WEBVTT

00:00:00.040 --> 00:00:03.200
[Autogenerated] Obviously logging depends on certainly consolidated

00:00:03.200 --> 00:00:07.200
logging across lots of systems depends on every system Agreeing that

00:00:07.200 --> 00:00:08.890
the time of day is the same time of day.

00:00:08.900 --> 00:00:09.250
Yes.

00:00:09.250 --> 00:00:11.670
Which is why we've got requirement 10 4.

00:00:11.670 --> 00:00:12.190
Yeah.

00:00:12.200 --> 00:00:18.650
On synchronizing time between devices, that's a lot of requirement in 10.4.

00:00:18.650 --> 00:00:20.360
There's a lot of stuff in 10.4,

00:00:20.360 --> 00:00:23.660
and and this is kind of my least favorite part of PC IDE s s,

00:00:23.640 --> 00:00:24.130
uh,

00:00:24.140 --> 00:00:29.450
because because there's a lot of stuff and a lot of it is kind of the minutia.

00:00:29.460 --> 00:00:29.970
Yes.

00:00:29.980 --> 00:00:33.440
You want time synchronization across all of your devices?

00:00:33.450 --> 00:00:35.750
Uh, you know, like I mentioned in the requirement nine stuff.

00:00:35.750 --> 00:00:39.700
I've encountered CCTV systems where the timestamp was wrong, right?

00:00:39.710 --> 00:00:41.930
And and that's really problematic, right?

00:00:41.930 --> 00:00:45.370
Or if you have differing time stamps across different system components,

00:00:45.440 --> 00:00:48.000
um, now these requirements were written,

00:00:48.010 --> 00:00:50.670
you know, in the earlier part of the last decade,

00:00:50.680 --> 00:00:56.340
find large on, have stayed more or less the same over over this period of time.

00:00:56.350 --> 00:01:01.190
So a lot of a lot of the requirement difficulty I find in the NTP

00:01:01.190 --> 00:01:06.190
section is the idea that the organization has its own one or two

00:01:06.190 --> 00:01:08.330
times servers that everything points to.

00:01:08.330 --> 00:01:11.450
And then they go out to some internet time source the,

00:01:11.540 --> 00:01:11.820
you know,

00:01:11.820 --> 00:01:16.350
NIST or pool or the Navy Observatory or whatever to get some time source on.

00:01:16.350 --> 00:01:17.330
Then propagate that back.

00:01:17.340 --> 00:01:19.490
And that was intended so that you don't have,

00:01:19.500 --> 00:01:22.960
you know, drift or like we don't care what PICO second it is.

00:01:22.970 --> 00:01:25.650
We care that the time stamps error All consistent, right.

00:01:25.660 --> 00:01:30.360
Uh, I think that is increasingly sort of a less relevant,

00:01:30.370 --> 00:01:33.270
uh, in terms of preventing sort of drift problems.

00:01:33.280 --> 00:01:37.390
And the need for an organization to have a couple of internal

00:01:37.390 --> 00:01:39.200
time service that everybody points to is increasingly

00:01:39.200 --> 00:01:41.810
problematic in a cloud based environment,

00:01:41.810 --> 00:01:42.650
for example,

00:01:42.660 --> 00:01:45.900
where spitting up your own time server than poking holes in

00:01:45.900 --> 00:01:48.770
the VPC to point everything to it right,

00:01:48.770 --> 00:01:53.790
rather than just getting the cloud providers time service,

00:01:53.800 --> 00:01:56.290
you know that every host can point to directly,

00:01:56.300 --> 00:01:58.510
uh, you know, sometimes causes problems.

00:01:58.510 --> 00:02:00.730
And so I've had clients have to basically concoct a

00:02:00.730 --> 00:02:04.360
compensating control for everybody goes out to,

00:02:04.840 --> 00:02:06.420
you know, the one time service, right?

00:02:06.420 --> 00:02:07.990
Every host, because as of the one time service,

00:02:07.990 --> 00:02:12.260
rather than they have their own internal time server.

00:02:12.260 --> 00:02:12.420
Right?

00:02:12.420 --> 00:02:15.560
Because the notion of internal network in cloud

00:02:15.560 --> 00:02:16.730
environment starts to break down.

00:02:16.740 --> 00:02:18.050
Absolutely only.

00:02:18.440 --> 00:02:19.320
And as you say,

00:02:19.320 --> 00:02:21.740
if you if you have to start punching holes in your virtual

00:02:21.740 --> 00:02:25.050
private networking or your CloudWatch for someone to go out to

00:02:25.050 --> 00:02:30.260
get a time zone getting a little yeah, yeah,

00:02:30.260 --> 00:02:32.550
I mean, I remember used to use that, you know.

00:02:32.560 --> 00:02:34.990
In fact, actually, recently I have Did, you know,

00:02:35.000 --> 00:02:39.020
make sure we had areas on the roof and we're picking up time from GPS,

00:02:39.040 --> 00:02:41.320
which has got the easiest way to do it for a data center,

00:02:41.320 --> 00:02:41.790
for example.

00:02:41.800 --> 00:02:46.860
Yeah, but this whole thing about having to have these times services a bit,

00:02:46.870 --> 00:02:49.210
okay, it can be, Yes.

00:02:49.220 --> 00:02:51.800
The requirement should just be that everything has the same time.

00:02:51.810 --> 00:02:53.050
I don't care how you do it, right?

00:02:53.060 --> 00:02:53.830
Yeah, right.

00:02:53.840 --> 00:02:55.460
In a more ideal world,

00:02:55.460 --> 00:03:00.790
that would be how you would approach it and finally stuck in 10 for some reason.

00:03:00.800 --> 00:03:04.500
That I don't really understand is fine integrity monitoring, isn't it?

00:03:04.510 --> 00:03:08.510
Well, there's a part of on the audit log on the audit logs, right?

00:03:08.510 --> 00:03:11.360
And I think that hearkens back to a day where the

00:03:11.360 --> 00:03:16.210
centralized repository is a collection of files,

00:03:16.220 --> 00:03:16.620
right?

00:03:16.630 --> 00:03:17.480
Right now,

00:03:17.490 --> 00:03:21.600
a very few clients where I don't mark 10.55 is not app

00:03:21.600 --> 00:03:25.240
clickable because for the most part, when they collect and centralized blog data,

00:03:25.240 --> 00:03:28.430
those log events are records in a database,

00:03:28.440 --> 00:03:29.200
right?

00:03:29.210 --> 00:03:33.330
And so file level integrity checking doesn't really

00:03:33.330 --> 00:03:35.070
work in a relational database.

00:03:35.080 --> 00:03:42.750
Uh, now it's still very possible to have, you know, files getting centralized.

00:03:42.760 --> 00:03:46.310
In which case, yeah, you want some sort of file integrity monitoring there.

00:03:46.310 --> 00:03:50.220
But that's that's not often the case,

00:03:50.230 --> 00:03:52.530
because putting want to database doesn't really work.

00:03:52.530 --> 00:03:54.120
No changes, No.

00:03:54.120 --> 00:03:54.310
I mean,

00:03:54.310 --> 00:03:56.070
if you want it to go off all the time and train

00:03:56.070 --> 00:03:59.100
your staff to ignore the similar, it's absolutely then do that,

00:03:59.100 --> 00:03:59.560
yeah.
