x
Employment

IETF Routing Director Resigns

The routing group at the Internet Engineering Task Force (IETF), a major Internet standards body, is now in need of next-generation leadership.

Citing fundamental issues with the way the standards body goes about its business, the director of the routing group within the IETF resigned on Monday.

Rob Coltun, a consulting engineer for Movaz Networks Inc., resigned as the IETF's Routing Area Director in part because he feels a once simple process has become far too politically driven, as business agendas have overtaken what's best for the Internet's growth.

In the IETF, which is made up of academic and industry volunteers the world over, Area Directors are appointed for two-year terms by a nominations committee. The Area Directors in a particular working group are generally expected to know more about the combined work of the working groups in that area than anyone else. In the Routing Area, for example, an Area Director would be expected to have a broad knowledge of Inter-Domain Routing, Border Gateway Multicast Protocol, and several other specific routing technologies being discussed by the various working groups there.

Coltun resigned his post as Routing Area Director just four months prior to the end of his term and just a couple of weeks before the IETF's 52nd meeting, to be held December 9 to 14, 2001, in Salt Lake City, Utah. But what was most surprising about his resignation was the somewhat barbed critique he wrote to colleagues.

"Over the last couple of years I have found it exceedingly difficult to work within the IESG [Internet Engineering Steering Group, the group of Area Directors that is responsible for technical management of IETF activities and the Internet standards process]," Coltun wrote in a memo emailed to several IETF members.

He tells Light Reading that he was thinking of resigning earlier in the year, but stayed on in support of his co-chair, Abha Abuja, who died on October 21. Coltun says it was Abuja who encouraged him to air out his views publicly.

In his memo, Coltun praises his colleagues in the IESG as "well-meaning" and "intelligent," but he worries that the IESG has "lost the big picture." His memo says his enthusiasm for the IESG has dampened over the years as the IETF's meetings have ballooned in size and have become weighed down with corporate politics. "The 'just do the right thing' principal that seemed to be there for so long is hard to come by," he writes.

"In my opinion, part of the problem is that the IESG's model of 'management by body' worked when a few hundred people showed up for IETF meetings and the scope and quantity of our work was much less, but this model is now antiquated.

"IETF meetings must be smaller to be more productive. One suggestion is to formalize the private meetings that happen between authors [of proposed technical standards] and co-chairs and provide tutorials and updates for the masses.

"The conversations of whether or not particular work should be taken on by the IETF are very important but should last at most a few months, not years.

"It is no coincidence that as the factors that are driving the current phase of the evolution of the Internet change, restructuring of the management of the IETF becomes inevitable."

Coltun, of course, is not the first person to complain about how large in size and loosely ordered many IETF working groups have become. Since the working groups don't take a vote to decide issues, the group has to come to what seems to be a general consensus, without any way of actually measuring that consensus.

This medieval structure has led to some carefully crafted proposals, but mostly to excessive delays on deciding whether some ideas will go on to be accepted practice amongst router makers and other equipment vendors (see MPLS: Keeping it Real).

In fact, stories abound of some working groups determining consensus by how loud people are humming in protest to a proposal. "It's usually not a very loud hum," says Coltun.

Some vendors have been known to throw off the informal working group consensus by crowding a meeting with their own staff in order to slow or stop a proposal by out-humming their rivals.

"It's so childish it's unbelievable," says one senior engineer at a European telecom equipment company. "You're not chosen for your technical knowledge. You're chosen for your singing voice."

The IETF's own informational guide acknowledges the problem that comes with its openness: "And, if you think about it, how could you have 'voting' in a group that anyone can join, and when it's impossible to count the participants?"

IETF meetings used to be based on developing technical ways to accomodate the Internet's growth. Some say the process is now clouded by the growing markets for Internet technology. Millions of dollars are spent implementing a standard once it's adopted because no vendor can afford to have a product that doesn't work well with anything else in a network.

"Frankly, I give the IETF a lot of credit for the standards they've put in place over the years," says an IETF member who works for a Massachusetts-based optical networking startup. "Now the situation has become such that if a vendor lobbies and successfully gets a certain standard adopted, it could make millions of dollars. Very early on in the Internet's development, the IETF was just what was needed. But times have changed."

"There are a lot of conversations going on in the IESG about how it should reconstitute itself," says Coltun, who is taking time away from standards work to help Movaz develop products. "Unfortunately, they tend to be more procedural than anything addressing the big picture."

A call to IETF Executive Director Steve Coya was not returned by press time.

