WEBVTT

00:00:01.238 --> 00:00:04.182
So requirement 6.1 and 6.2.

00:00:04.182 --> 00:00:09.460
Basically this is where in my experience, people lose cardholder data.

00:00:09.460 --> 00:00:13.120
Vulnerability management, patch management,

00:00:13.120 --> 00:00:15.135
so this is probably the most important part,

00:00:15.135 --> 00:00:17.522
and we've got actually a whole module,

00:00:17.522 --> 00:00:21.501
a whole separate module to talk about this.

00:00:21.501 --> 00:00:21.824
So yeah,

00:00:21.824 --> 00:00:23.279
not doing vulnerability management for me is the

00:00:23.279 --> 00:00:25.715
number one cause of data breaches.

00:00:25.715 --> 00:00:27.573
Every time I've seen forensics reports, it's like yeah,

00:00:27.573 --> 00:00:31.300
it was a known vulnerability of the operating system or the application.

00:00:31.300 --> 00:00:32.926
Same experience for you?

00:00:32.926 --> 00:00:34.348
Very much so, very much so,

00:00:34.348 --> 00:00:36.361
and it's very much where I see a lot of organizations

00:00:36.361 --> 00:00:38.546
struggling with their compliance requirements.

00:00:38.546 --> 00:00:42.680
Okay, as I talked about in the training module before this,

00:00:42.680 --> 00:00:48.183
this is not about running a vulnerability scanner against the system.

00:00:48.183 --> 00:00:50.126
That’s a whole different requirement.

00:00:50.126 --> 00:00:53.176
So you can't do this by scanning things remotely saying,

00:00:53.176 --> 00:00:55.321
oh look, that's got a vulnerability.

00:00:55.321 --> 00:00:57.141
This is about managed,

00:00:57.141 --> 00:01:00.040
proactively managing the software that you're running on your systems,

00:01:00.040 --> 00:01:00.644
isn't it?

00:01:00.644 --> 00:01:01.720
Yes.

00:01:01.720 --> 00:01:02.712
How do you do that then?

00:01:02.712 --> 00:01:04.479
Well, so I think you've got a couple of things, right?

00:01:04.479 --> 00:01:05.230
One,

00:01:05.230 --> 00:01:10.172
you want to know about vulnerability information for

00:01:10.172 --> 00:01:11.950
the software that you're using, right?

00:01:11.950 --> 00:01:16.166
So you need to subscribe to the relevant alerts from the

00:01:16.166 --> 00:01:18.220
software vendors or the software providers,

00:01:18.220 --> 00:01:20.970
or you know, an aggregator service like the U.S.

00:01:20.970 --> 00:01:22.883
Cert or the like.

00:01:22.883 --> 00:01:24.533
How are you assessing that?

00:01:24.533 --> 00:01:27.908
As an assessor, when you go into assess 6.1 and 6.2,

00:01:27.908 --> 00:01:30.135
what are you looking for from a organization?

00:01:30.135 --> 00:01:34.232
Show me bulletins that you're getting, right?

00:01:34.232 --> 00:01:35.441
Show me who you're getting it from,

00:01:35.441 --> 00:01:37.651
does that match up with your technology stack,

00:01:37.651 --> 00:01:40.945
and then show me what you do about it, right?

00:01:40.945 --> 00:01:46.120
So take those vulnerability details that you get and then tell me about

00:01:46.120 --> 00:01:49.978
how you organize that and then deploy those fix-its.

00:01:49.978 --> 00:01:56.194
So you're looking for prioritization based on CVSS base scores?

00:01:56.194 --> 00:01:56.887
You could.

00:01:56.887 --> 00:02:01.085
So the requirement that talks about assigning a risk ranking vulnerability,

00:02:01.085 --> 00:02:05.079
but the goal is that it needs to be objective and repeatable,

00:02:05.079 --> 00:02:05.478
right?

00:02:05.478 --> 00:02:08.310
So it's not just an opportunity for everybody to just sit

00:02:08.310 --> 00:02:09.662
around and backslap each other and say, yeah,

00:02:09.662 --> 00:02:11.353
we don't think that's that big of a deal,

00:02:11.353 --> 00:02:11.568
right?

00:02:11.568 --> 00:02:13.492
There's got to be some useful methodology.

