1
00:00:06,940 --> 00:00:13,760
So, anyways, I'm Dan Burrows, just a little bit about myself.

2
00:00:13,760 --> 00:00:19,200
I'm a professor at the University of Central Florida in the Engineering Technology Department.

3
00:00:19,200 --> 00:00:32,260
Specifically, I work in the Information Systems Technology Program, which is a, it's a very applied approach to engineering and specifically information technology systems.

4
00:00:32,260 --> 00:00:43,040
We're currently at the University of Central Florida, developing a undergraduate security program that I'm also, I'm the one that's mainly involved in developing that.

5
00:00:43,040 --> 00:00:46,640
And that's actually what my second talk a little bit later is on today.

6
00:00:46,660 --> 00:00:51,420
But this is the more, maybe more interesting, more technical one.

7
00:00:51,640 --> 00:00:59,780
My past research includes the correlation of distributed network sensors.

8
00:00:59,780 --> 00:01:16,960
If you have a, if you have a large scale network, you have intrusion detection systems, you have firewalls, you have other things that are recording information saying what's going on on the network, trying to correlate that information together to bring it all together and trying to see the big picture of what's going on,

9
00:01:16,960 --> 00:01:23,020
rather than getting these little kind of half views of what's really occurring out on the network.

10
00:01:23,020 --> 00:01:31,100
And also I'm working on some projects that involve combined data sharing networks for law enforcement in the state of Florida.

11
00:01:32,540 --> 00:01:37,480
You know, this little blurb ad about the University of Central Florida, it's down in Orlando, Florida.

12
00:01:37,480 --> 00:01:42,240
It's about, we've got about 42,000 students there, about 5,500 of which are in engineering.

13
00:01:42,280 --> 00:01:46,740
We're located about 13 miles east of Orlando, Florida.

14
00:01:48,480 --> 00:01:54,740
As I said, the Information System Technology Program, this is a really, it's an applied engineering degree.

15
00:01:54,740 --> 00:02:02,620
It's very hands on, very real world, very current tools, not so much the traditional theoretical approach to engineering.

16
00:02:03,700 --> 00:02:07,220
Heavy concentration on hands on learning in the program.

17
00:02:08,320 --> 00:02:19,040
A little bit about the background of what led me to this current idea is, there was a previous research that I worked on.

18
00:02:19,040 --> 00:02:37,200
I gave a couple of presentations back at DEF CON 9 and 10, and this was on using Bayesian multiple hypothesis to try and analyze reports from these distributed intrusion detection systems and other types of network sensors, with the idea of trying to determine if what you were seeing were correlated events,

19
00:02:37,200 --> 00:02:48,100
if this was an initial attack and a follow-up, or if the two events you were seeing in two different places might be two systems communicating who shouldn't be communicating together.

20
00:02:48,100 --> 00:02:54,040
And to try and use multiple sensors to distribute over large distributed networks to try and do this.

21
00:02:54,040 --> 00:03:13,600
The basic idea behind what I was doing with that was taking target tracking methods that came out of, actually out of radar target tracking systems, trying to track real objects in the sky with various radar systems, and taking that same approach and treating all of our sensors that we have on our network,

22
00:03:13,600 --> 00:03:23,040
whether it's an IDS, a firewall, a router, even a human, sat there making observations about what's going on, and trying to automate the correlation of that together.

23
00:03:23,040 --> 00:03:36,580
And this wasn't intended to totally throw the human analyst out of the loop, but just to kind of act as a sort of a first level filtering system to throw away a lot of the information that you don't need in order to come to a conclusion about what's going on,

24
00:03:36,580 --> 00:03:41,960
with the idea that we were going to kind of move this more towards a point of view.

25
00:03:41,960 --> 00:03:48,340
Instead of looking at this one section of the network or this one part of what was going on that's visible to this one sensor.

26
00:03:48,340 --> 00:03:52,260
Like, you know, you've got your firewall sat at the, say you've got it sat at the border of your network.

27
00:03:52,260 --> 00:03:53,020
It sees everything that's going on.

28
00:03:53,040 --> 00:03:54,480
It sees everything that's coming in and out of the border.

29
00:03:54,480 --> 00:03:57,240
But it doesn't see stuff that's entirely internal to the network.

30
00:03:57,320 --> 00:04:04,400
You might have an intrusion detection system sat on one segment of the network that sees what's going on there, but it doesn't see what's going on on another segment of the network.

