Hi Thomas,
I definitely understand your trepidation against all in one installers but,
fundamentally, this is the new way of the world.
All in one installer => Statically compiled bunch of stuff that's hard to
maintain and track
Docker Container => Statically compiled bunch of stuff that's hard to
maintain and track
Go => Statically compiled bunch of stuff that's hard to maintain and track
Rust => Statically compiled bunch of stuff that's hard to maintain and track
You get the picture, modern life is a whiz-bang set of uncontrolled and
untrackable nonsense currently.
Also, the main reason that Ruby doesn't get a ton of love on Red Hat
systems is that Red Hat went full bore Python. The number of esoteric
Python libraries that show up is amazing compared with the similar Ruby
environment. That said, both from the stock distributions are pretty much
useless without enabling Software Collections to get a newer version.
For the purposes that I was talking about, just running it locally or
against test systems, the OpenSSL issue doesn't really come into play. But
I do 100% agree that distributed RPMs should be linked against the system
OpenSSL libraries for FIPS compatibility. We can work on patching the
upstream spec files to do this or make it a request *if* the SSG community
decides to adopt the tool set.
Thanks,
Trevor
On Wed, Mar 22, 2017 at 10:36 AM, Albrecht, Thomas C <
thomas.c.albrecht(a)lmco.com> wrote:
My two cents on the tooling aspect
Based on this conversation, I started doing my own investigation of Inspec
on our systems. One thing that had me nervous was Ruby, since I know
there’s a dependency hell that goes along with Ruby gems, and I wasn’t sure
how much work would need to be done to get the right packages installed
from the Red Hat repos. However, the installation from the inspec.io
website was a single RPM that has all the dependencies bundled into it.
This obviously has the benefit of being simple to install and maintain, and
not needing to worry about which version of Ruby is installed on your
system, but it doesn’t really mesh with the Unix philosophy of “Do one
thing and do it well.”
For example, in the /opt/inspec/embedded/bin directory, there are 51
binaries, most of which are duplicates of binaries that Red Hat ships in
their own packages. I can’t see an easy path to altering the Inspec
packaging to rely on Red Hat binaries, but I definitely can’t see Red Hat
even considering allowing an application install its own version of the
dependent tools, if only from a security perspective, since they’re
embedding sensitive libraries like OpenSSL.
So, from an outsiders perspective, I can’t see an easy road to get the
tools integrated into official RHEL repositories.
Tom Albrecht III, CISSP-ISSEP, GPEN, RHCSA
*From:* Shawn Wells [mailto:shawn@redhat.com]
*Sent:* Wednesday, March 22, 2017 10:58 AM
*To:* scap-security-guide(a)lists.fedorahosted.org
*Subject:* EXTERNAL: Re: RFC: chef.io Inspec
Really like the simplicity of Inspect! In my own world view, the only
thing that really matters is automating the C&A compliance burden and
doing so with tools+content that people can actually use (SCAP or not,
don't really care).
There are roadblocks to using Inspec: not natively included in enterprise
operating systems, the US Gov doesn't recognize it as a configuration or
vulnerability scanner, and results cannot be ingested by mandatory tooling
like ACAS and various CDM tools. Same challenges that Ansible and Puppet
face, too.
Out of the Ansible/Puppet/Chef grouping, Ansible is the only one with
roadmap to be natively included in Linux. However Ansible 'check mode'
isn't nearly as easy/robust as Inspec. Makes me feel a little.... stuck.
If content was developed for Inspec, how would people be able to consume
it? Theoretically the @redhat.com'ers could finagle a way to ship the
*content*. We certainly could upload it to the NIST National Checklist
Program. But how would people get the Inspec *tooling*? I'm not sure if
downloading off the internet is viable for most system owners.
On 3/22/17 10:05 AM, Trevor Vaughan wrote:
So, in my case, I'm working on the application and validation of different
compliance profiles as part of the SIMP project.
In each of our configuration modules, I wanted to integrate an SSG scan
into the final phases of our acceptance testing.
I probably should have said 'development framework' instead of
'demonstration framework'. But, the idea is that we can run regular
acceptance tests as part of our CI process by integrating inspec into that
mix.
We started down a path similar to inspec but abandoned it because we
wanted to use the authoritative SSG materials so that we could show the C&A
authorities an apples-to-apples comparison.
Unfortunately, this proved to be simply too complex since we could not
extract micro data streams out of the SSG in an automated fashion. For
example, we have an auditd module. For this, we would want to apply our
settings against NIST 800-53 and then test against the 800-53 SSG profile.
We would then want to switch to STIG and test against the STIG profile,
etc... However, we *only* want to do it for the auditd portions of the SSG
and we just couldn't find an easy way to make this happen that didn't
require a LOT of upkeep.
Thanks,
Trevor
On Tue, Mar 21, 2017 at 11:37 PM, Jason Newton <nevion(a)gmail.com> wrote:
Hi Trevor,
The hope is that any work with SSG, rigor included would be 10x less
cost of any lines of the SSG. So it is a new cost but relative to
long term maintenance costs, it likely more than wins out - and
hopefully the SSG project would arrive at that point fast - in like a
month or two if someone puts in serious porting effort and starts
getting peer review going.