- Phil Harvey, Senior Editor, Light Reading
http://www.lightreading.com
random-reading 12/4/2012 | 7:30:38 PM
re: IETF Routing Director Resigns Well said, John!

Because there are so much at stake in the IETF
process, those with resources (money, people)
will do ANYTHING to steer the process to their
advantage. Serious work was not being done in
the haste of "inventing" "new" stuff. Anyone
care to count how many I-Ds produced on MPLS?

Perhaps with the burst of "Internet bubble", and
with much less at stake, things will settle.

RR

mu-law 12/4/2012 | 7:30:44 PM
re: IETF Routing Director Resigns "I see what you're upset about. I think I've fixed the problem by being more specific regarding the AD's role in the IETF."

Yes, but there is such a thing as too much caution; don't confuse accuracy with suppression of the healthy cynicism and subjectives ([email protected]%#$) that this piece warrants.

I have to agree with the other posts on the whole. Problems with the IETF "culture" are both manifold and well understood, it would seem.

I should offer that the "entropic" organization strategy of IETF has caused it to nucleate in a way that parallels the mindset and attributes of its very human membership, consisting of equal parts vendor and operation, from what I can tell. We have seen the enemy, and it is us.

If proposals for new technology are not subject to the same scrutiny as demanded by a simple refereed journal, then results in kind should not be expected. IETF is not (and should not) be the place to look for "good science", but if new work is to be undertaken or proposed, let there be systematically proven technical justification; this is where the "barrier to entry" should be set.

The question (IMO) is which inmates are to be in charge of the asylum, and how are they are to be chosen... Ideas?
skeptic 12/4/2012 | 7:30:49 PM
re: IETF Routing Director Resigns Many of the drafts and standards written by these front end men are bogus, sub-standard and with very little thought behind them.
------------

Then they should not be considered or immediatly
thrown out without any consideration at all.
But that doesn't happen. And part of the reason
it doesn't happen is that the front-end men
and their drafts are being protected by people
higher up in the IETF who are untouchable.

Your mention of BGP-MPLS VPN's is a good
example. It should have been stopped, but
look whose name is on it. And note as well
that its not even a standard, its "informational".
But all cisco wanted was the magic word "RFC"
which makes it (to the public) an open standard no
matter what the reality of the situation.

Its no different than BGP, where a tight group
of people have a vested interest in making the
"standard" utterly useless (except as a seal
of approval that their equipment is "open
standards".)

So why did it happen. Its not hordes from big
companies that did the deed on that one. It
was a very well-connected set of people high
up in the IETF (including Y. Rekhter) who can
do almost anything they want. And if anyone
dares to "upset" them, they will find themselves
on the outside looking in.

Thats the real lunacy of Rob's comments. The
problem at IETF isn't necessarly the size of the
organization. Its that the "insider" nature of
the organization and the financial rewards that
go along with being on the inside have corrupted
the process at every level.

froggy 12/4/2012 | 7:30:52 PM
re: IETF Routing Director Resigns Now it's business as usual and IETF is just becoming a tool for the most influential groups to protect the quasi-monopoly they have on routing.

If you still have illusions about the free Internet promoting values of freedom and innovation, you have not been at the IETF in quite a while
John Casper 12/4/2012 | 7:30:53 PM
re: IETF Routing Director Resigns I know Rob -he is a well meaning young man who would like to work in a certain way.

The problem is that the idealist view of what IETF should be, and what it has become now started diverging when big vendors like Cisco, Junipers, Nortel created an IETF attack force of their own in mid 90s. Many of the drafts and standards written by these front end men are bogus, sub-standard and with very little thought behind them. Many of the standards developed by IETF are outrightly dangerous to networking (BGP/MPLS VPNs widely promoted these days is a prime example.). Some of these guys are not only responsible for crafting bad standards but also defending them at the expense of other worthwhile ideas using the commercial muscle of their big company. The whole team of people in big three or four vendors justify their job by amount of crap they can publish in terms of drafts every month. They don't have quality in their work. If they are not able to publish in IEEE and ACM journals where serious work is published with suitable peer reviews? All their interst is in publishing drafts for solving imaginary problems.

There is a lot of total rubbish being published as Ineternet Drafts by guys who know little about the real networks. IPv6, IP Mobility, MPLS, Multicast Routing are the prime examples of areas with largely undeployable junks published as standards. Many aspects are not well thoughtout when first RFC is published then the second corrects the error, then the chain begins one after the other each one correcting the loopwhole of the previous one. The question is why these things can't be previously thought out with so much of the collective brain power in the group?