31
00:04:04,400 --> 00:04:13,820
And to take all these little pictures, pieces of information, and to put them together to get a more complete view of what someone was doing while they were attacking the network.

32
00:04:13,840 --> 00:04:21,440
And did this through using radar tracking algorithms, specifically Bayesian multiple hypothesis tracking.

33
00:04:22,740 --> 00:04:34,580
And making an analogy to the world of computer security where our firewalls, our IDS, our humans looking at what's going on on the network, anything that records information is a sensor.

34
00:04:34,580 --> 00:04:36,100
That's like a radar system.

35
00:04:36,180 --> 00:04:41,820
The actions that the person out there trying to break into the network are doing, that's the target.

36
00:04:41,820 --> 00:04:43,480
And that's the trace they leave behind.

37
00:04:43,480 --> 00:04:45,880
And we're trying to put these traces together.

38
00:04:45,980 --> 00:04:49,400
And we're working in a situation where we've got to do this in real time.

39
00:04:49,400 --> 00:04:50,920
We've got incomplete data.

40
00:04:50,920 --> 00:04:56,100
We've got inaccurate data, potentially, about what's going on, depending on how your sensors are operating.

41
00:04:57,620 --> 00:05:02,600
Just a little bit more about the system.

42
00:05:03,320 --> 00:05:16,080
The idea, how this works, this idea of this multiple hypothesis tracking is as events occur on the network, as an event is when a sensor recognizes that something's occurring.

43
00:05:16,080 --> 00:05:30,700
When something actually pops up on the sensor, when it gives a report, when you've got an IDS that tells you, hey, I think I see something, when you've got a firewall that decides it's not going to allow this packet or it sees something unusual, when you've got your sysadmin on your network saying,

44
00:05:30,700 --> 00:05:32,880
hey, you know, something weird's going on here.

45
00:05:32,880 --> 00:05:35,940
As these events arrive, they're analyzed.

46
00:05:35,940 --> 00:05:41,500
And say, for example, we're sat there and one of our systems sees a port scan happening out there.

47
00:05:41,780 --> 00:05:45,720
We build a hypothesis about what we think's going on.

48
00:05:45,720 --> 00:05:52,960
In this case, if we've got a single scan coming in from the outside world, hey, you know, that's all the information we have.

49
00:05:52,960 --> 00:05:56,960
That's the only hypothesis we can come to is that that's what's going on.

50
00:05:56,960 --> 00:05:58,960
We're seeing a port scan right now.

51
00:05:59,260 --> 00:06:03,060
Really, I guess, actually, technically, you can make two hypotheses at this point.

52
00:06:03,060 --> 00:06:11,200
You can say, hey, there's a port scan going on, or alternatively, my IDS isn't working properly, and this is a false positive.

53
00:06:12,560 --> 00:06:18,500
Then, as other events occur, say you see a buffer overflow attack against a machine following that.

54
00:06:18,500 --> 00:06:21,080
Well, what could be going on now?

55
00:06:21,080 --> 00:06:26,280
We could make the assumption that the port scan is being followed up by a buffer overflow.

56
00:06:26,280 --> 00:06:30,960
Someone scanned a machine for a vulnerability, then they attacked that particular vulnerability on it.

57
00:06:30,960 --> 00:06:43,000
Or, potentially, these could be two unrelated events, that the port scan could be part of one attack, the buffer overflow could be part of a different attack on the system.

58
00:06:44,640 --> 00:06:49,940
And what we were trying to do with this system was trying to automate the process of figuring these out.

59
00:06:49,940 --> 00:06:52,400
Now, this is a simple example.

60
00:06:52,400 --> 00:06:54,280
This is a pretty trivial case.

61
00:06:54,360 --> 00:07:09,200
In this particular case, if the port scan is a scan against, I don't know, against port 80, and the buffer overflow is a buffer overflow against an IIS server, chances are those two things are related.

62
00:07:09,200 --> 00:07:14,240
If the port scan doesn't match up with what the buffer overflow was, chances are they're not related.

63
00:07:14,240 --> 00:07:31,660
But, you can extrapolate from this idea into much more complicated systems and much more complicated series of events, of trying to piece together the pieces and figure out whether or not all of these events and all of these things that we see occurring across the network are related together,

64
00:07:31,660 --> 00:07:43,800
are they not related together, are they likely to be something real, are they likely to be a false positive caused by a imperfect sensor, imperfect IDS, or firewall, or human in the loop.

