WEBVTT

00:00:00.190 --> 00:00:01.040
Just prior to this,

00:00:01.040 --> 00:00:04.660
we started talking about host detection capabilities of the osquery

00:00:04.660 --> 00:00:06.750
and Sigma and different rules in use for seams,

00:00:06.750 --> 00:00:09.500
analyzing logs, and file activity, and now we're going to look at it

00:00:09.500 --> 00:00:11.500
from another perspective on the network indicators.

00:00:11.500 --> 00:00:14.210
And we got into it a little bit prior to this because we

00:00:14.210 --> 00:00:17.130
almost had to because once we, first of all,

00:00:17.130 --> 00:00:19.320
when we looked at what the service was doing, we saw that

00:00:19.320 --> 00:00:21.020
there was an interesting port being used.

00:00:21.130 --> 00:00:22.780
It was connecting out to the internet.

00:00:22.790 --> 00:00:22.930
Yeah,

00:00:22.930 --> 00:00:26.340
I think it' was 449 or 447, so I was kind of curious for

00:00:26.340 --> 00:00:28.210
a TCP traffic out to the internet, right?

00:00:28.690 --> 00:00:34.130
And then secondly, under the config, when we unencrypted the

00:00:34.130 --> 00:00:38.200
config, we identified additional ports and IPs and things

00:00:38.200 --> 00:00:38.970
that we should be looking for.

00:00:38.970 --> 00:00:41.430
So we know that those are command and control servers that are looking out.

00:00:41.700 --> 00:00:43.180
Yeah, yeah, those, you know,

00:00:43.540 --> 00:00:46.980
each plugin has its own configuration file, and when you decode them,

00:00:47.450 --> 00:00:50.480
at least we found about a dozen or so different IP addresses and ports that

00:00:50.480 --> 00:00:56.100
would be connecting over.  So we obviously could use those on a network to

00:00:56.100 --> 00:00:57.800
look to see if anybody was connecting to those.

00:00:58.090 --> 00:01:01.350
That would be a great indicator looking looking at those ports to see if

00:01:01.350 --> 00:01:04.280
anybody's connecting to those ports for the unusual ones.

00:01:04.280 --> 00:01:04.769
Obviously,

00:01:04.769 --> 00:01:08.035
for something like 80 or 443 you're going have a ton of traffic to that,

00:01:08.035 --> 00:01:10.650
but the rest of them would be a little bit unusual.

00:01:10.740 --> 00:01:11.650
Yeah, absolutely.

00:01:11.650 --> 00:01:14.500
And when we talk about doing detection, especially,

00:01:14.730 --> 00:01:15.030
well,

00:01:15.040 --> 00:01:19.280
really any type of detection, but talking about network detection, we can

00:01:19.280 --> 00:01:23.240
block those things, that's cool, and we can put detection on very

00:01:23.240 --> 00:01:25.520
specific domains, but then people are using DGAs,

00:01:25.520 --> 00:01:25.820
right?

00:01:25.820 --> 00:01:29.440
And so that's like cat and mouse game. And you can put detection on specific

00:01:29.440 --> 00:01:31.530
IPs they're going out to, but those change as well. Right,

00:01:31.540 --> 00:01:32.090
right..

00:01:32.090 --> 00:01:33.460
And the configs can change.

00:01:33.470 --> 00:01:39.150
I usually try to look for more kind of behavioral or maybe you call

00:01:39.540 --> 00:01:44.100
metadata about the network traffic that is going to be more consistent

00:01:44.100 --> 00:01:47.180
regardless of changing IP or a port or that kind of stuff. Right.

00:01:47.180 --> 00:01:51.540
Especially now that know a lot of traffic is encrypted using SSL or TLS,

00:01:51.650 --> 00:01:53.620
you can't just pure into the traffic.

00:01:53.630 --> 00:01:58.980
You've got to look for other indicators related to the SSL certificates or

00:01:58.980 --> 00:02:02.030
something like that in order to hopefully find some indicators.

00:02:02.140 --> 00:02:02.990
Yeah, absolutely.

00:02:02.990 --> 00:02:05.470
And if you could decrypt it, or even if you can't,

00:02:05.470 --> 00:02:08.729
there's still signatures, so we can use Snort or we can use Suricata, and we

00:02:08.729 --> 00:02:12.320
can alert on odd ports that we find, right? But first,

00:02:12.320 --> 00:02:14.520
we have to find the ports, so I think there's usually a ton of

00:02:14.520 --> 00:02:18.220
information to worry about with that. You gave me a pcap that was from

00:02:18.230 --> 00:02:21.770
that last VM where we detonated everything after it started downloading

00:02:21.950 --> 00:02:25.140
the main executable that was inside speedLan.

00:02:25.660 --> 00:02:29.860
So the speedLan executable was communicating out and downloading those

00:02:29.860 --> 00:02:32.290
modules, and so hopefully there's some of that data in there.

00:02:32.290 --> 00:02:36.320
And you gave that to me to or did some analysis in Bro, or Zeek now, right?

00:02:36.330 --> 00:02:36.730
Right.

00:02:36.770 --> 00:02:38.950
Well, it's Zeek now, but if you install it,

00:02:38.960 --> 00:02:42.520
you still have to install Bro, so. And all the commands are Bro.

00:02:42.530 --> 00:02:48.740
There's like a guide on zeek.org to tell you how to make aliases

00:02:48.740 --> 00:02:50.800
to call Zeek. And I'm like that's extra work.

00:02:50.800 --> 00:02:51.490
I'm not going to do it.

00:02:51.500 --> 00:02:54.140
Oh, yeah, just continue on with the way it is.

00:02:54.150 --> 00:02:54.540
Yeah, right.

00:02:54.540 --> 00:02:57.990
So you're not familiar with Bro, it's used heavily in a couple different

00:02:57.990 --> 00:03:00.280
courses, Security Event Triage: Signature and Session Analysis,

00:03:00.280 --> 00:03:04.200
as well as from the same path there's a network behavioral

00:03:04.200 --> 00:03:07.770
analysis course that uses Bro and malloc and a lot of this

00:03:07.770 --> 00:03:08.900
kind of session analysis stuff.