00:02:13.492 --> 00:02:13.777
Right?

00:02:13.777 --> 00:02:16.081
A lot of organizations do use CVSS, right,

00:02:16.081 --> 00:02:18.978
and maybe apply the temporal and the environmental scores,

00:02:18.978 --> 00:02:21.947
which are not sort of typically applied.

00:02:21.947 --> 00:02:22.711
You see,

00:02:22.711 --> 00:02:26.629
I have seen organizations just take the base scores and they just completely

00:02:26.629 --> 00:02:29.998
ignore their own environment and they just run it off base,

00:02:29.998 --> 00:02:31.709
and that's insane.

00:02:31.709 --> 00:02:33.288
It's kind of nuts, right?

00:02:33.288 --> 00:02:37.857
And very often the temporal and the environmental scores help bring

00:02:37.857 --> 00:02:41.901
some of those risks down from them and would conceivably make their

00:02:41.901 --> 00:02:44.829
lives easier than fewer sort of urgent,

00:02:44.829 --> 00:02:50.765
burning crises to deal with in terms of their patch management.

00:02:50.765 --> 00:02:53.594
So from an architectural perspective, I mean,

00:02:53.594 --> 00:02:57.962
should people be using tools for this?

00:02:57.962 --> 00:03:02.016
Yes, to the greatest extent possible, and where they cannot,

00:03:02.016 --> 00:03:07.904
they want a really, really detailed a list of the kind of one-off cases,

00:03:07.904 --> 00:03:12.229
right, those funky security appliances or the legacy systems,

00:03:12.229 --> 00:03:16.809
or the things that they cannot automate or tool effectively,

00:03:16.809 --> 00:03:17.807
right?

00:03:17.807 --> 00:03:18.951
Know where those are,

00:03:18.951 --> 00:03:22.384
make sure you've got a really sort of firm grip on what

00:03:22.384 --> 00:03:24.769
that is and what you need to do about it so that those

00:03:24.769 --> 00:03:27.411
don't become the kind of forgotten, unpatched corners.

00:03:27.411 --> 00:03:31.490
Right, and as an assessor, do you like it when you walk in and they say,

00:03:31.490 --> 00:03:34.515
well, here's my vulnerability management tool?

00:03:34.515 --> 00:03:37.312
Does that give you a warm feeling?

00:03:37.312 --> 00:03:40.964
I like to see if it they can then say not just here is our tool,

00:03:40.964 --> 00:03:44.025
but here's how we use it, and here's what we do,

00:03:44.025 --> 00:03:44.358
right?

00:03:44.358 --> 00:03:47.295
And then here is, you know, here is the outcome from that.

00:03:47.295 --> 00:03:49.696
You know, if there's just an oh yeah,

00:03:49.696 --> 00:03:52.539
here's the vulnerability management tool that we gave you know,

00:03:52.539 --> 00:03:55.029
the dumbest person in our department to operate so that we

00:03:55.029 --> 00:03:58.112
don't have to deal with either him or it,

00:03:58.112 --> 00:04:02.945
right, I mean that's not great, but if it's an integral part of their operations.

00:04:02.945 --> 00:04:07.346
I think the other sort of really important thing that I encountered time

00:04:07.346 --> 00:04:11.702
and time again is that from a sort of strategy basis,

00:04:11.702 --> 00:04:15.069
a lot of organizations aren't dealing with their

00:04:15.069 --> 00:04:19.185
software end of life in a useful fashion,

00:04:19.185 --> 00:04:25.052
and honestly, if you're not thinking about end of life for software components,

00:04:25.052 --> 00:04:26.628
particularly the really underlying stuff,

00:04:26.628 --> 00:04:29.600
the programming languages, the operating systems,

00:04:29.600 --> 00:04:31.890
you know, the major application software stuff.

00:04:31.890 --> 00:04:35.039
If you're not dealing with that at like a risk assessment level,

00:04:35.039 --> 00:04:37.527
then there's probably something broken in your risk assessment.

00:04:37.527 --> 00:04:40.798
Should that be in your PCI DSS risk assessment in requirement 12?

00:04:40.798 --> 00:04:41.238
Well,

00:04:41.238 --> 00:04:44.797
the requirement doesn’t stipulate as much so I cannot as

00:04:44.797 --> 00:04:50.063
an assessor compel you to do that, but I think it's a really good idea.

