[aprssig] Universal APRS messaging

[Heikki Hannikainen]

On Thu, 23 Oct 2008, Tyler Allison wrote:
> Steve Dimse wrote:
>> We'll have to agree to disagree on this. I think providing any
>> illusion that there is security is wrong. IGate operators need to know
>> they bear total responsibility for everything transmitted through
>> their IGate. Authenticating web access give the illusion of security
>> which does not exist.
>
> No we don't. I agree with you on a philosophical stand point :)
>
> What I meant is that if someone is bound and determined to implement
> something the only real thing I've heard that might stop some random acts
> of annoyance is the web access authentication. But it should not be
> cloaked in the guise of removing liability. It's stopping ankle biters,
> which are not the real liability risk.

I have to agree with Tyler on the practical side of things. I know of a 
few ways to break into my apartment without much trouble, but I still want 
to keep the front door locked so that anyone can't simply walk in. And 
yes, I know that the front door can be picked too.

If there is a security hole on one side of a production system, we 
shouldn't create more of them on the other side to just prove the point. I 
think we should work to plug the existing holes instead of creating more 
of them.

On the philosophical side, I do enjoy reading the bugtraq mailing list and 
see the value of releasing proof-of-concept code to force big companies to 
fix their security holes, if they're not willing to do so otherwise. As 
long as they're not publishing holes in *my* systems, it's fine. :)

And while there isn't security on APRS-IS, I wouldn't actually mind 
improving security on the web interfaces. It'll be useful when/if the 
security on APRS-ISv2 (or whatever) is implemented. OpenID would be nice 
for mutual sharing of verified license status information, I suggested it 
in the spring:

http://oh7lzb.blogspot.com/2008/05/how-to-authenticate-licensed-hams.html

It simply does not work for the APRS-IS case, though. RSA/DSA signatures 
by igates for every gated message, anyone? :)

(Seriously speaking, the CA mess would be a maintenance nightmare in such 
a decentralized system, and the CPU requirements would be noticeable. I 
doubt people would bother.)

   - Hessu, OH7LZB (aprs.fi)