] From:	HENRY::IN%"leichter-jerry%YALE.ARPA%SRI-KL.ARPA%relay.cs.net@rca.com" 15-NOV-1986 21:51  To:	MHJohnson@HI-MULTICS.ARPA # Subj:	Re: DECNET Security Questions   G       I have a series of questions about DECNET security.  Our customer E     would like to have our system be a part of their corporate DECNET J     network and questions have been raised about how the systems should beJ     secured.  We are concerned that proprietary information might be moved-     from system to system w/o being detected. K There are many different notions of "security".  You've described one:  You L want to DETECT violations.  Detection does not necessarily imply PREVENTION. Is that what you have in mind?  J There are also related questions.  For example, who do you want to protectL against?  Unprivileged users at terminals on the remote machine?  PRIVILEGEDJ users?  Users who can physically play around with a remote machine?  UsersJ who can tap the phone lines connecting the machines?  All of these requireG different amounts of effort to deal with; I've ordered them in order of L increasing difficulty.  (Protecting against wire tappers is quite difficult,$ given the present state of the art.)     L Similarly, the amount and kind of access you want to provide to remote users makes a big difference.   L       (1) Should I disable the default DECNET account completely or is thereK     a way to properly restrict access to files and other resources instead? L You can restrict access in various ways, at a lot of "grains of coarseness".M The UIC protection scheme is enough, if the UIC of the DECnet account puts it N in a unique group of its own, to ensure that only files that are accessible toN WORLD could be accessed via the default DECnet account.  With ACL's, you couldJ go further and explicitly exclude access by the default DECnet account, orN any network access (even using an explicit username/password or proxy, but notN a SET HOST remote process) by excluding all holders of the NETWORK identifier.K (The SET HOST processes have REMOTE rather than NETWORK.)  Note that all of N this can be done even while leaving the files accessible to the "world" on theH local machine.  (However, watch out for SUBMIT/REMOTE; I don't know what" characterizes such jobs off hand.)     J       (2) Can I reject SET HOST commands from some nodes?  I would like to>     implicitly reject all nodes & then allow access from some.J I don't think you can do this with NCP directly.  However, if you define aH system-wide SYS$SYLOGIN file, it will ALWAYS execute.  Within that file,M you can check the logicals SYS$REM_NODE and SYS$REM_ID to determine where the J connection is coming from, and from whom, and then act accordingly.  (Note) that these logicals are new in VMS V4.4.)      K Note that you can restrict access with the ACCESS parameter of the SET EXEC L and SET NODE NCP commands.  However, be aware that a remote user with SYSPRV/ is allowed in despite such access restrictions.   L       (3) Should I add a proxy entry of *::* -> XXX, where XXX is a usernameJ     that is disabled from network operations?  Does this gain me anything?H I can't think of anything at all that this gains; it's more or less likeK disabling the default DECnet account, except that if the default account is L still there, a remote user can always get to it (by using the NODE"":: form,M which specifies explicitly null access control, thus overriding any proxies).      H       (4) Should I dedicate a MicroVAX for this link & control access at.     that point or does it make any difference?L It all depends on what it is you are trying to accomplish.  You haven't been< clear enough on this for me to be able to answer.  It might.     K       (5) Must all systems on the network be protected or can a `fire wall' :     be set up (like on the system described in 4) instead?L Again, it depends on what threats you are concerned about, and on what kinds of access you want to allow.  M Given the state of the art, I would be very hesitant about relying completely M on a "fire wall", unless that fire wall was simple and "thick", allowing very E little through at all (i.e., the ALLOWED accesses were very limited).      D     The reason we wish to do this is for electronic mail, updates of6     documentation and source, and other data transfer.H Still not enough information to be sure.  Do you need on-line access, orI just the ability to transfer files?  Could you live with nothing but MAIL M access?  (That could be made highly secure by having an intermediate node, as M in (4), with "mail drops" - there is no direct DECnet connection "across" the K MicroVAX; it talks to one network at a time, picking up mail and forwarding : it.  This would obviously require some hacking to set up.)  " 						        We will have some ofL     our people at their site for several months and they will need access toI     some of our stuff on a regular basis.  If you have any suggestions on H     this, please let me know.  I doubt that this is the only customer weJ     wish to do this for, so if we don't find out now, we will have to find=     out later.  I'll summarize the responses in a week or so.    							-- Jerry  ------- 