WEBVTT

00:00:00.040 --> 00:00:02.280
[Autogenerated] so thinking now of requirement 10,

00:00:02.290 --> 00:00:04.650
you're going on a successor for the second year.

00:00:04.660 --> 00:00:05.420
What?

00:00:05.430 --> 00:00:05.660
What?

00:00:05.660 --> 00:00:06.820
What typically goes wrong?

00:00:06.820 --> 00:00:11.570
What people get wrong in this, Um, you know, if they've done it, right?

00:00:11.940 --> 00:00:12.580
Actually,

00:00:12.590 --> 00:00:16.490
the thing that I find that happens the most is that the person

00:00:16.490 --> 00:00:19.020
whose job it was to sort of deal with the alerts,

00:00:19.030 --> 00:00:20.440
um, you know,

00:00:20.450 --> 00:00:25.690
maybe stops doing it or doesn't have sort of good

00:00:25.700 --> 00:00:27.360
kind of records keeping from it, right?

00:00:27.360 --> 00:00:30.240
So if it's I have a task to go look at alerts, right?

00:00:30.250 --> 00:00:35.470
And then all the output was was I looked at the alerts today, right?

00:00:35.840 --> 00:00:38.890
You know, two weeks later, you're not gonna remember what you did, right?

00:00:38.900 --> 00:00:42.690
So when I say tell me about how you follow up on alerts that error significant,

00:00:42.700 --> 00:00:44.700
and they sort of look at you with that funny face and kind of

00:00:44.700 --> 00:00:47.420
scratch their head like I don't remember,

00:00:47.430 --> 00:00:51.590
um, because they didn't take any time to document what they did.

00:00:51.590 --> 00:00:55.470
Now, if there's a a ticket right or something like that,

00:00:55.470 --> 00:00:57.480
that says I had this alert here.

00:00:57.480 --> 00:00:58.470
Here's what it told me.

00:00:58.480 --> 00:00:59.850
Here's what I investigated.

00:00:59.850 --> 00:01:03.570
Here's what I found even if the end result is,

00:01:03.940 --> 00:01:06.400
you know, that wasn't really anything particularly relevant.

00:01:06.410 --> 00:01:07.770
Or it was a false alarm.

00:01:07.780 --> 00:01:09.130
Or, you know,

00:01:09.130 --> 00:01:11.770
it was just some ____ on the Internet scanning me with some crummy

00:01:11.770 --> 00:01:14.960
vulnerability scanning tool That happens 40 times a day,

00:01:14.970 --> 00:01:15.450
you know?

00:01:15.450 --> 00:01:20.490
Then then you have some record of the effectiveness of your loggers.

00:01:20.500 --> 00:01:21.160
Right?

00:01:21.170 --> 00:01:23.560
And you have some metric to track on that.

00:01:23.540 --> 00:01:27.820
So you look so the failure point is people stop doing it.

00:01:27.820 --> 00:01:29.830
And if they are doing it, they're not keeping good records.

00:01:29.840 --> 00:01:33.090
Show right that they can prove that they're actually doing something worthwhile.

00:01:33.090 --> 00:01:33.860
Is opposes going.

00:01:33.860 --> 00:01:34.970
Yeah, Yeah, the logs.

00:01:34.970 --> 00:01:35.040
Error.

00:01:35.040 --> 00:01:35.410
Fine.

00:01:35.420 --> 00:01:35.770
Right.

00:01:36.240 --> 00:01:39.990
What about changes to the c D and making sure new devices

00:01:39.990 --> 00:01:41.700
get brought into the logging family?

00:01:41.710 --> 00:01:43.450
That can happen.

00:01:43.460 --> 00:01:43.750
Right.

00:01:43.760 --> 00:01:49.590
Um, I actually don't see that as a failure particularly often,

00:01:49.590 --> 00:01:50.930
but it certainly can happen.

00:01:50.940 --> 00:01:51.570
Um,

00:01:51.580 --> 00:01:55.890
the other thing that I'm I'm expecting sort of more failures for in

00:01:55.890 --> 00:01:59.600
the future is the new requirement stuff in 10.8,

00:01:59.610 --> 00:02:00.000
right?

