Conference noted::hackers_v1

	It seems to me that the last thing could be done simply by
doing this:

		    $ MAIL NL: node::SYS$SYSDEVICE

Or do some systems (clustered?) not use SYS$SYSDEVICE?
		<_Jym_>

SYS$SYSDEVICE is always there.
							-- Jerry

	After playing with this a bit, I've come up with a few
things:
	o Before auto-forwarding came along, it was possible
	  to use system-wide logicals to forward mail (and it
	  still is).  If I were to do this:
		    $ define/system DYER KEN::OLSEN
	  OLSEN on node KEN would get my mail.  I don't know
	  if this was done on purpose (bug or feature?), but
	  I don't think it's needed anymore.

	o MAIL to TT: instead of NL: - If there is, for some
	  reason, a username identical to the logical you're
	  peeking at, sending mail to NL: will send an empty
	  and perhaps incriminating message to that username.
	  Sending it to TT: gives you a chance to ^C out of it.

	o Another good logical is LUSERS$REMOTE_ALLOW.

	o Sometimes sending to NODE::SYS$WELCOME will reveal
	  interesting things.  For some strange reason, if you
	  send it to your own node, and SYS$WELCOME points to
	  a file, it will give you a NOSUCHUSR message for each
	  record in the file; but if you send it to another node
	  using the network (including your own node if you
	  specify it as 0::), you'll only get the first record.

	Anyone else playing with this?
		<_Jym_>

                       -< determining remote time zones >-

RE: .0:

>The following dcl trickery will reveal what time zone a remote node is in:
>
>	$ MAIL NL: node::MAIL$TIMEZONE
>	or
>	$ MAIL NL: node::NOTES$TIMEZONE
>
>	$ MAIL NL: node::SYS$SYSTEM

Doesn't always work if the logical name isnt defined on the node (obvious),
and instead of SYS$SYSTEM one should use SYS$SYSROOT.

My interest derives from trying to figure out where all these @##$%%^*! nodes
actually ARE! Does a list exist showing where they are, who is connected to
whom, etc? Just curious...

- Giles

    There's a couple of files in ANCHOR::NET$NODES that give node names,
    location codes, system manager, type of system, ....  They are
    MININODE.LST and SITENODE.LST.

> My interest derives from trying to figure out where all these @##$%%^*! nodes
> actually ARE! Does a list exist showing where they are, who is connected to
> whom, etc? Just curious...

        Who  is  Connected  to whom? With over 10,000 nodes! I hope you 
        have better things to do with your time.  Try the suggestion of 
        .5 ( Better Still, I keep an up-to-date copy of those on my Vax 
        here  in   Sydney   -   Save   the   Net   traffic   ).    Also 
        Anchor""::net$maps:*  has  Maps  of the areas and area routers, 
        but I think that they may be dated as they were being  done  by 
        Stan Rabinowitz who has since left Digital.
        
        q

	10,000 nodes?  Last I heard it was ~7500.

	To determine what or where a remote node might be, I generally do
a SET HOST.  If there is an announce, it often tells.  Sometimes NCP
TELL [node] SHOW EXEC will have something useful in the ID field (or else
it will tell you the CPU type).  These will create LOGFAIL accounting
entries, if that bothers you, or if you plan to be malicious.

	If you have a username (from a Notes entry or a MAIL message) then
sometimes ELF can locate that person.

	_Scott

RE: .-1

Something I got recently:

Message-class: DECMAIL-MS
From:	NAME: MCCAULEY
	INITLS: BOB
	FUNC: DT/EASYNET MGT.
	ADDR: VRO5-2/C10
	TEL: 273-3063 <70002@DECMAIL@DONJON@VRO>
Posted-date: 20-May-1986

(long distribution list removed)

Subject: EASYnet Milestone

A few years ago (1978) we had to split up our internal DECsystem-10 
network because the ANF-10 networking products didn't work reliably on 
a network of 50 nodes.  That wasn't considered surprising because not 
many people had envisioned networks that "large" when the products 
were designed.

Last Wednesday, the EASYnet, Digital's corporate computer network,
registered its 10,000th system and nobody even noticed.  Measured in
terms of number of nodes, that makes EASYnet the WORLD'S LARGEST
PRIVATE COMPUTER NETWORK.  We are currently adding over 100 new
systems per week, so we don't expect to be overtaken for a while. 

