1 1 00:00:00,360 --> 00:00:02,220 Dynamic analysis. 2 2 00:00:02,220 --> 00:00:03,180 So up to this point, 3 3 00:00:03,180 --> 00:00:04,950 we've talked about static analysis 4 4 00:00:04,950 --> 00:00:07,290 and going through the code line by line 5 5 00:00:07,290 --> 00:00:08,940 using our decompiler. 6 6 00:00:08,940 --> 00:00:10,530 In this lesson, we are going to talk 7 7 00:00:10,530 --> 00:00:13,560 about why you might want to do dynamic analysis instead. 8 8 00:00:13,560 --> 00:00:15,210 Now when you do static analysis, 9 9 00:00:15,210 --> 00:00:17,490 we're going to be looking through the disassembled code, 10 10 00:00:17,490 --> 00:00:19,170 but that's far from perfect 11 11 00:00:19,170 --> 00:00:22,020 because a lot of times, it's been changed and modified 12 12 00:00:22,020 --> 00:00:23,970 so we can't see what's going on. 13 13 00:00:23,970 --> 00:00:26,040 So when a malware writer takes their code 14 14 00:00:26,040 --> 00:00:27,600 and tries to hide their intent, 15 15 00:00:27,600 --> 00:00:29,910 stack analysis can be really difficult. 16 16 00:00:29,910 --> 00:00:33,930 So what we might want to do instead is do a dynamic analysis. 17 17 00:00:33,930 --> 00:00:35,760 Now, with a dynamic analysis, 18 18 00:00:35,760 --> 00:00:38,700 this involves the execution of a compiled program 19 19 00:00:38,700 --> 00:00:40,710 and then we're going to analyze the way it executes 20 20 00:00:40,710 --> 00:00:43,920 and interacts with the host, the system, or the network. 21 21 00:00:43,920 --> 00:00:47,160 By doing that, we can figure out how this thing works. 22 22 00:00:47,160 --> 00:00:49,260 So if you think back to our different types of testing 23 23 00:00:49,260 --> 00:00:52,020 with known environment test and unknown environment test, 24 24 00:00:52,020 --> 00:00:53,490 dynamic analysis is more like 25 25 00:00:53,490 --> 00:00:55,350 unknown environment test, right? 26 26 00:00:55,350 --> 00:00:57,570 We're going to run it and see what comes out the other side, 27 27 00:00:57,570 --> 00:00:59,040 what changes it makes. 28 28 00:00:59,040 --> 00:01:01,080 With static analysis, we're trying to essentially 29 29 00:01:01,080 --> 00:01:02,550 do a known environment test, right? 30 30 00:01:02,550 --> 00:01:04,650 We're trying to get to something that looks like code 31 31 00:01:04,650 --> 00:01:07,740 so we can analyze all the moving pieces inside of it. 32 32 00:01:07,740 --> 00:01:09,750 Now, when you're doing dynamic analysis, 33 33 00:01:09,750 --> 00:01:11,970 there are a couple of different tools you're going to use. 34 34 00:01:11,970 --> 00:01:15,090 You might use a debugger, a stress test application, 35 35 00:01:15,090 --> 00:01:17,400 a fuzzing application, and we're going to talk about 36 36 00:01:17,400 --> 00:01:19,410 each of these three in this lesson. 37 37 00:01:19,410 --> 00:01:20,910 Now, when we talk about a debugger, 38 38 00:01:20,910 --> 00:01:22,380 this is a dynamic testing tool 39 39 00:01:22,380 --> 00:01:25,740 that's used to analyze software as it's being executed. 40 40 00:01:25,740 --> 00:01:28,080 Now essentially what the debugger allows you to do 41 41 00:01:28,080 --> 00:01:31,620 is step through the program instruction by instruction. 42 42 00:01:31,620 --> 00:01:33,450 If you looked in IDA, we had that pseudo code. 43 43 00:01:33,450 --> 00:01:35,520 There was 50 lines of code there. 44 44 00:01:35,520 --> 00:01:36,353 Now in a real program 45 45 00:01:36,353 --> 00:01:39,750 there might be 50, 500, or even 5,000 lines. 