On the good note, there are some of the best brains in networking also are there in IETF. But these guys are not politicized, many of them come from universities and reserach institution with no commercial axe to grind. These are the people who produce quality work, represent the best in IETF, and obviously cannot comptete with attack force of Big 3. They have a real job other than writing useless drafts. They will only submit a well thought out idea and proposal.

If we really wish to improve process, the size of IETF meeting should be reduced. No company should be allowed to send more than three reps to a particular working group (including the working chair if that belongs to a particular company).

With all its short comings, IETF meeting is still a fun place. But I would like to see less faces from big 3 or 4 who seem to dominate just about every group.


oo John




dre 12/4/2012 | 7:30:54 PM
re: IETF Routing Director Resigns
Niels Ferguson and Bruce Schneier did an
analysis of IPsec and had some interesting
insights into "the IETF problem" and a rough
solution. This paper is available at the
Counterpane Lab's research site:
http://www.counterpane.com/ips...


Finally, it has become clear that most of the
blame for this state of affairs lies not with
the people that worked on IPsec, but with the
process used to develop it. Committee designs
are bad enough when all you try to do is to get
something to work. The IPsec experience has
demonstrated that a committee design process is
wholly incapable of creating a useful design for
a security system. In our opinion, there is a
fundamental conflict between the committee process
and the property of security systems being only
as strong as their weakest link. Therefore, we
think that continuing the existing process and
fixing IPsec based on various comments in bound
to fail. The NIST-run process for selecing the
new AES might serve as an example of an alternate
process for developing and standardizing a security system.
We strongly discourage the use of IPsec in
its current form for protection of any kind
of valuable information, and hope that future
iterations of the design will be improved.


What Niels Ferguson and Bruce Schneier are
talking aout is the Advanced Encryption
Standard (AES). AES is the replacement for
DES. For more information, check out:

http://csrc.nist.gov/encryptio...
http://www.nist.gov/public_aff...

It is clear that the same issues are happening
elsewhere in the IETF. These are fixable
problems. The IETF does not have to become the
ITU, or destroy its credibility completely. I
think something can be made good out of all of
this, but I do hope it happens sooner than later.

-dre
skeptic 12/4/2012 | 7:30:56 PM
re: IETF Routing Director Resigns
What IETF does Rob Coluton go to? There is
almost nothing decided in the public meetings
anymore. Its decided away from the conference
by invitation only. And nowhere is this more
true in than in the routing area which he was
responsible for. As far as consensus in the
room goes, the people running the meeting get
to decide when thats done and they usually dont
do it unless it goes along with what they want.

They don't even use the mailing lists anymore
for anything real. And has it made the routing
are any more productive? No.

And you can complain about "industry pressure"
all you like, but as long as area directors
and chairs are openly using their IETF positions
in the job market to line their pockets with
money, what does anyone expect? They pay
the money based on getting something back.

And if the IETF goes to an even more select model
of membership, all its going to do is line the
pockets of the people (and their freinds) who
get made "insiders". And based on past
experience, some of the most disruptive and
useless people at IETF are going to get made
automatic insiders.

And as far as problems at the IETF go, the
real problem isn't the size of the meetings,
its that the upper reaches of the IETF are
dominated by a circle of people who don't care
about much of anything that matters. Van
Jacobson and friends can run the diffserv working
group into the ground without those people
saying a word. They are off being "important"
(while doing nothing).

The processes at IETF are already informal
and arbitrary. So what is holding things back?
If its the IESG itself, going to a smaller IETF
isn't going to fix it.

If the AD's and chairs can't say "NO" with the
power they have now, how is anything supposed
to get better with fewer people in the room?

dljvjbsl 12/4/2012 | 7:30:59 PM
re: IETF Routing Director Resigns The area director resigns and suggests that private meetings be formalized and that the masses be presented with tutorials on the decisions.

Firstly, isn't this the way the IETF works now? The are definite 'in-crowds' who make decisions in private and then present the working groups with joint drafts. Inputs from people outside of the in-group are poliitely or as I ahve seen in some cases impolitely, ignored. It only takes a short amount of time for someone to learn this and play along with the standard practice.

Secondly the draconian insistence on IPR licensing acts as a dramatic hindrance to anyone who wishes to comment. Someone makes an unguarded comment at an IETF meeting or in a posting to the mailing lsit. His company's IPR is now up for licensing by anyone. It is very wise to keep silent at an IETF meeting unless there is a clear strategy from your company to indicate doing so.