65
00:07:45,720 --> 00:07:53,920
Actually, when I was working on this, one of the most interesting sources of data that I had to work with came from the Capture the Flag games here at DEF CON.

66
00:07:53,920 --> 00:07:58,740
Luckily, the SHMOO group captures all of that data that goes on during the Capture the Flag game.

67
00:07:58,740 --> 00:08:00,540
They capture all of that network traffic.

68
00:08:00,820 --> 00:08:08,200
And that provided a great playground to work in, to try and see if that information could be correlated together.

69
00:08:08,240 --> 00:08:12,040
And this little graph I got up here, I mean, this is kind of just pulled right out of context.

70
00:08:12,040 --> 00:08:19,240
These are some slides I grabbed just to give a bit of background about the previous research in this.

71
00:08:20,400 --> 00:08:42,020
Basically, what this was showing is, as you let the number of hypotheses that you consider increase, in that previous example, I said that we had the two hypotheses, either these two events are unrelated, or these two events are part of a chain of a single attack.

72
00:08:42,680 --> 00:08:45,860
Well, we let those, we say, these are our two hypotheses.

73
00:08:45,860 --> 00:08:54,220
And then we look at the evidence that's going on, and we evaluate them, and we find out which hypotheses we think is currently the best one we have available.

74
00:08:54,640 --> 00:08:57,740
Then, we wait for more information to come in.

75
00:08:57,740 --> 00:09:02,060
We'll have a certain confidence level, and how confident do we believe in this hypothesis?

76
00:09:02,760 --> 00:09:11,760
As more information comes in, we'll build more hypotheses, and we re-evaluate the likelihood of our existing hypotheses.

77
00:09:11,760 --> 00:09:23,020
Eventually, at least in theory, this will eventually converge to a point where one hypothesis is clearly the winning hypothesis.

78
00:09:23,120 --> 00:09:34,940
Now, the problem with it, I said that's in theory, the problem with that is, you have to consider every possible hypothesis in order to do that, and the number of hypotheses grows at a super-exponential rate.

79
00:09:34,940 --> 00:09:37,340
So, it's practically impossible to do that.

80
00:09:37,340 --> 00:09:42,580
There's a lot of work on the pruning of them, deciding which ones you want to keep, which ones you want to throw away.

81
00:09:42,700 --> 00:09:58,820
And what this graph is just showing here was, along the bottom is a number of hypotheses that were allowed to remain in the system, going from, well, at the bottom one, all the way up to 1,000, and the accuracy of how well it correlated the different attacks together.

82
00:09:58,820 --> 00:10:12,040
And it never quite got up to 100%, but it got up into about 93, 94% accuracy on the capture-the-flag data, being able to piece together what we're seeing and what were related attacks.

83
00:10:12,040 --> 00:10:29,820
And now, of course, this was actually correlated against the baseline, was done by hand, taking, we, I, and a few others that helped me, took the data over the capture-the-flag game over about a, I think it was about an eight-hour period, and by hand,

84
00:10:29,820 --> 00:10:37,460
figured out what was related information, what were related attacks, and then built a system that could do to automate that process.

85
00:10:37,560 --> 00:10:40,060
So, this thing, it worked.

86
00:10:40,060 --> 00:10:42,540
This is a, it was a pretty decent idea.

87
00:10:42,540 --> 00:10:43,420
It worked.

88
00:10:43,420 --> 00:10:47,020
It didn't necessarily work as well as we had hoped, initially.

89
00:10:48,500 --> 00:10:55,620
If you've got everything on a same network or related network, where you're likely to see lots of related events, it works well.

90
00:10:55,660 --> 00:11:10,040
If you've got this distributed over large networks, where you're likely to see many more unrelated events than related events, it goes, the accuracy goes down a bit, or at least the amount of work you have to do to get good accuracy goes way up, is more accurate.

91
00:11:12,440 --> 00:11:15,840
So, that's what it had done in the past.

92
00:11:16,240 --> 00:11:25,800
And this got me thinking about, so, this was a system that had developed to try and detect related information on a network.

93
00:11:25,840 --> 00:11:29,120
And then, of course, you know, once you can do that, you always think about the next step.

94
00:11:29,120 --> 00:11:34,680
And once the next step, once you can detect something, you figure out how to hide it even better, right?

95
00:11:34,680 --> 00:11:38,180
You know, it's always kind of that one-upmanship game that we're playing.

96
00:11:38,180 --> 00:11:42,780
So, it got me thinking, how do we defeat a system that works like that?