00:02:00.010 --> 00:02:02.200
Which is about tracking failures, right?

00:02:02.210 --> 00:02:04.360
That That's to say, Hey, if something has gone wrong,

00:02:04.940 --> 00:02:08.949
do some, you know, identify it do some root cause analysis,

00:02:08.960 --> 00:02:12.780
figure out what your window of exposure was and then fix it right on.

00:02:12.780 --> 00:02:15.730
That's a really interesting sort of set of steps,

00:02:15.740 --> 00:02:20.850
Uh, in that new requirement on, I think you get kind of two sorts of failures.

00:02:20.860 --> 00:02:23.370
This is a requirement for service providers only GUID.

00:02:23.370 --> 00:02:23.960
Is IT?

00:02:23.940 --> 00:02:24.360
It is.

00:02:24.360 --> 00:02:26.200
It is requirement specifically for service providers.

00:02:26.270 --> 00:02:28.340
I think you get two sorts of failures, though, right?

00:02:28.340 --> 00:02:30.880
You get the kind of active failure, right?

00:02:30.890 --> 00:02:32.740
A device goes down, right?

00:02:32.750 --> 00:02:36.080
A thing doesn't respond to health or uptime Check.

00:02:36.090 --> 00:02:36.530
Right.

00:02:36.540 --> 00:02:40.530
Um, and and those error relatively obvious.

00:02:40.530 --> 00:02:40.790
Right?

00:02:40.790 --> 00:02:44.110
And sussing that out and going and doing something about it.

00:02:44.120 --> 00:02:46.000
Pretty straightforward, right?

00:02:46.010 --> 00:02:48.880
Then there's the more sort of silent failure, right?

00:02:48.880 --> 00:02:53.970
Your intrusion detection sensor doesn't detect anything, right?

00:02:53.980 --> 00:02:55.840
Does that mean that there was nothing to detect or it

00:02:55.840 --> 00:02:57.890
just isn't doing anything harder to say?

00:02:57.900 --> 00:03:00.270
So So So what we're talking about really,

00:03:00.270 --> 00:03:02.360
is to sort of fairly we've got failure,

00:03:02.360 --> 00:03:04.450
effectiveness and the failure of operations.

00:03:04.450 --> 00:03:04.860
Right?

00:03:05.440 --> 00:03:07.750
Operational effectiveness is pretty easy, isn't it?

00:03:07.760 --> 00:03:10.260
It's like it's not working anyway.

00:03:10.270 --> 00:03:11.710
Someone turned it off.

00:03:11.720 --> 00:03:11.990
Right?

00:03:12.000 --> 00:03:14.300
Effectiveness is quite hard to test, although, isn't it?

00:03:14.310 --> 00:03:18.480
I think it can be hard to test for, um, there are.

00:03:18.480 --> 00:03:21.020
There are some cases where I think it's relatively easy, right?

00:03:21.020 --> 00:03:23.590
At a client a tous point many years ago,

00:03:23.600 --> 00:03:27.220
who is paying a third-party to do their sort of intrusion detection.

00:03:27.540 --> 00:03:27.820
Right.

00:03:27.820 --> 00:03:28.620
And I was on site.

00:03:28.620 --> 00:03:29.400
I was saying, Well, you know,

00:03:29.400 --> 00:03:31.380
what kind of things are they providing you with them?

00:03:31.380 --> 00:03:33.150
Like, Oh, you know, we haven't heard anything from them.

00:03:33.840 --> 00:03:34.300
What do you mean?

00:03:34.300 --> 00:03:35.350
You haven't heard anything from them?

00:03:35.550 --> 00:03:37.650
What we do, they just haven't set up any alerts.

00:03:38.540 --> 00:03:43.680
Well, like our CSV services just did a vulnerability scan last week did that.

00:03:43.690 --> 00:03:46.590
I mean, you're not getting alert.

00:03:46.590 --> 00:03:48.310
Like we've never seen an alert from your scans.

00:03:48.310 --> 00:03:49.840
Well, we did a pen test before.

