Re: System Functional Requirements

Karsten Hilbert <[email protected]> Tue, 22 Mar 2016 10:58:46 +0100
Newsgroups gmane.comp.gnu.medical.devel
Message-ID <[email protected]>
Hello Holger,

your response is more directly to the point. I agree that we
agree that privacy concerns cannot be taken lightly in
healthcare.

> I think what Alejandro was looking for was what normally is supposed to some sort of security classification.

I did not understand what he was looking for. That's why I
asked whether he would like to elaborate.

> Besides: I would strongly disagree with your statement, that timeouts are the job of higher layers. 

I expressed my opinion on those two examples, not on timeouts in general.

> Within my regular job, this kind of data would be ranked
> into highes security customer data, which would imply, that
> any system  storing or using this data would need to use
> encryption for data storage,

I agree. That's the job of PostgreSQL or the filesystem (IMO
the latter).

> at least one way encryption for passwords,

Passwords are encrypted by PostgreSQL.

> two factor authentication for admins

PostgreSQL can be set up for two factor.

> and definitely  timeouts for userlogins within the
> application to make sure, that no data can be viewed by any
> person without permission.

What is the exact difference between a screen locker that
kicks in after one minute and an application lock that kicks
in after one minute ?

I am pretty sure you aren't suggesting that I individually
unlock all 4 applications I have running alongside ?  A
desktop session lock would lock/unlock all 4 at the same
time, some of which I may not even have control over due to
them being proprietary.

> But I know that most of the times the developers simply
> fail in offering these people the tools they need to have
> to implement these security levels in a usable way.

I am sure Alejandro could elaborate on - if he so wishes -
which threats he wishes to mitigate. Define the threat
scenario, identify the attack surface, look for the tools --
fine with me.

> What I saw lately as a good example was some sort usb-token
> for every waiter. Without a key, the terminals were locked,
> inserting and some sort of 4 character pin unlocked the
> terminal „app“.

This can be done with a pam_usb enabled desktop session
locker (say, screensaver).

> I know of some people using ubikey for this kind of 2
> factor auth, some are using google authenticator, which i
> would not find usable here, but sure there are more of these
> possibilities. 
> This is, what I would like to see at a doctors place... 

Sure, and they should be implemented at the desktop session level.

This is _not_ to say the application level should _not_ offer
any security measures.

And, BTW, this is FLOSS so patches are welcome and forks
possible.

$> grep -H -n -A 10 'def _on_user_activity_timer' Projekte/gm-git/gnumed/gnumed/client/wxpython/gmGuiMain.py

Projekte/gm-git/gnumed/gnumed/client/wxpython/gmGuiMain.py:3517:	def _on_user_activity_timer_expired(self, cookie=None):
Projekte/gm-git/gnumed/gnumed/client/wxpython/gmGuiMain.py-3518-
Projekte/gm-git/gnumed/gnumed/client/wxpython/gmGuiMain.py-3519-		if self.user_activity_detected:
Projekte/gm-git/gnumed/gnumed/client/wxpython/gmGuiMain.py-3520-			self.elapsed_inactivity_slices = 0
Projekte/gm-git/gnumed/gnumed/client/wxpython/gmGuiMain.py-3521-			self.user_activity_detected = False
Projekte/gm-git/gnumed/gnumed/client/wxpython/gmGuiMain.py-3522-			self.elapsed_inactivity_slices += 1
Projekte/gm-git/gnumed/gnumed/client/wxpython/gmGuiMain.py-3523-		else:
Projekte/gm-git/gnumed/gnumed/client/wxpython/gmGuiMain.py-3524-			if self.elapsed_inactivity_slices >= self.max_user_inactivity_slices:
Projekte/gm-git/gnumed/gnumed/client/wxpython/gmGuiMain.py-3525-#				print("User was inactive for 30 seconds.")
Projekte/gm-git/gnumed/gnumed/client/wxpython/gmGuiMain.py-3526-				pass
Projekte/gm-git/gnumed/gnumed/client/wxpython/gmGuiMain.py-3527-

Karsten
-- 
GPG key ID E4071346 @ eu.pool.sks-keyservers.net
E167 67FD A291 2BEA 73BD  4537 78B9 A9F9 E407 1346

_______________________________________________
Gnumed-devel mailing list
[email protected]
https://lists.gnu.org/mailman/listinfo/gnumed-devel