46 46 00:01:39,750 --> 00:01:42,390 With the debugger, I can go and run one line at a time 47 47 00:01:42,390 --> 00:01:45,000 through that code or one instruction at a time 48 48 00:01:45,000 --> 00:01:47,970 based on the assembly code that that binary has. 49 49 00:01:47,970 --> 00:01:50,850 Now, I can stop it at any point and then start looking at it 50 50 00:01:50,850 --> 00:01:52,920 and that's what a debugger allows us to do. 51 51 00:01:52,920 --> 00:01:55,560 The debugger is going to allow us to pause the execution 52 52 00:01:55,560 --> 00:01:57,810 and monitor or adjust the value of variables 53 53 00:01:57,810 --> 00:02:00,180 at different stages through its execution. 54 54 00:02:00,180 --> 00:02:02,250 Again, this allows us to control a little bit 55 55 00:02:02,250 --> 00:02:04,050 what's inside that black box 56 56 00:02:04,050 --> 00:02:05,700 so we can see what comes out the other side 57 57 00:02:05,700 --> 00:02:06,990 and then we might be able to figure out 58 58 00:02:06,990 --> 00:02:08,580 how it actually works. 59 59 00:02:08,580 --> 00:02:11,250 Now when you use a debugger, it looks something like this. 60 60 00:02:11,250 --> 00:02:13,200 In this example, you can see a breakpoint 61 61 00:02:13,200 --> 00:02:16,380 that little stop sign was set at line 209. 62 62 00:02:16,380 --> 00:02:18,030 This means the program will start running 63 63 00:02:18,030 --> 00:02:20,910 and when it gets to line 209, it's going to stop. 64 64 00:02:20,910 --> 00:02:22,770 Now once it stops there and pauses, 65 65 00:02:22,770 --> 00:02:24,990 at this point I can highlight that line 66 66 00:02:24,990 --> 00:02:27,150 and I can look at the different values of the command 67 67 00:02:27,150 --> 00:02:29,040 or the value of a particular variable. 68 68 00:02:29,040 --> 00:02:31,500 In this case, it's letting me see what the value 69 69 00:02:31,500 --> 00:02:36,500 of wcoef[1] is going to be in this case minus 0.72. 70 70 00:02:38,220 --> 00:02:39,810 Now, does that mean anything to us right now? 71 71 00:02:39,810 --> 00:02:41,370 Not really, but if we were going through 72 72 00:02:41,370 --> 00:02:43,500 and analyzing this program, it might be helpful 73 73 00:02:43,500 --> 00:02:45,420 to know what all those things are. 74 74 00:02:45,420 --> 00:02:47,880 Now debuggers are legitimately used by programmers 75 75 00:02:47,880 --> 00:02:49,140 to find errors in their code 76 76 00:02:49,140 --> 00:02:50,970 and then identify how to fix them, 77 77 00:02:50,970 --> 00:02:53,790 but as a malware analyst, we can use them too. 78 78 00:02:53,790 --> 00:02:55,890 Now, we're not the only ones who use 'em, though. 79 79 00:02:55,890 --> 00:02:57,450 Some attackers and reverse engineers 80 80 00:02:57,450 --> 00:03:00,030 can use debuggers as well to step through a binary 81 81 00:03:00,030 --> 00:03:02,250 and determine how to best exploit any weakness 82 82 00:03:02,250 --> 00:03:03,900 or vulnerability in it. 83 83 00:03:03,900 --> 00:03:07,140 This is used by both sides, defense and attack 84 84 00:03:07,140 --> 00:03:08,430 and so it's really something that's useful 85 85 00:03:08,430 --> 00:03:10,740 for you to understand if you happen to be a programmer 86 86 00:03:10,740 --> 00:03:13,530 and you're working as a reverse malware engineer. 87 87 00:03:13,530 --> 00:03:14,730 Now the next time we're going to talk about 88 88 00:03:14,730 --> 00:03:16,260 is known as a stress test. 89 89 00:03:16,260 --> 00:03:18,750 Now a stress test is a software testing method 90 90 00:03:18,750 --> 00:03:21,990 that evaluates how software performs under extreme load. 91 91 00:03:21,990 --> 00:03:24,930 Now that extreme load can come as lots of different things. 