97
00:11:42,780 --> 00:11:59,520
How could you avoid detection on a network that was well-observed, well-monitored, and where the information was well-correlated, where you couldn't rely on the guy sat over on this part of the network not knowing what the guy sat over on this other part of the network was doing,

98
00:11:59,520 --> 00:12:02,060
or seeing, or knew about what was going on.

99
00:12:02,060 --> 00:12:08,060
If you assume that there was complete coverage of the network, how could you hide what was going on on there?

100
00:12:08,060 --> 00:12:23,120
And I'm not necessarily, I'm not talking about so much the detection of a specific attack, like someone does a buffer overflow against a web server, they do, you know, whatever other type of initial attack to gain access to the system.

101
00:12:23,120 --> 00:12:29,360
But assuming they're already on there, they've got on to the system and they want to keep communicating back and forth.

102
00:12:29,360 --> 00:12:33,140
They've got their remote machine they're sat at.

103
00:12:33,140 --> 00:12:35,140
They've got access to a machine on the network.

104
00:12:35,140 --> 00:12:38,040
They want to communicate back and forth, but they don't want to be detected while they're on the network.

105
00:12:39,760 --> 00:12:44,540
How could you go about that and avoid detection on a well-monitored network?

106
00:12:46,460 --> 00:12:51,460
So, you have to look at this from the point of view of where do we have the choices about what we do?

107
00:12:51,700 --> 00:12:55,840
When you're talking about the initial attack, you don't have a whole lot of choices.

108
00:12:56,040 --> 00:13:03,400
If you've got a, you know, I guess, I mean, I guess, unless you get lucky and you see a system out there that's got a ton of vulnerabilities on it, then you've got some choice about what to do.

109
00:13:03,400 --> 00:13:07,880
But assume that you have limited choices of how a system is broken into.

110
00:13:07,880 --> 00:13:10,480
There's limited ways into the system to begin with.

111
00:13:10,480 --> 00:13:12,300
You don't have many choices there.

112
00:13:12,640 --> 00:13:21,680
So, any tool that's written to, to initially gain access to a system usually has a fairly limited life span.

113
00:13:21,680 --> 00:13:27,980
At least it's limited, limited life span in the realm of people who actually care about security.

114
00:13:27,980 --> 00:13:33,220
You always find some bozo out there who never updates anything, never puts any security on their system.

115
00:13:33,220 --> 00:13:34,540
You can always get into them.

116
00:13:34,540 --> 00:13:37,860
But talking about the people who are serious about it, you have a limited life span.

117
00:13:37,880 --> 00:13:40,540
With any sort of tool for initially getting in there.

118
00:13:41,460 --> 00:13:48,560
However, once you have access to a machine and all you're worried about doing is communicating with that machine, you've got a lot of choices.

119
00:13:48,680 --> 00:13:51,400
You know, we're not stuck with one type of communication protocol.

120
00:13:51,400 --> 00:13:56,080
We're not stuck with any traditional communication protocol that's out there whatsoever.

121
00:13:57,540 --> 00:14:03,800
The trick to this is figuring out which ones work and which ones won't be seen.

122
00:14:07,960 --> 00:14:09,740
So, here's, here's the question.

123
00:14:09,740 --> 00:14:13,320
How do you communicate on a network without letting anyone know you're doing it?

124
00:14:13,820 --> 00:14:15,280
First thing's first.

125
00:14:16,620 --> 00:14:18,860
How do they detect what's going on?

126
00:14:18,860 --> 00:14:21,960
How does someone see what you're doing on a network?

127
00:14:21,980 --> 00:14:24,880
You've got IDS, you've got firewalls, you've got observers.

128
00:14:24,920 --> 00:14:29,320
All of these things are sat there looking at what's going on on the network.

129
00:14:30,980 --> 00:14:42,440
In general, I pick the IDS and I say software human because, you know, whatever's looking at your network, whether it's automated or whether it's a person, they're trying to determine if someone's intruding on your network.

130
00:14:42,980 --> 00:14:44,420
You get two basic forms of this.

131
00:14:44,420 --> 00:14:58,500
And this, this applies, um, I think the IDS system in at least one sense is the most general case and you can kind of, uh, find derivative cases that cover this, the same concept for firewalls and routers and everything else out there.

132
00:14:58,500 --> 00:15:00,920
There's basically two forms of detection out there.

133
00:15:01,000 --> 00:15:02,860
Anomaly and signature detection.