00:04:50.063 --> 00:04:53.557
Maybe an easy example.

00:04:53.557 --> 00:04:58.036
I encountered a client, I don't know, five or six years ago at this point,

00:04:58.036 --> 00:05:00.998
where all of their applications were written in Java 6,

00:05:00.998 --> 00:05:01.403
right?

00:05:01.403 --> 00:05:05.035
And Java 6 had been dead for a good solid while,

00:05:05.035 --> 00:05:09.845
even Java 7 was either end of life or very nearly so,

00:05:09.845 --> 00:05:11.685
and you know,

00:05:11.685 --> 00:05:15.910
moving things to Java 8 to a supported version of Java

00:05:15.910 --> 00:05:18.979
would've required an enormous engineering effort,

00:05:18.979 --> 00:05:19.505
right?

00:05:19.505 --> 00:05:24.052
Total re-write of not just a single sort of application, but several, right?

00:05:24.052 --> 00:05:26.201
Maybe even as many as a dozen applications.

00:05:26.201 --> 00:05:26.700
Right?

00:05:26.700 --> 00:05:29.689
I mean, that's a huge undertaking.

00:05:29.689 --> 00:05:30.145
Right?

00:05:30.145 --> 00:05:35.790
And Java 6 and Java 7 frankly are riddled with you know,

00:05:35.790 --> 00:05:39.905
nasty security bugs right, many of which don't even have any specification,

00:05:39.905 --> 00:05:42.541
right, since they were you know,

00:05:42.541 --> 00:05:45.134
acquired where a lot of the really severe vulnerabilities are,

00:05:45.134 --> 00:05:49.376
we're not going to tell you, we're just going to say it's really bad,

00:05:49.376 --> 00:05:49.770
right?

00:05:49.770 --> 00:05:49.875
Yeah,

00:05:49.875 --> 00:05:52.208
so I don't even know how you deal with a compensating

00:05:52.208 --> 00:05:56.150
control for unknown vulnerability, like ehhhh,

00:05:56.150 --> 00:05:58.595
it's bad, right?

00:05:58.595 --> 00:05:59.629
Sandbox.

00:05:59.629 --> 00:06:03.037
Yeah, put the application, you know, in a literal sand pit,

00:06:03.037 --> 00:06:08.553
right, and then pour concrete over it, like there's your control.

00:06:08.553 --> 00:06:15.132
So that not dealing with your end of life causes a severe set of problems.

00:06:15.132 --> 00:06:18.306
So year two, you go back to an organization,

00:06:18.306 --> 00:06:22.475
and I'm asking this question because I sort of guess I know the answer to this,

00:06:22.475 --> 00:06:27.148
having been an assessor, having worked in cards.

00:06:27.148 --> 00:06:29.331
Is this is something that people commonly fail on in year two?

00:06:29.331 --> 00:06:30.433
Yes.

00:06:30.433 --> 00:06:31.518
Yeah.

00:06:31.518 --> 00:06:32.845
Yeah, well, it's very easy.

00:06:32.845 --> 00:06:37.360
You know, I think we mistake an idea.

00:06:37.360 --> 00:06:39.507
We talk about patching as a fundamental,

00:06:39.507 --> 00:06:42.220
right, a security fundamental and we've been talking about it since whatever,

00:06:42.220 --> 00:06:48.603
at least the 90s, and I think we conflate that with it's easy.

00:06:48.603 --> 00:06:49.094
Right.

00:06:49.094 --> 00:06:50.089
And it's not.

00:06:50.089 --> 00:06:51.029
No.

00:06:51.029 --> 00:06:53.093
It's a lot of labor, it's a lot of difficulty,

00:06:53.093 --> 00:06:54.755
it's a lot of complexity,

00:06:54.755 --> 00:06:58.490
it's a lot of aggravating the rest of your organization,

00:06:58.490 --> 00:07:02.246
and you know, it doesn't bring in new revenues,

00:07:02.246 --> 00:07:02.533
right?

00:07:02.533 --> 00:07:05.858
So no, it's hard work.

00:07:05.858 --> 00:07:10.572
Now the move to DevOps, is this our savior from vulnerability management?

00:07:10.572 --> 00:07:11.299
Oh, no.

