RE: [802.21] On the MIHEP initiative in IETF
Ajoy,
I guess we're somehow monopolizing the mailing list with this, but I suspect that this discussion would come up anyway very soon.
> > Ajoy,
> > please forgive me if I'm wrong, you seem to have the impression that
> > somebody is stating that CARD
> > cannot be the protocol selected for L3 transport of 802.21 services.
> >
> Ajoy-> This is not my point. IETF has no background in IEEE 802.21 work so
> I am not sure if we can make good choice in IETF about IEEE problem.
Stefano: that's why 802.21 needs to define clear requirements for IETF. At the same time, previous experiences in IETF indicated clearly that if another standardization body goes to IETF just looking for a stamp of approval of what they have done, IETF has always reacted quite negatively (and understandably so).
> During CARD development, IETF has extensively discussed pros and cons of
> various approaches required for information service protocol. So, I am not
> sure if starting the same discussion again will add much value. So, if we
> think we need an information service protocol then perhaps we should justify
> why existing protocol won't be able to address the problem or cannot be
> extended to address the problem.
Stefano: absolutely, and that's exactly what I stated previously. The chain is quite straightforward: 802.21 requirements and guidelines => IETF reviews and adopts requirements => IETF evaluates existing candidate protocols to determine if any can satisfy requirements (in the current form or with some extensions) => based on determination, either design a new protocol, extend an existing one, or define how an existing one can be used to satisfy the requirements.
> > Please keep in mind that nobody is saying that CARD cannot be the protocol
> > that will be selected.
>
> Ajoy-> I am not sure why IETF should define multiple similar protocols and
> then make a determination about which one to use. Sorry if I am not
> following your argument here. CARD was made a work item based IETF consensus.
Stefano: I guess I'm not being able to explain myself. I agree that we should not define multiple protocols and then choose. At the same time, just because CARD exists does not mean it should be adopted. IETF needs to verify it satisfies the requirements.
> > What is being said is that the evaluation of existing protocols must be
> > made against a set of requirements. Exactly because you worked in IETF you
> > should be well aware that IETF must always start from requirements and then
> > go through a process of evaluating existing protocols against the
> > requirements before making a decision. That is the only way to guarantee
> > that an optimal solution is found.
>
> Ajoy-> Note that IETF has extensively discussed information service protocol > earlier so I am not sure why there is need to do similar exercise again. If
> you think existing protocol would require change or enhancement please send
> an email to the authors so that proposed changes are discussed.
Stefano: this is not for me to do, and it is not to be discussed on this list before an appropriate location for work in IETF exists.
> If there is a need to start a working group then that can addressed as well.
Stefano: this is exactly why there has been a MIHEP BOF proposal. Anyway, I would not continue the discussion here, since it does not belong to this list.