The 10,000th node is a MicroVAX in MKO2, a system named PLANIT.  A 
small memento and ceremony is planned to recognize this milestone in 
the growth of EASYnet.

The fact that the network functions so well in this complex and
dynamic environment is a tribute to the quality of the Digital product
set (hardware and software, particularly DECnet) and the support teams
in Digital Telecommunications (both headquarters and field) and Field
Service.  We have every confidence that the network will continue to
operate smoothly as we move forward to our next target of 20,000
nodes.  At the present rate of growth, we expect to be there around
the end of FY87. 

The significance of this event is two fold.  First, as the largest
user of Digital's products our experience operating a large corporate
network can provide insight into future customer requirements and
provides a "stress test" for the products.  Secondly, we serve as a
reference account and an example of how Digital's networking products
and services can be used to solve real business problems - in all
parts of the corporation, every day - another example of how "Digital
has it now". 

    Somewhere around (the Tools clearinghouse?) there is a tool called
    TRACER which figures out what the path is between two nodes.  One
    can figure out topology that way.  One can also do things like
    TELL xxx SHOW KNOWN CIRC, although that gets rather wordy on ethers.
    
    
    Burns
    

.0>These strike me as mild security holes.
    
    I agree.  I think MAIL should return the NOSUCHUSR error WITHOUT
    remote translation.  Translation should only be performed locally.
    Comments?
    
.3>	o Before auto-forwarding came along, it was possible
  > 	  to use system-wide logicals to forward mail (and it
  >	  still is).  If I were to do this:
  >		    $ define/system DYER KEN::OLSEN
  >	  OLSEN on node KEN would get my mail.  I don't know
  >	  if this was done on purpose (bug or feature?), but
  >	  I don't think it's needed anymore.
    
    Actually, it is still useful in cases such as
    
    	MAIL X.TXT MY_UNIT/SUBJECT="Hooey"
    
    where 
    
    	MYUNIT = "COAR,JONES,WILSON,PETERSON,FOO,BAR"
    
    or
    
    	MY_UNIT = "@MAIL$:STLC_DELIVERY.DISTRIBUTION"
    
.3>	o Another good logical is LUSERS$REMOTE_ALLOW.
    
    What is this all about?
    
.9> Somewhere around (the Tools clearinghouse?) there is a tool called
  > TRACER which figures out what the path is between two nodes.
    
    There's one called NETPATH as well; are these actually the same
    thing under different names?
    
    #ken	:-)}

    re .10:
    
    The bug mentioned in .0 is supposed to be fixed in V5.
    
    	Jon

>re : Note 164.0   determining remote time zones 2-OCT-1985 16:27
>
>The following dcl trickery will reveal what time zone a remote node is in:
>
>	$ MAIL NL: node::MAIL$TIMEZONE
>	or
>	$ MAIL NL: node::NOTES$TIMEZONE
>

    I realize the original note is somewhat dated, but......
    
    How does a 'hacker' determine system time on a remote node??

    I am considering an application that must provide international
    clocks.  With all the 'mumblemumble' standard times out there there
    is no fixed relationship to local time in all time zones.
    
    (BTW NOTES/MAIL$TIMEZONE don't seem to be defined anymore)
    
    All help/ideas appreciated -- 
    

    Well, here's one answer.  It uses the creation date and time of
    a file created in the default DECNET directory.
     
    Note: This will not work if the default DECNET user's directory
    does not exist, and there are some nodes out there like this.
    

    .
    .
    .
$         if p1 .eqs. "" then inquire p1 "Node "
$         node = p1
$
$         open/write tempfile 'node'::timestamp.log
$         close      tempfile 
$         rem_time = f$file_attributes("''node'::timestamp.log","cdt")
$         delete/nolog 'node'::timestamp.log;*
$
$         say     "Time on node ''node' is ''rem_time'"
$
$

Another possibility is to use the LISP logicals which may be defined if
VAX LISP is installed on the remote node.  eg. on QED:

(LNM$SYSTEM_TABLE)

  "LISP$DAYLIGHT_SAVING_TIME_P" = "YES"
  "LISP$TIME_ZONE" = "5"

    
    Thanks, I appreciate your quick, accurate, and hack of a response.