For running Inspec on production nodes - I went back and forth on my
reply here and my final thoughts are, well this is the double edge
sword of solving all integration and parsing problems by hosting the
DSL in ruby. If all concerning functionally cannot be hidden (using
ruby-fu) prior to module loading without breaking functionality, then
there's not really much that can be done with this approach
fundamentally - perhaps they could make some validatable Inspec
DSL-based document (that is not ruby) for the InSpec DSL that is
runnable
Was not quite sure what you meant regarding validating demonstration
frameworks and utility although the condition of peer reviewed came
through clearly.
Any devs care to weigh in?
-Jason
On Mon, Mar 13, 2017 at 8:36 PM, Trevor Vaughan <tvaughan(a)onyxpoint.com>
wrote:
> I've also been looking at Inspec and it shows quite a bit of promise to
fix
> a lot of the usability issues of the SSG project.
>
> However, to be an integrated enhancement/replacement, it will need to go
> through the same level of rigor as the SSG checks.
>
> Let's face it, while that rigor can be painful, it is at least
validatable
> while Inspec is truly just whatever code you wish to run on a system.
>
> I'd be a bit scared of running Inspec on production nodes since it's
> basically a root-level command and control system with almost nothing in
the
> way of checks and balances.
>
> That said, for validating demonstration frameworks, it may be useful if,
and
> only if, the community content can become peer reviewed and concrete.
>
> Thoughts about integrating it into the SSG project?
>
> Trevor
>
> On Sun, Mar 12, 2017 at 8:17 PM, <nevion(a)gmail.com> wrote:
>>
>> Hi folks,
>>
>> I've watched the project for a couple of years now and done some work
with
>> it. I'm grateful many of you took the time to get SSG where it's at.
>> Recently I came across chef.io's Inspec and I thought hey, this seems
to
>> solve alot of the engineering challenges the SSG project seems to have,
>> especially with multi-distro support.
>> Github:
https://github.com/chef/inspec
>> Product:
https://www.chef.io/inspec/
>>
>> Here are some of the benefits my eyes see:
>>
>> * Content is centralized (less mutli-file editing for the same tests aka
>> less XMHell :-)
>> * Content is encouraged to be less distribution specific but allowed to
be
>> as specific as it needs, and this is encouraged inside a module, meaning
>> less versions are floating around in the codebase
>> * Declarative/Programmatic as opposed to Declaritive/document oriented -
>> I'd say with the later case has led to somewhat awkward and hard write
>> tests, in some cases
>> * Can test active configuration as well as configuration file state
>> * Looks significantly less verbose - as in saving 10x lines of
sourcecode
>> (not counting XML output) and developer productivity would be much
higher
>> * and as a possible benefit, I wouldn't be surprised if tests ran much
>> faster - though I have done no extensive tests to show this.
>>
>> Most importantly I believe inspec could pave the way to all distros
>> working off of a common and smaller codebase than the existing SSG
>> implementation and be far less maintenance burden and as such I wanted
to
>> raise awareness and get a topic going on the ML.
>>
>> Inspec doesn't seem compatible out right with openscap (or maybe it is -
>> they have bidirectional oval/xccdf consumers/emitters), but it IS
aiming to
>> replace OVAL/XCCDF as something more sane and just plain old better for
the
>> tasks. They also have tools for running it without scap tools involved
of
>> course so it doesn't have to lean on openscap et al, but for some people
>> being seperate from SCAP tools is going to be a con and not a pro.
>> _______________________________________________
>> scap-security-guide mailing list --
>> scap-security-guide(a)lists.fedorahosted.org
>> To unsubscribe send an email to
>> scap-security-guide-leave(a)lists.fedorahosted.org
>
>
>
>
> --
> Trevor Vaughan
> Vice President, Onyx Point, Inc
> (410) 541-6699 x788
>
> -- This account not approved for unencrypted proprietary information --
>
> _______________________________________________
> scap-security-guide mailing list --
> scap-security-guide(a)lists.fedorahosted.org
> To unsubscribe send an email to
> scap-security-guide-leave(a)lists.fedorahosted.org
>
_______________________________________________
scap-security-guide mailing list -- scap-security-guide@lists.
fedorahosted.org
To unsubscribe send an email to scap-security-guide-leave@
lists.fedorahosted.org
--
Trevor Vaughan
Vice President, Onyx Point, Inc
(410) 541-6699 x788 <(410)%20541-6699>
-- This account not approved for unencrypted proprietary information --
_______________________________________________
scap-security-guide mailing list -- scap-security-guide(a)lists.fedorahosted.org
To unsubscribe send an email to scap-security-guide-leave(a)lists.fedorahosted.org
--
Shawn Wells
Chief Security Strategist
U.S. Public Sector
shawn(a)redhat.com | 443.534.0130 <(443)%20534-0130>
_______________________________________________
scap-security-guide mailing list -- scap-security-guide@lists.
fedorahosted.org
To unsubscribe send an email to scap-security-guide-leave@
lists.fedorahosted.org
--
Trevor Vaughan
Vice President, Onyx Point, Inc
(410) 541-6699 x788
-- This account not approved for unencrypted proprietary information --