00:03:49.850 --> 00:03:52.070
Did you see any of that?

00:03:52.080 --> 00:03:52.650
No.

00:03:52.900 --> 00:03:56.870
Okay, so So you're writing somebody a check every month,

00:03:57.540 --> 00:03:59.460
and then they're not really doing anything for you.

00:04:00.640 --> 00:04:01.480
Yeah, great.

00:04:01.490 --> 00:04:03.000
But they bought something, right?

00:04:03.000 --> 00:04:04.350
But but But we paid them right.

00:04:04.350 --> 00:04:07.570
So So clearly that must get us an in place, right?

00:04:08.440 --> 00:04:10.040
10.8 for service providers.

00:04:10.050 --> 00:04:12.820
Do you try and encourage merchants to do it.

00:04:12.830 --> 00:04:14.560
I mean non service providers to do as well.

00:04:14.570 --> 00:04:17.910
Um, e can't insist on it.

00:04:17.920 --> 00:04:18.940
I you know,

00:04:18.940 --> 00:04:21.320
sometimes it comes up and I say this is a really interesting practice.

00:04:21.320 --> 00:04:24.340
You should consider it, you know,

00:04:24.340 --> 00:04:28.910
And usually they give me that sort of glazed overlook Knowing

00:04:28.920 --> 00:04:30.860
the council on the way the council work.

00:04:31.540 --> 00:04:33.750
If a requirement goes in like this,

00:04:33.760 --> 00:04:37.100
that says for service providers only on it's not about

00:04:37.100 --> 00:04:38.720
things that only a service provider could do,

00:04:38.720 --> 00:04:41.050
like keep separate logs for different organizations.

00:04:41.640 --> 00:04:45.110
There would have been a discussion around the table saying,

00:04:45.250 --> 00:04:47.990
Should we put this to everybody, right?

00:04:48.000 --> 00:04:51.900
I think that's a very realistic possibility that yeah,

00:04:51.900 --> 00:04:53.840
I mean, that's typically what happens, right?

00:04:53.850 --> 00:04:57.370
And people look at it and they say that XAML put it to

00:04:57.370 --> 00:05:00.360
service providers to begin with, see how it gets adopted,

00:05:00.370 --> 00:05:03.990
see if it produces results on if it's successful,

00:05:04.000 --> 00:05:06.310
will roll it out to the rest of the community s.

00:05:06.310 --> 00:05:07.310
So my my view,

00:05:07.310 --> 00:05:10.250
whenever I see a new requirement in the standard for a service provider,

00:05:10.250 --> 00:05:12.780
is I need to be thinking of in three years time,

00:05:12.780 --> 00:05:14.520
that's gonna be on everyone's in everyone.

00:05:14.530 --> 00:05:17.730
That seems like a likely I can right Is really just a question of timing.

00:05:17.740 --> 00:05:21.070
Yes, it is, Um, and will have a notice period,

00:05:21.070 --> 00:05:23.580
as usual, but but but it's actually quite a good thing.

00:05:23.580 --> 00:05:26.890
I mean, you know, incorporate in corporate information security.

00:05:26.900 --> 00:05:29.920
I am known as someone who is just really annoying.

00:05:29.920 --> 00:05:34.310
I just annoy people about control, effectiveness and control operations.

00:05:34.320 --> 00:05:36.960
You know, it's not good enough just to buy it.

00:05:36.940 --> 00:05:40.030
It has to be turned on and working, and we have to know it's turned out well,

00:05:40.030 --> 00:05:43.660
you have to do something with it and has to It has to improve things.

00:05:43.840 --> 00:05:46.570
Eso yeah, if you're If you've got some other control,

00:05:46.570 --> 00:05:48.070
we've got some intrusion, detection or whatever,

00:05:48.070 --> 00:05:51.790
and it's not catching intrusions or it's not compelling you to take action.

00:05:51.800 --> 00:05:53.940
Yeah, I mean, whatever, you know, open family,

00:05:53.950 --> 00:05:54.230
you know,

00:05:54.230 --> 00:05:56.220
you commission pan tests and you get the pen test

00:05:56.220 --> 00:05:58.160
results on you go to the I T people.