92 92 00:03:24,930 --> 00:03:26,610 It might be extreme processor load, 93 93 00:03:26,610 --> 00:03:29,340 extreme memory load, extreme network load. 94 94 00:03:29,340 --> 00:03:30,173 That's up to you 95 95 00:03:30,173 --> 00:03:32,670 and what you're trying to stress inside your test. 96 96 00:03:32,670 --> 00:03:34,650 Now a stress test can be used to determine 97 97 00:03:34,650 --> 00:03:36,930 what could trigger a denial of service. 98 98 00:03:36,930 --> 00:03:39,480 So a lot of times we'll stress test our own applications 99 99 00:03:39,480 --> 00:03:41,490 and our own networks because for instance 100 100 00:03:41,490 --> 00:03:43,020 when I'm setting up my new website 101 101 00:03:43,020 --> 00:03:44,970 and I'm going to have all my students come onto it, 102 102 00:03:44,970 --> 00:03:46,650 I want to know how many students can I take 103 103 00:03:46,650 --> 00:03:48,390 before the site crashes. 104 104 00:03:48,390 --> 00:03:50,190 Well, I can run a stress test against it 105 105 00:03:50,190 --> 00:03:51,030 and when we've done this, 106 106 00:03:51,030 --> 00:03:52,860 we figured out that we can support 107 107 00:03:52,860 --> 00:03:54,360 10,000 students simultaneously 108 108 00:03:54,360 --> 00:03:56,160 or a hundred thousand students simultaneously 109 109 00:03:56,160 --> 00:03:58,110 or 10 students simultaneously. 110 110 00:03:58,110 --> 00:03:59,250 We have to know what that number is 111 111 00:03:59,250 --> 00:04:01,740 so we can actually sell the right number of students 112 112 00:04:01,740 --> 00:04:03,660 without giving you bad quality of service 113 113 00:04:03,660 --> 00:04:06,390 and that's what a stress test allows you to do. 114 114 00:04:06,390 --> 00:04:09,150 Now as a cybersecurity analyst, you might do a stress test 115 115 00:04:09,150 --> 00:04:12,240 of your systems to see what would make them fall down, 116 116 00:04:12,240 --> 00:04:14,430 what would make them have a denial of service 117 117 00:04:14,430 --> 00:04:15,510 and so you can use a program 118 118 00:04:15,510 --> 00:04:17,400 like this web application stress. 119 119 00:04:17,400 --> 00:04:19,260 You can run this program against your server 120 120 00:04:19,260 --> 00:04:21,270 and it will run a bunch of different scripts against it 121 121 00:04:21,270 --> 00:04:22,890 and that way it'll be able to figure out 122 122 00:04:22,890 --> 00:04:25,560 what would cause that denial of service for you. 123 123 00:04:25,560 --> 00:04:27,840 Again, this might be resource exhaustion. 124 124 00:04:27,840 --> 00:04:29,340 It could be something like you're using up 125 125 00:04:29,340 --> 00:04:32,400 all the processor, all the memory, all the disk space, 126 126 00:04:32,400 --> 00:04:35,100 all the network bandwidth, whatever the limitation is, 127 127 00:04:35,100 --> 00:04:37,020 you can find it by doing a stress test 128 128 00:04:37,020 --> 00:04:39,450 and so often when you're deploying a new network 129 129 00:04:39,450 --> 00:04:40,800 or a new web application, 130 130 00:04:40,800 --> 00:04:42,750 you would be asked as a cybersecurity analyst 131 131 00:04:42,750 --> 00:04:43,920 to test that system 132 132 00:04:43,920 --> 00:04:45,900 by putting it through some stress testing. 133 133 00:04:45,900 --> 00:04:48,090 The next area we're going to talk about is fuzzing. 134 134 00:04:48,090 --> 00:04:50,850 Now, fuzzing is a dynamic code analysis technique 135 135 00:04:50,850 --> 00:04:52,830 that involves sending a running application, 136 136 00:04:52,830 --> 00:04:54,810 random and unusual input 137 137 00:04:54,810 --> 00:04:57,390 to evaluate how the application responds. 