On 12/04/2017 00:03, Dave Dolson wrote:
Regarding "benefits", these devices clearly have perceived benefits to those
who deploy them. My view is that the document should explain that perspective
for readers who lack the operator perspective.
The intent was not to mandate or recommend deployments.
I'm not against facts, which was one of the reasons that I co-authored
RFC 3234. But in that document we worked hard to be as neutral as
possible. Revisiting the topic 15 years later seems like a good idea,
and focussing on transport-layer snooping is fine. But the tone would
need to be very neutral and the draft would need to present pros
and cons for this to be published by the IETF, I think. The topic is
one of the major ongoing tussles and the IETF needs to be completely
"This document advocates for transport connections to be measured and
managed by the network..."
That's opinion. We may describe, but we shouldn't advocate. Give the
doucment a good scrub to eliminate all advocacy, perhaps, and mention
the downsides for each upside?
One downside that doesn't seem to be mentioned is *finding* the transport
layer information. That will become increasingly difficult in future.
BTW I don't see ECMP or load balancing listed in section 3. Those
seem to be major applications of transport layer snooping.
Sent: Tuesday, April 11, 2017 7:48 AM
To: Stephen Farrell; Martin Thomson; ietf(_at_)ietf(_dot_)org
Subject: RE: draft-dolson-plus-middlebox-benefits (was RE: Review of
Please see inline.
De : Stephen Farrell
Envoyé : mardi 11 avril 2017 10:51
À : BOUCADAIR Mohamed IMT/OLN; Martin Thomson; ietf(_at_)ietf(_dot_)org
Cc : draft-dolson-plus-middlebox-benefits(_at_)tools(_dot_)ietf(_dot_)org
Objet : Re: draft-dolson-plus-middlebox-benefits (was RE: Review of draft-
On 11/04/17 09:15, mohamed(_dot_)boucadair(_at_)orange(_dot_)com wrote:
I hope that the IETF never publishes
draft-dolson-plus-middlebox-benefits; it makes claims about the
benefits of specific solutions for different use cases with the
goal of justifying those solutions.
[Med] I'm afraid this is speculating about the intent of
draft-dolson. Assured this is not the purpose of that document. The
motivation is to document current practices without including any
recommendation or claiming these solutions are superior to others.
Just to note that I completely agree with Martin's interpretation
of the thrust of this draft and I totally fail to see how your
argument above can be justified given that draft title, abstract
and even filename (and also the content;-).
[Med] "beneficial" is derived from the initial request that motivated this
draft (excerpt from the abstract):
At IETF97, at a meeting regarding the Path Layer UDP Substrate (PLUS)
protocol, a request was made for documentation about the benefits
that might be provided by permitting middleboxes to have some
visibility to transport-layer information.
When the abstract
says "This document summarizes benefits" then I cannot interpret
that as other than being intended to justify the uses described.
[Med] I would prefer if we can avoid to "interpret", but raise questions to
the authors if there is a doubt. The document does not provide a
recommendation or claims this is the only way to achieve the technical goals.
It does only reflect some deployment reality together with some motivations.
A fairly thorough re-write to aim to describe the pros and cons
would be a different and more useful document.
[Med] There are already many RFCs that discuss the issues/cons (I can cite
this RFC I co-authored https://tools.ietf.org/html/rfc6269 for the CGN case).
What is needed IMHO is something else: understand the requirements that led
to deploy some of these functions.
Similarly a draft
that strives to neutrally describe existing reality could maybe
be useful (*)
[Med] This is the intent of draft-dolson.
but one that only describes middlebox friends with
"benefits" is not IMO beneficial ;-)
[Med] The intent is not to "sell something" but to understand the technical
needs so that hopefully we can have a reference for future solution-oriented
If a given function can be provided without involving an on-path device, this
would be great for operators (optimize CAPEX/OPEX is our motto).
(*) That is the argument for draft-mm-effect-encrypt, for which I
do support publication (apparently in disagreement with Martin in