The practices above tend to make the IETF political. Large players who can make de facto standards are the ones who can demand entry to the in-groups. These large players have strategies in the working groups that make the IPR provisions of the IETF moot. I have seen them simply claim IPR rights over all the work done by a working group after the fact. Who is going to say no to them?

The recommendations by the resigning area director far from making the IETF an open place would formalize these arrangements. Perhaps having the real procedures for the IETF formalized and made public would be a good thing. However it would not make it anymore open than it is now.
DCITDave 12/4/2012 | 7:31:00 PM
re: IETF Routing Director Resigns Thanks for your post.

I see what you're upset about. I think I've fixed the problem by being more specific regarding the AD's role in the IETF.

Please hum loudly if you have any further questions or concerns.

ph
hardcore router 12/4/2012 | 7:31:01 PM
re: IETF Routing Director Resigns the best man.
Pseudopersonality 12/4/2012 | 7:31:08 PM
re: IETF Routing Director Resigns This is the 21st century right?

This group is developing the next generation internet right?

Humming? hmmmmmmmmmmmmmm

Learn something every day but this takes the cake.



AlexZ 12/4/2012 | 7:31:09 PM
re: IETF Routing Director Resigns There's no single and perfect solution as there's
no perfect democracy. However, there still are
many things that can be done.

First, regarding meetings, votings and hummings.
We almost forgot (and few people read IETF
guideliness these days) that the main means of
discussions in the WGs is the mailing list.
Votings and hummings are required to be verified
on the mailing lists, even though I do not recall
this happening lately. When objecting on a mailing
list, it is a common practice to bring reasons
than say "I just don't like it" and, surprisingly,
there are not so many active WG members on the
mailing lists as there are people in the meeting
room.

Second, regarding more productive work within a
large community. IETF guidelines already provide
the notion of a design group which can work on a
specific mechanism, achieve local consensus and
bring the results to the WG. The WG still needs
to approve the results, however.

Third, regarding the subject of votings and
hummings, usually competing proposals. IETF rules
suggest that finding a common ground should be
encouraged, but it is ok to have multiple
mechanisms at once and let the time and market
choose which one should live.

There's many that can be added here. In short,
we need to start following the IETF guidelines
as much as possible to begin with.

Regarding the sad departure of Rob, I would
consider this as a very chance to initiate a wider
discussion about the situation in the IETF and
actually change (or at least start changing)
things to the better. On the other hand, I fear
that it might unfortunately take many bright
people like Rob before this happens.
before
veemee 12/4/2012 | 7:31:10 PM
re: IETF Routing Director Resigns Though I agree IETF meetings have become politically motivated, and the vote and hum methods are are screwed up. But what could be a possible solution for this, even while allowing the IETF to stay as open as it is at present?

IETF being an open group(voting doesnt make sense)
mu-law 12/4/2012 | 7:31:10 PM
re: IETF Routing Director Resigns "In the IETF, an Area Director manages a working group"

No, a WG Chair manages a working group. Area Directors lead the standards process for all groups in an area.

"(such as routing, switching, or security)"

This is just uninformed diarrhea... these are not working groups. "Switching" is neither an area nor a working group, and by definition, it can't be because it is a detail of implementation does not involve IP.

The sentient future of transport will be determined by what happens in IETF... it would behoove you to understand it especially if you insist on reporting about it.

All it takes is time, and all materials are free; this stuff is right on the IETF web page. If you're in a hurry, start with RFC3160, and then 2026:

http://www.ietf.org/rfc/rfc316...
http://www.ietf.org/rfc/rfc202...
Belzebutt 12/4/2012 | 7:31:13 PM
re: IETF Routing Director Resigns Weren't IETF people always looking down on the ITU as a huge bloated organization, corporate-driven, where company agendas drive everything? Now people are looking down at the IETF and pointing to the ITU as a good example?

I guess it's the end of an era, if you want to go back to this informal decision-making, you should head along to Internet 2 or whatever next-gen thing academia is working on right now.
froggy 12/4/2012 | 7:31:13 PM
re: IETF Routing Director Resigns IETF cannot continued to be managed by "well-meaning" researchers, all of them being now directly or indirectly on Cisco's payroll.

Time for a real shake-up of that medieval organization, driven by arrogant and aging engineers whose only intellectual contribution was BGP... that's a scary thought !!!
HOME
Sign In
SEARCH
CLOSE
MORE
CLOSE