134
00:15:03,060 --> 00:15:06,620
Signature detection works on known attacks, known events.

135
00:15:06,620 --> 00:15:13,460
You know, if you, if you know exactly what you're looking for, you know, you see this packet go by, it's something bad, you want to throw up an alert about it.

136
00:15:13,460 --> 00:15:19,000
Hey, that's great, you can catch it every time, it'll never get by, you know, you've got that one nailed.

137
00:15:19,500 --> 00:15:29,020
The only way to avoid that type of detection is to use either an unknown method or something perhaps that changes constantly, never use the same method twice.

138
00:15:30,580 --> 00:15:37,800
The anomaly detection, it doesn't detect misuse of the system, it doesn't detect something malicious, it just detects something unusual.

139
00:15:37,800 --> 00:15:48,580
So, it's, it's not necessarily really great at trying to pick up something bad happening on your network, but it'll pick up something unusual.

140
00:15:48,820 --> 00:15:55,040
And it's really, if the signature detection doesn't work, that's the only chance you have left, is looking for something unusual.

141
00:15:55,040 --> 00:15:58,420
If you don't know what to look for, you've got to look for things that stand out.

142
00:15:59,500 --> 00:16:03,500
So, we can't really avoid the signature detection.

143
00:16:03,500 --> 00:16:10,160
If you've got a way of communicating or a way of doing anything on the network and someone can write a signature to detect that, that's it.

144
00:16:10,160 --> 00:16:12,080
You can't, you can't do anything with it.

145
00:16:12,080 --> 00:16:15,580
It's got a limited lifespan, eventually it's going to be seen.

146
00:16:17,200 --> 00:16:32,600
So, the only alternative is to use something that you can't write a signature for, something that changes constantly, something that is quote unquote random, or at least appears random to anything that's observing the network.

147
00:16:32,720 --> 00:16:36,760
And we have to be able to do this without looking unusual.

148
00:16:36,780 --> 00:16:37,760
And that's the key.

149
00:16:37,760 --> 00:16:49,880
Not only do we have to do something, it's kind of a catch-22 in a way, because we've got to do something that's very unusual, you know, this kind of random changing communication protocol, but we've got to do it without looking unusual.

150
00:16:49,880 --> 00:16:52,580
You can't stand out on there.

151
00:16:52,580 --> 00:16:57,280
So, the first stage of this is to detect the baseline of the network.

152
00:16:57,480 --> 00:17:01,640
What does the traffic on the network look like to begin with?

153
00:17:01,640 --> 00:17:04,420
What does, what is occurring on the network?

154
00:17:04,420 --> 00:17:06,760
What happens out there all the time?

155
00:17:06,780 --> 00:17:08,700
And this can be done passively.

156
00:17:08,700 --> 00:17:24,660
You can sit back and discover what type of traffic is allowed on the network, what does normal traffic look like, and kind of build yourself a map of both what is going on on the network, what the structure of the network is, and what's allowed in and out of this network.

157
00:17:26,140 --> 00:17:28,000
This is not a new concept.

158
00:17:28,180 --> 00:17:30,160
This just, this general idea of doing this.

159
00:17:30,160 --> 00:17:31,240
This is done all the time.

160
00:17:31,240 --> 00:17:36,760
This is how many tools and, and other fun little things are used.

161
00:17:36,780 --> 00:17:38,020
They're written and developed.

162
00:17:38,040 --> 00:17:46,140
But the key thing is, is that this is what, generally, this is what people, people do this process.

163
00:17:46,200 --> 00:17:50,380
Someone looks at a network or you think about what's going on on the network.

164
00:17:50,640 --> 00:17:53,200
You know, a good example of this is a reverse WWW shell.

165
00:17:53,200 --> 00:17:54,740
That's an old tool.

166
00:17:54,740 --> 00:17:59,060
The idea of this was that you've got a computer inside the network.

167
00:17:59,660 --> 00:18:02,780
It's, you're not allowed to make connections from the outside to it.

168
00:18:02,780 --> 00:18:06,720
You know, the firewall's gonna drop any connection, inbound connections to it.

169
00:18:06,780 --> 00:18:09,780
So you put a piece of software on that computer inside the network.

170
00:18:09,860 --> 00:18:15,480
That, at certain intervals, perhaps at random intervals, it looks like it's going out to a web browser.

171
00:18:15,480 --> 00:18:17,100
It goes, makes a web request.

172
00:18:17,100 --> 00:18:25,960
Except that, what that web request is doing is coming to your computer on the outside, getting the next command that you've told, told it to do.