00:07:11.299 --> 00:07:14.593
But we don't patch, we just redeploy a new hardened,

00:07:14.593 --> 00:07:17.964
secure, up-to-date version of the OS?

00:07:17.964 --> 00:07:24.400
Well, if what you're doing is taking your sort of template,

00:07:24.400 --> 00:07:24.758
right,

00:07:24.758 --> 00:07:27.981
and fixing the template and then pushing that deployed

00:07:27.981 --> 00:07:29.489
template out across your environment,

00:07:29.489 --> 00:07:34.570
because you've got some sort of cloud infrastructure or the like,

00:07:34.570 --> 00:07:35.252
okay.

00:07:35.252 --> 00:07:35.911
Right?

00:07:35.911 --> 00:07:40.329
Sure, that can be of use, but yeah,

00:07:40.329 --> 00:07:42.158
and that could conceivably remedy vulnerabilities,

00:07:42.158 --> 00:07:42.829
right?

00:07:42.829 --> 00:07:43.963
Yeah, I met someone who said that we never patch,

00:07:43.963 --> 00:07:45.393
and it's like, you never patch?

00:07:45.393 --> 00:07:48.956
He's like okay, I'm going to write 6.1 and 6.2, so no, here's what we do.

00:07:48.956 --> 00:07:50.917
We take the operating system,

00:07:50.917 --> 00:07:54.108
we take a new copy of the operating system fully up to date,

00:07:54.108 --> 00:07:56.992
fully patched, and then we just redeploy that from here,

00:07:56.992 --> 00:07:59.252
and it's like, yep, that would work.

00:07:59.252 --> 00:08:00.017
Yeah, that would work.

00:08:00.017 --> 00:08:04.793
If you capture all of the relevant software components, sure.

00:08:04.793 --> 00:08:06.902
And the libraries and components?

00:08:06.902 --> 00:08:10.870
This is hard to do, isn't it, because of everything that goes into it.

00:08:10.870 --> 00:08:11.179
Right,

00:08:11.179 --> 00:08:14.818
and I think it gets really tricky in situations where you have

00:08:14.818 --> 00:08:18.709
different teams responsible for different components,

00:08:18.709 --> 00:08:19.082
right,

00:08:19.082 --> 00:08:22.194
or if you're using like a hosted provider that's

00:08:22.194 --> 00:08:24.185
providing some managed services and they say,

00:08:24.185 --> 00:08:26.854
we'll patch the operating system, but you patch the application.

00:08:26.854 --> 00:08:30.723
Well, you know, when you're talking about a Linux or a Unix system,

00:08:30.723 --> 00:08:35.463
where does the operating system end and the application layer begin?

00:08:35.463 --> 00:08:37.441
I don't know.

00:08:37.441 --> 00:08:39.887
And I know, you know, from looking,

00:08:39.887 --> 00:08:40.897
and I speak to a lot of people in the industry still,

00:08:40.897 --> 00:08:42.171
but you know,

00:08:42.171 --> 00:08:44.128
not patching applications like if you're using content

00:08:44.128 --> 00:08:48.397
management systems or a shopping cart, a very popular shopping cart application,

00:08:48.397 --> 00:08:51.160
you know, the fact that that's not patched,

00:08:51.160 --> 00:08:51.710
you know?

00:08:51.710 --> 00:08:53.949
The bad guys can tell when you've not patched.

00:08:53.949 --> 00:08:56.726
It's not, you can't hide this from them.

00:08:56.726 --> 00:08:58.891
And so the second you're at the we have not patched and

00:08:58.891 --> 00:09:01.455
the bad guys can use Google or Showdown and they see

00:09:01.455 --> 00:09:06.739
you're running version Unpatched, you're just advertising your vulnerability.

00:09:06.739 --> 00:09:09.492
You are, right, and if you've sandbagged yourself by,

00:09:09.492 --> 00:09:10.184
you know,

00:09:10.184 --> 00:09:13.980
not fixing Struts or running an old version of Tomcat or

00:09:13.980 --> 00:09:17.654
something like that where the fix is not easy,

00:09:17.654 --> 00:09:17.962
right?

00:09:17.962 --> 00:09:22.170
The fix takes a lot of coordination and development effort, right?

