLMM crowed 
> is willing to take the risk then I do not think you can do 
> much about it.
> 
> I still think that having two competing solutions in the IETF 
> is a bad idea, unless one of them becomes "experimental" so 
> that the effort is captured but it is clear that the IETF has 
> only fully defined one thing!
> 
> In fact what I think is that the IETF does not need a 
> solution to any of this but I guess I have said this already :-)
> 
> Regards
> George
> 
> 
> > -----Original Message-----
> > From: Julien Laganier [mailto:[e-mail]]
> > Sent: Friday, October 06, 2006 9:08 AM
> > To: [e-mail]
> > Cc: [e-mail]
> > Subject: Re: NETLMM and PMIPv6 (RE: [netlmm] Consensus call
> tomakedraft-
> > giaretta-netlmm-dt-protocol-02.txt a WG draft)
> > 
> > Hi Vidya,
> > 
> > On Friday 06 October 2006 00:51, Narayanan, Vidya
> > wrote:
> > > (Cross-posting to MIP6)
> > 
> > [ Perhaps this discussion should happen on a single
> >   ML? I am not though. ]
> > 
> > > All,
> > > I would like to understand how making this document a WG document 
> > > impacts any Proxy MIP6 solution that may be standardized at the 
> > > IETF. Jari's email on the
> > > MIP6 ML indicated that he would like to see this 
> discussion happen 
> > > before a decision is made about the NETLMM DT document.
> > 
> > My understanding of Jari's note was slighty different.
> > I understood that amongst other things he asked:
> > 
> >  - what are the requirements that are not fullfiled
> >    with today's solutions?
> > 
> >  - people interested in PMIP6 to talk to NetLMM WG
> >    *now* if their requirements are close to what
> >    NetLMM does.
> > 
> > > Personally, I think that the fundamental requirements are almost 
> > > identical in both cases.
> > 
> > Agree.
> > 
> > > I think summarizing the differences in requirements and 
> discussing 
> > > whether standardizing both solutions makes sense would be 
> worthwhile 
> > > at this time.
> > 
> > Discussing differences in requirements would definitely make sense.
> > 
> > However, because as you noted they seems to be almost 
> identical I do 
> > not think it makes sense to standardize both solutions -- that 
> > wouldn't be good from a standardization point of view.
> > 
> > Rather, those (supposedly few) requirements that are presently not 
> > fullfilled by a NetLMM solution should be worked on so that we 
> > satisfies all of them with a single protocol.
> > 
> > Best regards,
> > 
> > --julien
> > 
> > _______________________________________________
> > netlmm mailing list
> > [e-mail]
> > https://www1.ietf.org/mailman/listinfo/netlmm
> 
> _______________________________________________
> netlmm mailing list
> [e-mail]
> https://www1.ietf.org/mailman/listinfo/netlmm
> 

_______________________________________________
netlmm mailing list
[e-mail]
https://www1.ietf.org/mailman/listinfo/netlmm