173
00:18:25,960 --> 00:18:33,900
It carries it back as the HTTP response to the, to the target computer, to the one that, the one that you got access to.

174
00:18:33,900 --> 00:18:36,160
And then it carries out that particular command.

175
00:18:36,160 --> 00:18:36,760
It's masquerade.

176
00:18:36,780 --> 00:18:40,400
It's masqueraded looking like the, the computer on the inside is just surfing the web.

177
00:18:40,840 --> 00:18:44,160
That's a great way of hiding it, unless someone knows to look for it.

178
00:18:44,160 --> 00:18:49,540
Once someone knows to look for it, if you stick with it too long, they're gonna look for it, they're gonna see it, it's gonna stand out.

179
00:18:50,380 --> 00:18:56,820
Similarly, carrying data commands and ICMP echo, echo response packets, same sort of, same sort of idea.

180
00:18:57,160 --> 00:19:10,140
What I'm trying to do with this project is to automate this process and make it run on the fly and may, and have it adapt itself to whatever network it's currently working on.

181
00:19:10,140 --> 00:19:17,300
Look at what's going on on the network, decide what to do on the network, decide what will work there, and then carry it out.

182
00:19:19,220 --> 00:19:26,580
So, some aspects of network traffic are gonna have a low entropy, meaning a very low randomness.

183
00:19:26,580 --> 00:19:31,680
They're gonna be very kind of, the same thing's always gonna happen.

184
00:19:31,680 --> 00:19:35,240
Certain communication protocols are only gonna go to certain computers.

185
00:19:35,280 --> 00:19:37,780
DNS requests are only gonna go to your DNS servers.

186
00:19:37,780 --> 00:19:40,280
They're not gonna be jumping around to every computer on your network.

187
00:19:40,860 --> 00:19:43,700
Incoming web requests are probably only gonna come to your web server.

188
00:19:43,700 --> 00:19:45,720
They're not gonna come to every computer on your network.

189
00:19:46,020 --> 00:19:48,980
Other things are gonna look much more random.

190
00:19:48,980 --> 00:19:56,540
Maybe outgoing web requests, you know, people, different people surfing the web, outgoing FTP, all, all sorts of things.

191
00:19:56,540 --> 00:19:58,300
Other things are gonna be very random.

192
00:19:58,300 --> 00:20:00,460
There's gonna be a lot more entropy in that.

193
00:20:01,100 --> 00:20:05,400
And entropy is the key to hiding information.

194
00:20:05,740 --> 00:20:12,720
The, the more entropy you have on the system, the more information you can hide in that.

195
00:20:12,720 --> 00:20:22,080
This is like if you look at, if you look at steganography, for example, you know, you're trying to hide information in something else.

196
00:20:22,080 --> 00:20:28,940
If you've got a bitmap, a bitmap has lots and lots of redundant information in it, versus a JPEG image.

197
00:20:28,940 --> 00:20:33,700
A JPEG has taken a lot of that redundant information, that's why a JPEG's smaller than a bitmap.

198
00:20:33,860 --> 00:20:37,280
You can hide a lot more information than a bitmap than you can in a JPEG.

199
00:20:37,280 --> 00:20:40,220
Higher entropy, you can hide more information inside it.

200
00:20:40,320 --> 00:20:51,040
So, if we find what has the highest entropy on that network, what communication channels have the highest degree of randomness, we can then pick what we want to work in.

201
00:20:51,040 --> 00:20:54,980
And we've got all sorts of, all sorts of things that we can work with here.

202
00:20:54,980 --> 00:21:03,860
We can work with the timing of the packets, we can work with packets, sending out packets that have bad checksums, malicious checksums, malformed checksums, all sorts of different things that we can play with.

203
00:21:03,860 --> 00:21:10,180
We don't have to use a SSH or a Telnet or some other standard type of communication format.

204
00:21:10,180 --> 00:21:18,040
We can send crafted packets where the information is carried in all sorts of the different little pieces of information inside that packet.

205
00:21:18,040 --> 00:21:23,360
The key is, we just have to make sure it doesn't look unusual when that packet goes out.

206
00:21:23,680 --> 00:21:28,780
So the overall concept here is we detect what the network baseline condition is.

207
00:21:28,780 --> 00:21:33,060
On the specific network that we're looking at, what does normal look like?

208
00:21:33,060 --> 00:21:34,680
What doesn't look unusual?