00:09:22.170 --> 00:09:25.666
Then you're advertising, I'm going to be vulnerable for a good, long while.

00:09:25.666 --> 00:09:29.883
So now your web application file wasn't optimism.

00:09:29.883 --> 00:09:35.768
And probably drinking heavily.

00:09:35.768 --> 00:09:38.611
I'm not going to ask you about compromises because we know this is like

00:09:38.611 --> 00:09:41.409
the number one cause of compromise in the industry.

00:09:41.409 --> 00:09:43.881
A big Vegas arrow.

00:09:43.881 --> 00:09:45.754
Yeah, I've no other way of saying, you know,

00:09:45.754 --> 00:09:49.006
patch all the things other than patch all the things.

00:09:49.006 --> 00:09:51.510
Patch all the things now and manage the vulnerabilities.

00:09:51.510 --> 00:09:51.775
Right.

00:09:51.775 --> 00:09:54.269
What I would say just in terms of schedule,

00:09:54.269 --> 00:09:56.692
and this isn't even me, you know,

00:09:56.692 --> 00:10:02.436
crazy ranting; this was in the Verizon DBIR from a handful of years ago.

00:10:02.436 --> 00:10:02.715
Right?

00:10:02.715 --> 00:10:05.511
The mediums and the lows are the things that should

00:10:05.511 --> 00:10:08.641
get caught in your monthly cycle, right?

00:10:08.641 --> 00:10:12.480
The criticals and the highs are the drop everything and patch now.

00:10:12.480 --> 00:10:14.560
Because although the standard says 30 days for a critical,

00:10:14.560 --> 00:10:18.568
the bad guys are not going to wait 30 days for you,

00:10:18.568 --> 00:10:19.301
you know?

00:10:19.301 --> 00:10:22.610
They have a database, they know what version of Struts you're running, you know?

00:10:22.610 --> 00:10:23.754
My client got hit, you know,

00:10:23.754 --> 00:10:27.247
within a day and a half of that vulnerability coming out.

00:10:27.247 --> 00:10:30.728
People know about that; they're looking for you not to patch and then they're

00:10:30.728 --> 00:10:34.042
going to come and attack you; there's no way around it.

00:10:34.042 --> 00:10:36.677
The standard can't say two days because the whole world would be noncompliant,

00:10:36.677 --> 00:10:38.585
but it means two days or two hours.

00:10:38.585 --> 00:10:40.215
Yeah, hustle up!

00:10:40.215 --> 00:10:44.560
Absolutely!

00:10:44.560 --> 00:10:47.042
Compensating controls around this area?

00:10:47.042 --> 00:10:50.230
These are very, very context specific, right?

00:10:50.230 --> 00:10:52.960
So if there's a particular vulnerability that you cannot

00:10:52.960 --> 00:10:56.378
patch because you have some constraint, it breaks some functionality or whatever,

00:10:56.378 --> 00:11:00.659
then my question is, okay, what is that vulnerability,

00:11:00.659 --> 00:11:01.098
right?

00:11:01.098 --> 00:11:06.123
What is the leverage the attacker has, right?

00:11:06.123 --> 00:11:07.936
And then how do you mitigate against that?

00:11:07.936 --> 00:11:08.180
Okay,

00:11:08.180 --> 00:11:10.650
and so that's going to be very case specific and

00:11:10.650 --> 00:11:14.749
trying to do a one-size-fits-all, I think fails sort of categorically.

00:11:14.749 --> 00:11:17.703
But it's an option, if you really have something that will break,

00:11:17.703 --> 00:11:19.704
you can use compensating controls.

00:11:19.704 --> 00:11:20.823
Yes, yes, you can.

00:11:20.823 --> 00:11:22.071
So talk to your QSA early?

00:11:22.071 --> 00:11:23.197
Right.

00:11:23.197 --> 00:11:23.638
Okay, great.

00:11:23.638 --> 00:11:26.048
I think we've wrapped up what I actually think are probably

00:11:26.048 --> 00:11:28.699
the two most important requirements in DSS.

00:11:28.699 --> 00:11:31.411
In fact, in any cyber security exercise, this is it.

00:11:31.411 --> 00:11:32.228
This is it.

00:11:32.228 --> 00:11:35.603
This is it.

00:11:35.603 --> 00:11:41.228
This is the good stuff.
