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

Re: DNSEXT future

Kill it.
The rest of the message says as much, so if you are into the "bottom line" you can stop reading already. 

At 20:36 -0400 3/26/07, Ólafur Guðmundsson /DNSEXT co-chair wrote:


With the conclusion of the NSEC3 work there seems to be consensus
that DNSSEC work is done.
Oh, the DNSSEC work is not done, but what remains does not need a WG to discuss the protocol.
There's plenty of DNS work out there - implementations, performance, etc. But the protocol is cooked though and through.
Yes, DLV probably ought to be allowed to come to a conclusion first, but on the other hand, it's a shame we have to do it. 


All the remaining work in the group are close to completion,
with the exception of
	ECC DNSKEY specification.
I have nothing against this, but it has fallen so far off my radar, I can't see me wanting to volunteer any time to work on it. Will this be what puts DNSSEC over the top and get interest in its deployment? If not, I wouldn't keep a WG together for it. 


	advancing RFC's up the standards track.
Does anyone care about the latter? I have seen that no one wants to bother. This has become a strictly bureaucratic hurdle.
Draft Standards is supposed to mean that interoperability has been demonstrated by relying on the specifications. That's a noble goal. But the way things work in the DNS industry today we mostly just check to see if the new implementations work with BIND and that's enough. (At least one open source implementation touted that it's testing consisted of replaying queries and checking to see if the relies matched that of a BIND server's real responses to the same queries.) 

As long as no one demands Draft Standard compliance, it's not worth the effort.
What about DNAME? That's the only topic that caused any stir of engineering at the meeting last week. 


Both of these items have not had great interest or activity.
A potential new item for the working group is the provide input on the
processing of RFC2929bis templates while that process is stabilizing.
I'll note that RFC 2929bis has nothing to do with engineering and is a (much needed) bureaucratic process. It's not about engineering. 


For this reason we asked the meeting and our AD for guidance on what to do
with the working group. The three options are
	A. keep the working group open
	B. Put the group to sleep/hibernate
	C. Close the working group



Draft charter for DNSEXT in hyphenated state:
    The DNSEXT WG group will actively maintain the DNS protocol and is
    available for advancing DNS protocol related RFCs on the standard
    track, while defending against further enhancements of questionable
    value.
I really think it is a folly to maintain a dormant WG. (For those who don't know, Olafur and I go back 25 years back before I had to help him with the English of his thesis - I say this because: "Olafur, *hyphenated*? Is the group getting married? Like DNSOP-DNSEXT? Sorry folks, sorry for the sarcasm - yes it doesn't scale.) Oh, back to the cynical comment I *was* going to make... 

A "hibernating" anything cannot "actively" do anything.
Returning to being serious...DNSEXT was formed to engineer specific extensions for DNS, not be a bit bucket for the DNS protocol changes.
The DNS protocol is something that should not be undergoing continuous measurable improvement. It's infrastructure, at some point it is done. People who aren't DNS experts ought to be able to see a stable platform.
Yes there are features we want out of a naming system that are not in DNS. But there's no way we are going to get them features with what's running on port 53 today. "Super wildcards" - hat's off to the idea, but it can't happen with this protocol, we've plum tweaked the protocol out. 

Here's why we need to kill this outright:
1) Tell the ICANN folks we are done. They can count on the DNS as the basis of much of the work going on there. I'd be happy if we could engineer a better naming system (DNS2) that better fits the budding regulatory environment growing there.
2) Tell the folks that are not signing the root that we've done the best we can, it's in their hands now. The extensions are done.
3) Let folks know that the IETF is still an engineering organization with dynamic components, not yet a crusty bureaucracy that can't seem to let go of bailiwicks.
4) Give the impression that new features for a naming system are going to need a major overhaul of the naming system.
Engineering is meant to solve other people's problems. Fortunately for us, there are always problems somewhere.
PS - There's no reason to shut down the list. The PROVREG WG as shut down (effectively after the spring 03 IETF, officially about a year more as documents got passed through the gastro-editorial tract), the ietf-provreg list stayed open (and still is) and has seen the documents get to Draft Standard. The latter thanks to there being a desire to see it happen by participants. (And no WG as needed!) 

--
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
Edward Lewis                                                +1-571-434-5468
NeuStar

Sarcasm doesn't scale.

--
to unsubscribe send a message to namedroppers-request@ops.ietf.org with
the word 'unsubscribe' in a single line as the message text body.
archive: <http://ops.ietf.org/lists/namedroppers/>