209
00:21:35,040 --> 00:21:38,820
And then select from that the potential communication channels.

210
00:21:38,960 --> 00:21:43,480
Determine which ones have highly random information on the network.

211
00:21:43,640 --> 00:21:47,880
Trim it down, figure, we've got, initially we can select from anything.

212
00:21:47,880 --> 00:21:50,440
We can craft a packet that does anything we want.

213
00:21:50,440 --> 00:21:58,400
We're going to look at what happens on the network, figure out what's allowed, limit it to that, and then limit it to what has a lot of entropy within that.

214
00:21:58,780 --> 00:22:02,380
And then we have to somehow communicate this method to the receiver on the outside.

215
00:22:02,380 --> 00:22:07,020
We have to get the information of what communication scheme we're going to use.

216
00:22:07,020 --> 00:22:09,340
We have to get that to the receiver outside.

217
00:22:09,340 --> 00:22:13,800
And then maintain a, you know, I put almost undetectable presence on the network.

218
00:22:13,920 --> 00:22:15,900
You never should think you're undetectable.

219
00:22:15,900 --> 00:22:18,140
You should never think anything's absolute, because it never is.

220
00:22:18,140 --> 00:22:20,020
That's how you always get yourself into trouble.

221
00:22:20,300 --> 00:22:24,520
But we can monitor and update the network conditions to stay hidden.

222
00:22:24,520 --> 00:22:28,400
If network conditions change, we can change the scheme we're using to communicate.

223
00:22:30,240 --> 00:22:35,320
On the network, if we just sit there and listen, what can we learn by listening on the network?

224
00:22:35,320 --> 00:22:43,780
There's all sorts of things that we can find out by just sitting there and listening, not doing anything active, not doing anything to be discovered.

225
00:22:43,780 --> 00:22:48,820
You can figure out what types of packets are allowed inside the network, what types of packets are allowed outside the network.

226
00:22:48,820 --> 00:22:51,060
We can figure out the architecture of the network.

227
00:22:51,060 --> 00:22:51,940
Where are the switches?

228
00:22:51,940 --> 00:22:53,060
Where are the routers?

229
00:22:53,060 --> 00:22:55,420
Where are other machines on this network?

230
00:22:55,540 --> 00:22:58,640
And then look at what types of packets are flying around.

231
00:22:58,640 --> 00:23:01,920
What types of packets does the machine that I'm on send and receive?

232
00:23:01,920 --> 00:23:05,000
What types of packets do other machines out there send and receive?

233
00:23:05,000 --> 00:23:21,640
And from this, get an idea of what is allowed to move around this network, what is moving around this network, and what we can inject into the data flowing around without being, without sticking out like a sore thumb and being immediately identified.

234
00:23:22,800 --> 00:23:28,360
So if we narrow down these choices, we start with the ability to send any type of packet, you know, send it to any port whatsoever.

235
00:23:28,640 --> 00:23:45,180
We start paring it down based on what's available, figure out what has the highest entropy, look at this, look at the type of packets based on the protocol, the ports they're using, et cetera, and look at the entropy based on how often do these packets fly,

236
00:23:45,180 --> 00:23:55,660
how different is the data in these packets whenever they go by, what different machines do these types of packets go to, and then pick a channel that we know we can get from point A to point B.

237
00:23:55,660 --> 00:24:05,500
We can get from inside the network, we've got access to the machine network, out of that network over to the place where we want to get the information to, or back and forth, vice versa, either way.

238
00:24:05,500 --> 00:24:11,080
We can figure out what can get into one, out of one network, into another, and then back the other way.

239
00:24:11,080 --> 00:24:13,920
It doesn't necessarily have to be a symmetric communication channel.

240
00:24:13,980 --> 00:24:21,140
The packets that you send out, the packets that you send going out one way, don't have to look a heck of a lot like the packets coming back in the other way.

241
00:24:21,160 --> 00:24:37,700
As long as they're getting to the right segment on the network, as long as they're getting down to the right branch off of your switch so that it can be picked up by a, by a network card in promiscuous mode, as long as you can get down that far and you know what to look for,

242
00:24:37,700 --> 00:24:42,980
each guy on each end knows what to look for, you don't have to have it even be a symmetric communication channel between them.

243
00:24:42,980 --> 00:24:47,040
You can have essentially two different channels going back and forth.

244
00:24:49,040 --> 00:24:55,380
One of the big challenges with this system is starting the conversation.

