Re: improving WG operation (was Re: Voting (again))

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Hi Keith,

Keith, you have been advocating a model where the IETF would be stricter in allowing what work be taken up, in order to ensure that we can actually deliver. But I share the same opinion as John L that we should rather try to shape the IETF so that it can deliver what the world needs.


My primary criterion when arguing whether IETF should or should not take up a WG was always, in some sense, whether the Internet needed IETF to be involved in and supporting this effort. It involved both an assessment of how much harm would result from a botched design (in particular, a design that didn't respect the Internet environment and other protocols on the net), and of whether IETF could expend the resources necessary to manage the group and whether it could bring the necessary expertise to the table. It also involved an assessment of whether the proposed protocol would actually be of benefit to the Internet long-term.

All good criteria! I would probably add "assessment of whether lack of the protocol would be of
harm to the Internet long-term" (assuming the protocol falls within our scope, as you correctly
point out below). Here's an example: a protocol that is within IETF scope, but we suddenly
stop maintaining it to respond to changing requirements, or open it up to vendor extensions
without providing good abstractions that maintain interoperability.


Part of the reason why I believe so is that despite its problems, I think the IETF produces the best technology and highest quality. I want to use IETF multimedia, IETF network access control mechanisms, IETF security and not something else. This won't be easy of course, but I think we can do it. We are extremely good engineers and we've been able to produce scalable technology and useful, complexity reducing abstractions. Maybe time to apply some of that for our organization as well?


I don't think that IETF inherently produces the best technology and highest quality in every area of Internet protocol design. We cannot be good at everything. I may be dated in my awareness of our participants' expertise, but I doubt we have enough of the best designers of cryptographic algorithms, audio or video codecs, forward error correction codes, radio transmission methods, etc. There's a reason we leave valuable technical work to IEEE, 3GPP, W3C, etc. We have to specialize, as they do. The Internet is too vast and diverse for all of its technical work to be done by one organization. For me the selection criteria (in brief) have to do with whether the protocols in question impact the core Internet protocols or protocols traditionally developed in IETF, or whether the protocols in question need input from those with the most expertise from core or traditional IETF protocols. Those are fairly elastic criteria that cover a lot of ground, but not everything. For instance, we don't need to be involved much in B2B transaction processing as long as those guys can use existing protocols like TCP or HTTP in a way that works well for them and doesn't adversely impact the Internet. We might say things like "don't run everything over port 80" or "don't place too much faith in perimeter security" but we don't need to try to take over all of their protocol design.

I am in agreement with all what you say here. Just pointing out that even with
specialization, we may have quite a lot to do.


--Jari


_______________________________________________ Ietf@xxxxxxxx https://www1.ietf.org/mailman/listinfo/ietf

[Index of Archives]     [IETF Annoucements]     [IETF]     [IP Storage]     [Yosemite News]     [Linux SCTP]     [Linux Newbies]     [Fedora Users]