> > To have an official standard.
>
> I guess I don't see why an official standard is needed for a one-liner.

For partly automation.

> Plus, I would think that each administrator would have their own
> set of standards that they should enforce; an "official standard" would
> lead to automatic reactions instead of thinking for three seconds.

Well, depending on what type of architecture you have to manage, those 3 sec 
can accumulate rather quickly. I don't say "take human brains out of 
decision", but it should have a form that is managable by software frontends.

>
> I mean, how is ssh going to authenticate you to the remote administrator?

Rather meant that ssh(d) should look in a defined place for the public key and 
then inform the admin that a new one's waiting for auth. I wouldn't wanna 
have it 100% automated either. 

> (This is where PKI comes in, I suppose... you and the administrator should
> share a common certificate authority, so he can look up/verify your cert.
> Do we want to start putting ssh public keys into a PKI framework?)

Sounds tempting doesn't it...

> > > computer, and your key is compromised, just like a fixed password would
> > > be.
> >
> > If all that can be read is my public key?
>
> So you'd walk up to the administrator and say "I'd like an account on
> machine $FOO", hand 'em your company badge and a USB stick?  (And, one
> hopes, the fingerprint to you key on a sheet of paper.)

Pardon? I mean, my public key can travel as much as it wants, can it not? 
After all, it's a *public* key.

>
> It won't help you log in from an arbitrary machine, which is what I
> thought you were trying to do.  If you're just trying to distribute
> your public key so you can access various machines from your primary
> box, that's different.

That's pretty much what I want. Example: got three machines here. (Yet :) )
So I'd like all three of them to hold my public key and when I connect my usb 
stick to *one* machine I'd like to able to ssh into the others, too, without 
providing a password.

> Smart Cards aren't around because they aren't widely used; but they're
> being used in more and more systems.

I'll have a look at that when smart card readers are as common as usb ports. 

>
> Technology that doesn't _actually_ solve your problem isn't worth much.

Technology that nobody has either.

> > Laptop on my keyring will have me lose my pants a lot in public. not
> > good. ;)
>
> Isn't that all the rage among kids these days?

Been to a disco lately. Wrong evening, thought 80s party was on, instead was 
new metal / hardcore. Man I felt old.

> > If that key - and the locks! - are sufficiently secure - alright.
>
> Ah, well, this is where our comfort levels differ.
>
> In a corporate environment especially, one-key-fits-all-locks is a lousy
> design, despite being highly desired.

Well, one key fits all works in my case since I granted myself access to all 
my stuff :) - different thing in a company of course, but at the CeBit 
Toshiba hat those neat 4GB SD cards - can hold a lot of keys, I'd say...


-- 
-----BEGIN GEEK CODE BLOCK-----
Version: 3.12
GCS d--(+)@ s-:+ a- C+++(++++) UL+>++++ P+>++ L+++>++++ E-- W++ N o? K-
w--(---) !O M+ V- PS++(+) PE(-) Y++ PGP t++(---)@ 5 X+(++) R+(++) tv--(+)@ 
b++(+++) DI+++ D G++ e* h>++ r%>* y?
------END GEEK CODE BLOCK------

http://www.stop1984.com
http://www.againsttcpa.com


-- 
[email protected]
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-list

Reply via email to