138 138 00:04:57,390 --> 00:04:58,860 Now let me give you a simple example 139 139 00:04:58,860 --> 00:05:02,250 of how fuzzing is used to see if it can crash something. 140 140 00:05:02,250 --> 00:05:04,890 For example, on my website we have a checkout page 141 141 00:05:04,890 --> 00:05:06,630 and you're going to enter your credit card. 142 142 00:05:06,630 --> 00:05:08,730 Now we expect what to be entered there? 143 143 00:05:08,730 --> 00:05:09,600 Your credit card, right? 144 144 00:05:09,600 --> 00:05:11,310 It should be 16 digits. 145 145 00:05:11,310 --> 00:05:12,720 What if you start adding in things 146 146 00:05:12,720 --> 00:05:14,850 like special characters or letters? 147 147 00:05:14,850 --> 00:05:16,290 Will that crash our system? 148 148 00:05:16,290 --> 00:05:18,060 Well, if we do a fuzz test, 149 149 00:05:18,060 --> 00:05:20,040 we're going to send some of those random things in there 150 150 00:05:20,040 --> 00:05:21,480 to see if we can crash it 151 151 00:05:21,480 --> 00:05:22,410 and then we can find out 152 152 00:05:22,410 --> 00:05:24,300 have we done proper input validation 153 153 00:05:24,300 --> 00:05:26,130 to prevent this from crashing? 154 154 00:05:26,130 --> 00:05:28,680 That's what fuzzing does, it test your input validation 155 155 00:05:28,680 --> 00:05:30,510 and your other systems to make sure 156 156 00:05:30,510 --> 00:05:32,370 you have the right protections in place. 157 157 00:05:32,370 --> 00:05:33,630 Now, fuzzing is a technique 158 158 00:05:33,630 --> 00:05:36,270 that's designed to test your software for various bugs 159 159 00:05:36,270 --> 00:05:37,680 and vulnerabilities. 160 160 00:05:37,680 --> 00:05:39,570 Like I said, I can test your application 161 161 00:05:39,570 --> 00:05:40,890 and see if it's going to crash 162 162 00:05:40,890 --> 00:05:43,470 when you send it some kind of wacky input. 163 163 00:05:43,470 --> 00:05:45,180 That's what fuzzing is all about. 164 164 00:05:45,180 --> 00:05:46,590 Now, when you use a fuzzer, 165 165 00:05:46,590 --> 00:05:48,330 there are three different ways to inject 166 166 00:05:48,330 --> 00:05:51,000 that manipulate input into your application. 167 167 00:05:51,000 --> 00:05:53,040 You can use the application user interface 168 168 00:05:53,040 --> 00:05:54,510 which identifies input streams 169 169 00:05:54,510 --> 00:05:56,130 that are accepted by the application, 170 170 00:05:56,130 --> 00:05:59,190 such as an input box or a form or a command line switch 171 171 00:05:59,190 --> 00:06:01,110 and then send information that way. 172 172 00:06:01,110 --> 00:06:03,360 Another way you can do it is by using the protocol 173 173 00:06:03,360 --> 00:06:05,430 and you can transmit the manipulated packets 174 174 00:06:05,430 --> 00:06:06,840 into the application. 175 175 00:06:06,840 --> 00:06:08,820 For instance, maybe you're used to accepting things 176 176 00:06:08,820 --> 00:06:11,610 over the web so I can go ahead and modify the headers 177 177 00:06:11,610 --> 00:06:14,250 or the payloads and see if that'll crash you. 178 178 00:06:14,250 --> 00:06:16,830 And then the third way is to use a different file format. 179 179 00:06:16,830 --> 00:06:18,120 Now when you're doing a file format, 180 180 00:06:18,120 --> 00:06:19,860 we're going to try to open some kind of file 181 181 00:06:19,860 --> 00:06:21,720 but we're going to manipulate that file. 182 182 00:06:21,720 --> 00:06:24,360 So you might think that you're uploading a PNG file 183 183 00:06:24,360 --> 00:06:27,030 and when I actually send it to you, it's labeled as a PNG 184 184 00:06:27,030 --> 00:06:28,290 but it might be an executable 185 185 00:06:28,290 --> 00:06:30,390 and we'll see if it causes issues with your system. 