00:05:58.160 --> 00:05:58.310
Go.

00:05:58.320 --> 00:05:58.730
Look.

00:05:58.740 --> 00:06:01.920
Look at the pen test result we have We have controls against that.

00:06:01.920 --> 00:06:02.600
How did they get through?

00:06:02.600 --> 00:06:04.990
That is like it wasn't configured properly or Oh,

00:06:04.990 --> 00:06:06.970
yeah, Fred was going to update it next week.

00:06:06.980 --> 00:06:08.960
There's always sort of a reason behind it.

00:06:08.940 --> 00:06:13.710
A control failure, you know, a nonworking control is like having no control,

00:06:13.720 --> 00:06:14.040
right?

00:06:14.050 --> 00:06:15.450
You know, right, Exactly.

00:06:16.740 --> 00:06:22.640
Um so obviously favors of logging are not are not causes of compromises,

00:06:22.650 --> 00:06:26.850
but failures of logging on responding to blog events

00:06:27.340 --> 00:06:30.810
really lengthen the time it takes to detect something

00:06:30.810 --> 00:06:33.100
that's not right in your right well.

00:06:33.100 --> 00:06:33.810
And so,

00:06:33.810 --> 00:06:38.100
looking again at those sort of statistics about weeks or months

00:06:38.100 --> 00:06:40.450
until the organization finds out about it,

00:06:40.460 --> 00:06:43.730
in many cases these organization is never the one who detects it,

00:06:43.740 --> 00:06:43.960
right?

00:06:43.970 --> 00:06:47.260
It's the it's the bank or the payment processor.

00:06:47.340 --> 00:06:49.750
Who says I think you're a common point of purchase for

00:06:49.750 --> 00:06:51.270
all of these fraudulent transactions.

00:06:51.270 --> 00:06:53.130
I think maybe you've been breached, right?

00:06:53.130 --> 00:06:55.740
So, like at some point it stops being you that finds out right?

00:06:55.750 --> 00:06:58.870
It starts being somebody else who complains about it.

00:06:59.440 --> 00:07:00.640
I think you know we should.

00:07:00.650 --> 00:07:05.320
We should hark back to the very, very public breach of a large retailer in the U.

00:07:05.320 --> 00:07:05.790
S.

00:07:05.800 --> 00:07:07.980
Where they're logging system was telling them there

00:07:07.980 --> 00:07:10.180
was bad activity on the network, you know,

00:07:10.180 --> 00:07:13.350
And that came out in the reports that were very public afterwards.

00:07:13.360 --> 00:07:15.790
You know, they just the response wasn't there is like,

00:07:15.800 --> 00:07:18.070
you know, it's just not gonna do anything about it.

00:07:18.040 --> 00:07:21.370
Well, and what I read about that case in particular,

00:07:21.640 --> 00:07:24.280
uh, you know, somebody was responding to an interview and saying,

00:07:24.280 --> 00:07:25.740
Yeah, okay, there were alert saying, Hey,

00:07:25.740 --> 00:07:30.990
there's this malicious activity but I get 50 of those alerts every day,

00:07:31.000 --> 00:07:31.460
right?

00:07:31.470 --> 00:07:36.760
And picking out the good stuff for the most salient stuff

00:07:36.760 --> 00:07:39.680
from the less salient stuff is hard right?

00:07:39.690 --> 00:07:40.980
And And that, I think,

00:07:40.980 --> 00:07:42.680
points you to the idea that you really want to

00:07:42.680 --> 00:07:47.330
figure out how to tune your analysis, you know,

00:07:47.330 --> 00:07:50.990
to your organization so that you can find the really bad stuff right?

00:07:51.000 --> 00:07:52.180
And that's easier said than done.

00:07:52.180 --> 00:07:55.600
But it's worth doing so the end of the day we're talking,

00:07:55.600 --> 00:07:57.890
having a scene, managing it properly.

00:07:57.900 --> 00:07:58.340
Response.

00:07:58.340 --> 00:07:58.520
DLL.

00:07:58.520 --> 00:08:00.180
That's properly IT great
