🚀 go-pugleaf

tja@cs.mu.OZ.AU (Tim Arnold) writes:

>mikec@mundil.cs.mu.OZ.AU (Michael CIAVARELLA) writes:

>>Hi all,
Hi Mike!

>>I have a (hypothetical) situation which I'd appreciate
>>some comment on.  I've tried to be as exact as possible,
>>but if you want any clarification just mail away.

>>Machines XX and YY belong to different departments - each
>>machine's maintenance, support and associated staff being  
>>paid for by the relevant department.  Whilst each machine
>>is run separately, they do have some users in common.  
>>There are no shared filesystems between XX and YY though
>>they are on the same network, and it is possible to 
>>login to each machine from the other.  

>>Let's say A is a non-administrative user who has an acount
>>on both XX and YY.  The administration on XX has specifically
>>forbade the use of program I on their machine, either directly
>>or while remote-logged in to program YY.  There is no such
>>restriction on machine YY.  So, user A logs into
>>host YY, compiles up program I, and leaves the binary in his
>>account on YY.  The next day, A logs on to YY to use his
>>newly-installed program, only to find that the binary has
>>been deleted during the night.

>>So, person A questions the administrator on YY as to why his 
>>binary was deleted.  YY's admin is surprised at the question
>>because he certainly didn't do it.  In fact, on checking
>>through the system logs, it looks like A deleted his own file.
>>A knows he didn't do it either, and that the chances of his
>>password being cracked are infinitesimal, since he only changed
>>it yesterday, and it was a reasonable password.

>>He does, however, have a .rhosts entry on YY allowing him to 
>>login from XX, and indeed, the system logs show that this 
>>took place last night. However again, A didn't perform this login.
>>In fact, he'd changed his password on machine XX yesterday, 
>>and on checking through the system logs on XX which just 
>>happen to be publicly readable, the only person using XX 
>>at the time of the remote login, was XX's system administrator.

>>Now, all questions of technical security aside, the evidence
>>seems to point to the admin on XX impersonating A and removing 
>>the file belonging to A.

>>If this is in fact the case, then:
>>	1)	What are the legal implications (if any)
>>		of the XX admin's actions?
>Probably a breach of some Cth legislation (Criminal) concerning communications.
>This is pretty precarious and given the prosecution's competence not likely to
>get to court.
>Possibly an offence under the Victorian forgery provisions but again I suspect
>judges will be quite lenient on sys.ops. who have authority to do an awful lot
>on the system. THere may be informal understandings about mutual management of
>resources. (These are pretty normal at Melbourne)
  Yeah, but common courtesy would be for XX to send mail to YY saying:
  "A has a binary on YY called I. The use of this program is forbidden on
  XX. He also has a .rhosts file between YY and XX. I am going to {} and
  would you take steps to []?

  XX
  " where {} and [] are XX's problem, not mine.

>>	2)	If A logs in to YY from XX and uses his
>>		program, does this give the XX admin the
>>		right to act as he did?
XX has the authority to do what they want on their system. The binary is
not on their system.
>If YY sysop deletes the binaries (or denies access by a group) there is a
NONONO! Read the problem. It's XX who denies the use of the program. YY has
no such restrictions. It was also XX who deleted the binary, not YY.
>clear prohibition to that group using that service on that machine.
>Given the Victorian provision on computer trespass (s.9 Computer Crimes Act)
>A is in a lot of hot water if A goes to court so I suggest that A should cut
>their losses. :-) 1/2
Garbage. IF it was YY who had deleted the file, A would have no recompense.
The question is: if XX (henceforth called the paranoid sysadmin, or PS)
deletes a file ON A SYSTEM NOT UNDER THE PS's CONTROL, to enforce restrictions
that the PS imposed on THE SYSTEM UNDER THIER CONTROL, is it moral/legal?

Well, I'm not qualified to give a legal opinion, but morally:
  * the sysadmin is entitled to take action within the limits of their power
  and their domain to enforce their policies.
  * YY's system is not within the domain of XX.
  * Remote logins from XX to YY are within the domain of XX.
  * The prohibition is on RUNNING program I.
  Therefore, PS had no right (morally speaking) to remove the binary. The
  only right the PS had was to remove the .rhosts file. If A had actually
  run I from a session on XX, he would have been in trouble; but the
  things that PS can do should be restricted to the things under their control.

>If XX sys.op. is employed to maintain the system and the use on the machine
>is provided for course purposes then the sysop has every right to ensure that
>the system operates the way he authorizes it. If there is some implied
>authority but no other way for him to remove the file then why not?
   Because the PS has no authority on YY's system. I understand the meaning
   of a .rhosts file to be that account M has the same access priveleges that
   account N has, but not that they are the same account.

>>	3)	What would you do in the XX admin's place?
>Alert the YY admin of the breach and suggest they delete the binary and suspend
What breach? Has A a) used program I on XX? b) used program I via a remote
login from XX?
>the students account indefinitely.
& get ready for serious flamage from the ENTIRE computing community.
>>	4)	What would you do in A's place?
>I wouldn't be in A's place. I'm not that stupid.
I'd get YY to do the same thing to my account on XX as XX did to my account
on YY. I'd arrange to be talking with XX at the time, and then complian to
them about it. If XX starts breathing fire [someone touched MY system]
then gently remnind them, in the presence of witnesses (preferably
including YY) what they did....

>>	5)	What would you do in the YY admin's place?
>Delete the binary and suspend the student's account pending an explanation.
Why? what rules of YY's system has A broken?
>After making the student sweaty a little I'd restore the account. If it
After that, I'd get my CV together & start looking for a new job.
>happened again I'd refer the student to the Head of Department after
>suspending the account recomending disciplinary action or prosecution.
And the student union refers you to the Head of Department after suspending
the account recommending discipliinary action or dismissal.
  I'd request XX to refrain from treading on my system, or better still,
  report a security violation on his machine.... preferably on an international
  newsgroup.
  First, I'd try and duplicate the incident.
>>If the admin didn't in fact "do it" then the question remains:
>>		"whodunnit?" ;-)
>Unix is not secure. There are a lot of backdoors.
True. But one that works when the sysadmin is the only one logged on? One
that "just happens" to be targetted on A? Strange things do happen.
Probalbly someone logged into machine YY? (Given that it is none of
A, XX, or YY who actually deleted the file).

># This is not legal advice. If you intend to act on these statements you should
>obtain appropriate legal advice from a qualified solicitor. I am not
>responsible for damages arising from reliance on these  statements.
>-----------------------------------------------------------------------------
>Tim Arnold        | Law/Science (Computer Science)   |  simul iustus
>tja@cs.mu.oz.AU   | The University of Melbourne      |  et peccator
>=============================================================================
* These opinions belong to pdt@mundil.cs.mu.OZ.AU unless otherwise specified.
"Such energies correspond to velocities at almost 95% of the speed of light,
far faster than the speed of sound." - Scientific American, Nov. 1991.