186 186 00:06:30,390 --> 00:06:33,960 Now, fuzzers may craft the input using semi-random input 187 187 00:06:33,960 --> 00:06:35,640 or specific inputs. 188 188 00:06:35,640 --> 00:06:37,170 If we use semi random input, 189 189 00:06:37,170 --> 00:06:38,700 we call this a dumb fuzzer. 190 190 00:06:38,700 --> 00:06:40,830 If we use specific inputs on the other hand, 191 191 00:06:40,830 --> 00:06:43,050 we're using things based on what we know works 192 192 00:06:43,050 --> 00:06:44,730 based on other exploit vectors 193 193 00:06:44,730 --> 00:06:45,720 or certain vulnerabilities 194 194 00:06:45,720 --> 00:06:48,165 that we're trying to target specifically. 195 195 00:06:48,165 --> 00:06:50,340 We're not just using anything that's random in that case. 196 196 00:06:50,340 --> 00:06:52,470 Now, one of the examples of a fuzzer 197 197 00:06:52,470 --> 00:06:54,300 is what's known as Peach Fuzzer. 198 198 00:06:54,300 --> 00:06:56,130 The Peach Fuzzer platform is one 199 199 00:06:56,130 --> 00:06:58,500 of the most advanced commercial fuzzers out there. 200 200 00:06:58,500 --> 00:07:00,570 It is used by a lot of people in the industry. 201 201 00:07:00,570 --> 00:07:01,920 Now if you want to play with it, 202 202 00:07:01,920 --> 00:07:03,780 there's also a free community version, 203 203 00:07:03,780 --> 00:07:05,400 like a lot of the tools we've talked about. 204 204 00:07:05,400 --> 00:07:07,350 There's the commercial side that makes the money 205 205 00:07:07,350 --> 00:07:08,820 and then they have these free tools 206 206 00:07:08,820 --> 00:07:10,170 for us to use as we're learning 207 207 00:07:10,170 --> 00:07:12,720 and protecting our own home-based networks. 208 208 00:07:12,720 --> 00:07:14,430 Now as you go and use something like this, 209 209 00:07:14,430 --> 00:07:16,710 you can use it against your own systems. 210 210 00:07:16,710 --> 00:07:17,543 Don't take this 211 211 00:07:17,543 --> 00:07:19,860 and target other people's websites without their permission. 212 212 00:07:19,860 --> 00:07:20,700 That's hacking. 213 213 00:07:20,700 --> 00:07:22,320 That can land you in jail. 214 214 00:07:22,320 --> 00:07:24,690 Instead, you can target your own websites 215 215 00:07:24,690 --> 00:07:26,640 or sites you've placed in your lab. 216 216 00:07:26,640 --> 00:07:29,430 Now this platform has a lot of built-in test cases. 217 217 00:07:29,430 --> 00:07:32,070 Now test cases are these different preconfigured things 218 218 00:07:32,070 --> 00:07:33,090 that we're going to work against 219 219 00:07:33,090 --> 00:07:34,950 different types of products and systems. 220 220 00:07:34,950 --> 00:07:36,420 So we have some test cases that will go 221 221 00:07:36,420 --> 00:07:37,710 against network appliances. 222 222 00:07:37,710 --> 00:07:40,110 We have test cases that go against web applications. 223 223 00:07:40,110 --> 00:07:42,480 We have test cases that go against financial data, 224 224 00:07:42,480 --> 00:07:44,400 test cases that go against healthcare 225 225 00:07:44,400 --> 00:07:46,920 and we can use these to test these different applications 226 226 00:07:46,920 --> 00:07:48,300 and they have different data sets 227 227 00:07:48,300 --> 00:07:49,890 that we would expect to use 228 228 00:07:49,890 --> 00:07:52,790 and different things that we would expect to cause issues.