Re: [Spud] Putting Network-Layer Information in the Network Layer

[Ted Hardie <[e-mail]>]

On Fri, Jul 10, 2015 at 3:44 AM, Brian Trammell <[e-mail]> wrote:

>
> I've been thinking about whether it would be worth the effort to do an
> updated IP options transparency study.
>
> We revisited the safety of ECN in part because there were plausible
> processes for the situation to have changed -- the biggest problems that
> caused the Internet to be ECN-unsafe involved what were basically
> single-packet exploits against routers, and needed to be fixed for reasons
> *other* than making the feature work. (ECN *support* is another question,
> the answer to which is less cheerful.)
>
> I can't think of a comparable process making new/experimental options
> support less linked to connectivity risk in edge networks -- if nobody uses
> them (because they don't work) then if you try to use them and they don't
> work, it's your fault. That would unfortunately seem to be a rather stable
> state.
>
>

​From a research design point of view, I think you could construct such a
study so that your test nodes tried both IP options​ and SPUD, so I think
it is worth thinking about quite seriously.  While it is not the main point
of them, you might be able to talk the RIPE and MLAB folks into adding code
to their test devices to enable these checks.

More data is always good,

Ted