[Info-vax] VMS Privileges Versus Linux Capabilities
Stephen Hoffman
seaohveh at hoffmanlabs.invalid
Wed Jun 22 09:38:32 EDT 2016
On 2016-06-22 05:57:24 +0000, lawrencedo99 at gmail.com said:
> On Wednesday, June 22, 2016 at 3:20:09 AM UTC+12, Stephen Hoffman wrote:
>
>> As implemented, nobody in their right mind would want to use SEVMS, or
>> any other traditional mandatory access control system for that matter.
>> Some folks — certainly of their right mind — do have to use mandatory
>> access controls, because of their environment and the sorts of data
>> they have stored on their servers. Mandatory access control security
>> is not easy to manage, nor to use.
>
> The Linux ones have a logging mode, I understand, where they just
> generate reports about what would have been blocked, instead of
> actually blocking it. That may be of some help in figuring it out.
And if not, it's a simple matter of writing a few shell scripts and a
trip or two through awk, I'm sure.
On OpenVMS:
$ HELP SHOW AUDIT /ENABLE
Been around for decades.
But the reason for the denial is only a small part of why folks would
prefer to avoid using mandatory access controls, unnecessarily. The
difficulty involved in managing and using these systems goes far past
the auditing and figuring out what failed, and right to operations that
users of non-mandatory access control systems have become accustomed
to. Like sending around mail. On SEVMS and some other systems, mail
and other forms of information transfer only works on the same or
higher direction. Toward more and higher security users, or among
users of equal security. Using mandatory access control only gets more
interesting from there. You're managing information flows, which is
something that just isn't considered on discretionary access control
systems.
Sandboxing / jails avoids the information levels and information labels
and information sensitivity settings and information flow
considerations that arise with SEVMS and similar, and uses mechanisms
very much akin to mandatory access controls to block everything — file
access and transfers, various system calls, networking, etc — not
expressly authorized for the application. In OpenVMS, this'd have to
block system-wide logical names — or give each app its own
copy-on-write local copy of the system tables? — and would have to
figure out how to deal with WSAu: workstation devices, and a whole host
of other gnarly details. This is how you keep apps — innocent,
buggy/compromised or simply malicious — from getting tangled, when you
go to implement app stacking.
Almost nobody that has used and has managed mandatory access controls
is going to want to use it again, unnecessarily. Not without a very
good reason. Sandboxes and entitlements and app stacking are also a
pain in the arse, but more isolated and aren't trying to manage the
flow of sensitive information within the system. Sandboxing in a
mandatory access controls environment would be a whole new realm of
hurt, but at least that'd include the system developers and the local
system administrator and not (just) the sandbox app developers.
--
Pure Personal Opinion | HoffmanLabs LLC
More information about the Info-vax
mailing list