Re: [perpass] Tiny stacks

[Richard Barnes <[e-mail]>]

On Mon, Dec 9, 2013 at 4:58 PM, Brian E Carpenter <
[e-mail]> wrote:

> Hannes,
>
> On 10/12/2013 09:56, Hannes Tschofenig wrote:
> > Many of us are actually thinking about how to get the IP stack on these
> > devices.
> >
> > The IAB had a workshop in 2011 on smart objects and the report can be
> > found here: http://tools.ietf.org/html/rfc6574
> >
> > We then had a workshop specifically dedicated to security in 2012:
> >
> http://tools.ietf.org/html/draft-gilger-smart-object-security-workshop-02
> > (pending publication as an RFC).
>
> Fair enough, but did you consider specifically the privacy and
> surveillance aspects? I'm concerned that counter-measures that can
> be easily incorporated in full size devices may be too heavy for
> tiny devices. If this is not a real concern, I will be delighted
> of course.
>
> And there is the usual problem of converting workshop conclusions
> into WG action.
>

As I recall, the major upshot of the workshop, from a security point of
view, was that (1) security protocols are tough but tractable, and (2) the
really hard problem is the introduction problem.  By which I mean:
Smart/IoT devices are going to, by their nature, talk to something else
(otherwise they wouldn't need connectivity).  The "introduction problem" is
the challenge of telling devices whom they should talk to, and how to
authenticate them, in such a way that doesn't allow an attacker to insert
himself, with the very limited interfaces that IoT devices tend to have.
 At one layer, it's just an authentication / authorization problem, but
it's one that has much more impact on hardware/software configuration than
on protocol.

That is my understanding, at least, of how we arrived at the current state,
where most of the protocol work is focused on making the security protocols
nicer (e.g., CoAP, DICE).  Nobody has found an approach to the introduction
problem that applies everywhere^Wa lot of places.

In point of fact, most of the interesting IoT vulnerabilities we've seen so
far have not been due to either of the above problems, but rather to
manufacturers making stupid decisions that couldn't have been fixed by any
number of RFCs.

--Richard




>     Brian
>
> > There is even an IAB document in development that touches this topic:
> > http://tools.ietf.org/html/draft-iab-smart-object-architecture-03
> > (Comments welcome)
> >
> > [Recent comments indicated that there is a desire to talk more about
> > IPv6, and the transition mechanisms. Great that we worked on so many --
> > will for sure make it easier to fit them all on these devices.]
> >
> > As you know, we even have the IETF LWIG group that discusses these
> issues.
> >
> > If you look at recent events, like the Internet census
> > http://internetcensus2012.bitbucket.org/paper.html, then it should be
> > clear that even "small device" need security since otherwise we are
> > building the next generation botnet. This would not be good (tm).
> >
> > Ciao
> > Hannes
> >
> >
> > On 12/09/2013 07:47 PM, Brian E Carpenter wrote:
> >> On 09/12/2013 11:04, Stewart Bryant (stbryant) wrote:
> >> (on a different list and under a differeny Subject header)
> >> ...
> >>
> >>> Remembering of course that some platforms which wish
> >>> to use the Internet simply do not have the capability for
> >>> other than a very tiny very basic stack.
> >>>
> >>> I always use the PIC and the Arduino to remind myself what the
> >>> lower end of the franchise looks like.
> >> It seems to me that perpass should think a little bit about
> >> privacy and anti-surveillance issues for devices with tiny
> >> stacks, and see if that calls for any specific IETF work items.
> >>
> >>     Brian
> >> _______________________________________________
> >> perpass mailing list
> >> [e-mail]
> >> https://www.ietf.org/mailman/listinfo/perpass
> >
> >
> _______________________________________________
> perpass mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/perpass
>