Hi Stephen,
I oppose the change for reasons I state below:
On 1/13/14 8:28 PM, Stephen Farrell wrote:
Hi all,
Sam raised an issue that we didn't handle in the -03
version. [1] After some offlist mail between us I think
that his issue would be resolved with a change such as
the one below, adding a new paragraph.
Should text like this be added to the document? FWIW,
I'd be ok with that, or with leaving this out and for
a slightly longer discussion of the topic to be part
of Richard et al's problem statement draft. [2]
Please say what you think.
The proposed change in OLD/NEW form is:
OLD
Those developing IETF specifications need to be able to describe how
they have considered pervasive monitoring, and, if the attack is
relevant to the work to be published, be able to justify related
design decisions. This does not mean a new "pervasive monitoring
considerations" section is needed in IETF documentation. It means
that, if asked, there needs to be a good answer to the question "is
pervasive monitoring relevant to this work and if so how has it been
considered?"
NEW
Those developing IETF specifications need to be able to describe how
they have considered pervasive monitoring, and, if the attack is
relevant to the work to be published, be able to justify related
design decisions. This does not mean a new "pervasive monitoring
considerations" section is needed in IETF documentation. It means
that, if asked, there needs to be a good answer to the question "is
pervasive monitoring relevant to this work and if so how has it been
considered?"
In particular, architectural decisions, including which existing
technology is re-used, significantly impact the vulnerability of
a protocol to pervasive monitoring. For example, if a protocol
uses DNS to store information, then a passive attacker can observe
the queries made to the DNS. Those developing IETF specifications
therefore need to consider mitigating pervasive monitoring when
making these architectural decisions and be prepared to justify
their decisions. Getting adequate, early review of architectural
decisions including whether appropriate mitigation of pervasive
monitoring can be made is important. Revisiting these architectural
decisions late in the process is very costly.
While the sentiment isn't bad it's a bit of a truism, and the example
isn't good, since in large part the information shared through DNS can
often and more accurately (albeit not always) be gleaned through other
means (flow state, for instance). Once again I return to what I have
said previously: developers of specifications should examine this on a
case by case basis (at least for now), weigh whatever tradeoffs they
uncover, and move forward from there.
Eliot