245
00:24:55,380 --> 00:25:09,220
If you've got a piece of software running on the inside of this network and it's gonna start communicating with you on the outside and you don't know what particular communication scheme it's gonna be using, there's some difficulties in getting, in getting that first packet out of there,

246
00:25:09,220 --> 00:25:15,780
that first piece of information to the outside without having it be something that can easily be written as a signature.

247
00:25:15,780 --> 00:25:23,760
Because it's hard to, it's hard to change things or inject randomness into the first bit of communication.

248
00:25:23,760 --> 00:25:39,700
There's some, there's some ways that we can limit how signature vulnerable that first bit of communication is, but that's one of, that is, that's one of the points that needs, that presents some challenge and needs some work on in this.

249
00:25:40,420 --> 00:25:52,020
Having the communication software itself remain undetected on the remote machine, not talking about what the software is doing, but actually having the software itself avoid detection, of course that's another challenge, but that's a completely different story.

250
00:25:52,900 --> 00:26:03,400
That's, whether it's something that's using a polymorphic code, that is, that changes from one installation to another, or even changes itself constantly.

251
00:26:03,400 --> 00:26:07,960
There are ways around that, but it is a different, but it is another challenge in there.

252
00:26:08,300 --> 00:26:13,780
The third one, that's actually I think a fairly interesting one, is the problem on a low entropy network.

253
00:26:13,820 --> 00:26:17,540
What do you do when there's not a whole lot of randomness going on on the network?

254
00:26:18,000 --> 00:26:22,000
You know, if there's not a lot of randomness in the packets, how do we stay hidden in it?

255
00:26:22,020 --> 00:26:38,500
We have to ensure that every packet that we send out doesn't significantly affect the statistical baseline of the network, so that an anomaly detection system sat on the network will not recognize this as something unusual.

256
00:26:38,500 --> 00:26:39,560
It will not stand out.

257
00:26:39,560 --> 00:26:48,080
It won't push the statistical analysis of the network beyond the limits, beyond the thresholds that we currently have, and it will avoid becoming an anomaly.

258
00:26:48,080 --> 00:27:05,420
This is actually the area right now that I'm doing, concentrating on, trying to figure out the best methods for communicating and picking channels, and picking well-hidden channels on a low entropy network.

259
00:27:05,540 --> 00:27:25,880
So, just as a summary in the system, you know, what's real, what's not real, so far the allowed protocols discovery, the network architecture discovery, the statistical profiling of the network, those are, those are working prototypes.

260
00:27:25,880 --> 00:27:28,920
Those modules, those parts of the system are working prototypes.

261
00:27:28,920 --> 00:27:39,740
Next stages that we're moving towards are the selection of the communication scheme, implementation of the communication, and the adaptation to changing network conditions.

262
00:27:39,920 --> 00:27:50,180
I hope that by this time next year this will actually be a working package that I can give maybe a bit better presentation and demonstration of what it's actually doing.

263
00:27:50,180 --> 00:27:51,200
But this is the current idea.

264
00:27:51,980 --> 00:27:59,540
Anybody that has any interest in this, or any ideas, or feedback, or wants to tell me why I'm crazy for thinking this, feel free.

265
00:27:59,540 --> 00:28:01,260
There's my contact information.

266
00:28:02,300 --> 00:28:05,780
And that's it for a, about a 25-minute presentation.

267
00:28:13,260 --> 00:28:14,520
There's one question.

268
00:28:23,990 --> 00:28:26,430
That, that's a, that is a great question.

269
00:28:26,870 --> 00:28:29,730
I'll tell you, I'll take that just, I'll try and answer one more real quick.

270
00:28:29,730 --> 00:28:39,850
He asked, is there any concept of trying to actually change the baseline of the network so that you can do something that used to look unusual, make it not look unusual anymore?

271
00:28:39,850 --> 00:28:43,570
That is, it's not one of the first-level items on there.

272
00:28:43,570 --> 00:28:45,870
That's a pretty advanced way of working it.

273
00:28:45,870 --> 00:28:56,730
But actually changing, injecting stuff in slow enough that the system, the anomaly detection system adapts to it and thinks that's normal, that is one of the long-term ideas on there.

274
00:28:56,730 --> 00:29:00,210
That's a great idea to put into this.

275
00:29:00,210 --> 00:29:02,490
So, that's

276
00:29:36,500 --> 00:29:38,160
fair enough.

277
00:29:38,400 --> 00:29:40,160
Thank you.

278
00:29:42,370 --> 00:29:43,330
Sorry.