Re: Fwd: [Nea] Re: use of a design team to develop requirements

[Keith Moore <moore at cs.utk.edu>]

> I think I can safely say that the purpose of NAC is NOT to impose 
> anything on a host.

It's not the explicit purpose, but it is certainly a consequence of NEA 
if it is successful.  If you know something is going to do harm, and you 
do it anyway, you're still culpable for the harm even if doing harm 
wasn't your explicit purpose.

> The purpose of NAC is to allow a network owner to control access to 
> it's resources, which is well within the network owner's right. 
> Nothing more and nothing less.

Well, let's look at an analogy.  Let's say that you work for me, and I
want to establish whether you're trustworthy.  So I require you to
consent to have your house searched, at any time, for evidence of
anything that you might have or do that would cause me to question my
trust in you.  Would that be well within my rights as an employer?  And
if it's not okay for me to search your house, why is it okay for me to
search your laptop?

The real question is something like this - how invasive a search is
reasonable?


> If a host is not upto snuff and the network owner requires certain
> actions to take place, the host can *choose* not to participate in
> the network. There is no forcing anything. There is no other social,
> racial, national, or political agenda. And no, I don't agree with
> your claim that NAC is harmful to anyone

I strongly disagree.  Saying that a host can choose whether to use a 
particular network is like saying that a person can choose whether or 
not to breathe the local air, drink the local water, buy electricity 
from the local power company.  In many cases you have to use a 
particular network in order to function.  So the choice is to a large 
degree already taken away from the user.  When the user doesn't have 
choice other mechanisms are needed to protect the user's privacy.

> I can't speak for the IETF at large, but off the top of my noggin 
> here are some other examples of potentially draconian standards,

(RADIUS, PKIX, PPP, IKE)

None of these requires arbitrary information from a host - only a small 
amount of information that specifically exists for the purpose of 
authentication.  And if you can reduce NEA to a protocol that just gives 
a host a way to say to a network "XYZ virus scanner certifies that this 
host is good up to version X.Y of its virus tests" then I wouldn't have 
a problem with it.  The problems come from the potential that NEA can be 
used by a network to ask arbitrary questions of a host.  Standardizing 
spyware is not in the best interest of the Internet community.

Keith


_______________________________________________
Nea mailing list
Nea at ietf.org
https://www1.ietf.org/mailman/listinfo/nea