1 00:00:00,090 --> 00:00:00,960 In this lesson, 2 00:00:00,960 --> 00:00:03,420 we're going to cover cryptographic attacks. 3 00:00:03,420 --> 00:00:04,740 Now, cryptographic attacks 4 00:00:04,740 --> 00:00:07,470 refer to techniques and strategies that adversaries employ 5 00:00:07,470 --> 00:00:10,230 to exploit vulnerabilities in cryptographic systems 6 00:00:10,230 --> 00:00:12,450 with the intent to compromise the confidentiality, 7 00:00:12,450 --> 00:00:15,240 integrity, and authenticity of your data. 8 00:00:15,240 --> 00:00:16,410 These attacks are there 9 00:00:16,410 --> 00:00:18,330 to defeat the cryptographic protections, 10 00:00:18,330 --> 00:00:19,950 either by deciphering encrypted data 11 00:00:19,950 --> 00:00:22,740 without the appropriate key, impersonating another user, 12 00:00:22,740 --> 00:00:25,230 or creating forgeries that cryptographic systems 13 00:00:25,230 --> 00:00:26,850 would deem authentic. 14 00:00:26,850 --> 00:00:28,920 These methods can range from exploiting weaknesses 15 00:00:28,920 --> 00:00:30,900 in cryptographic algorithms and protocols 16 00:00:30,900 --> 00:00:33,270 to leveraging flaws in software implementations 17 00:00:33,270 --> 00:00:36,270 or simply using brute-force techniques to guess the key. 18 00:00:36,270 --> 00:00:38,250 For our purposes in this lesson, though, 19 00:00:38,250 --> 00:00:40,950 we're going to focus on attacks related to downgrade attacks, 20 00:00:40,950 --> 00:00:43,530 collision attacks, and the threat of quantum computing 21 00:00:43,530 --> 00:00:44,363 that is looming over 22 00:00:44,363 --> 00:00:46,770 all of our existing encryption algorithms. 23 00:00:46,770 --> 00:00:49,140 Now, first we have downgrade attacks. 24 00:00:49,140 --> 00:00:51,510 At its core, a cryptographic downgrade attack, 25 00:00:51,510 --> 00:00:53,580 also known as a version rollback attack, 26 00:00:53,580 --> 00:00:55,500 will aim to force a system to use a weaker 27 00:00:55,500 --> 00:00:57,750 or older cryptographic standard or protocol 28 00:00:57,750 --> 00:00:59,520 than what it's currently using. 29 00:00:59,520 --> 00:01:01,230 The idea here is that by rolling back 30 00:01:01,230 --> 00:01:02,670 to these outdated versions, 31 00:01:02,670 --> 00:01:05,220 an attacker can exploit known vulnerabilities or weaknesses 32 00:01:05,220 --> 00:01:07,770 in those older protocols that have now been addressed 33 00:01:07,770 --> 00:01:09,960 in newer or more secure versions. 34 00:01:09,960 --> 00:01:11,370 For example, let's pretend 35 00:01:11,370 --> 00:01:13,290 that we have two people, Alice and Bob, 36 00:01:13,290 --> 00:01:15,150 and they want to establish a secure connection channel 37 00:01:15,150 --> 00:01:16,590 between their systems. 38 00:01:16,590 --> 00:01:18,510 Now, most modern cryptographic systems 39 00:01:18,510 --> 00:01:20,790 will support multiple versions of a protocol 40 00:01:20,790 --> 00:01:22,440 to ensure for backwards compatibility 41 00:01:22,440 --> 00:01:24,240 with older legacy systems. 42 00:01:24,240 --> 00:01:26,280 When Alice and Bob begin their handshake 43 00:01:26,280 --> 00:01:28,050 to establish that secure connection, 44 00:01:28,050 --> 00:01:29,910 their systems will typically communicate 45 00:01:29,910 --> 00:01:30,900 the version of the protocols 46 00:01:30,900 --> 00:01:32,490 that each one of them can support, 47 00:01:32,490 --> 00:01:33,870 and then their systems will agree 48 00:01:33,870 --> 00:01:35,340 to use the most secure version 49 00:01:35,340 --> 00:01:37,650 that both of these systems can understand. 50 00:01:37,650 --> 00:01:40,200 For example, if I speak English and Spanish 51 00:01:40,200 --> 00:01:42,090 and you speak Spanish and Chinese, 52 00:01:42,090 --> 00:01:43,710 we would agree to speak Spanish 53 00:01:43,710 --> 00:01:46,230 because both of us understand that language. 54 00:01:46,230 --> 00:01:47,940 So going back to our example, 55 00:01:47,940 --> 00:01:50,310 let's pretend that we have an attacker named Eve, 56 00:01:50,310 --> 00:01:52,500 and she wants to be able to intercept this communication 57 00:01:52,500 --> 00:01:54,060 between Alice and Bob. 58 00:01:54,060 --> 00:01:55,170 To execute this attack, 59 00:01:55,170 --> 00:01:57,210 Eve will try to manipulate the messages being sent 60 00:01:57,210 --> 00:02:00,090 between Alice and Bob's computer to establish this handshake 61 00:02:00,090 --> 00:02:02,010 and decide on a version to use. 62 00:02:02,010 --> 00:02:03,960 This way, they can manipulate the message 63 00:02:03,960 --> 00:02:06,120 to say the highest version that both parties support 64 00:02:06,120 --> 00:02:08,910 is in fact an older, less secure version. 65 00:02:08,910 --> 00:02:12,390 Now, Alice and Bobs' systems, unaware of Eve's interference, 66 00:02:12,390 --> 00:02:14,400 will proceed with the outdated protocol, 67 00:02:14,400 --> 00:02:16,830 and this makes all their communications susceptible 68 00:02:16,830 --> 00:02:19,200 to vulnerabilities inherent in that older version 69 00:02:19,200 --> 00:02:20,550 of the encryption algorithm 70 00:02:20,550 --> 00:02:22,410 that Eve can now take advantage of 71 00:02:22,410 --> 00:02:25,230 to be able to see everything they're sending back and forth. 72 00:02:25,230 --> 00:02:26,760 One infamous real-world example 73 00:02:26,760 --> 00:02:28,200 of this type of downgrade attack 74 00:02:28,200 --> 00:02:30,450 involves the use of the POODLE attack, 75 00:02:30,450 --> 00:02:31,770 known as the Padding Oracle 76 00:02:31,770 --> 00:02:33,990 On Downgraded Legacy Encryption attack, 77 00:02:33,990 --> 00:02:36,660 which targeted SSL version 3.0. 78 00:02:36,660 --> 00:02:38,040 Now, despite the availability 79 00:02:38,040 --> 00:02:41,580 and widespread use of a more secure protocol like TLS, 80 00:02:41,580 --> 00:02:44,010 many systems still supported SSL 3.0 81 00:02:44,010 --> 00:02:45,930 for backward compatibility. 82 00:02:45,930 --> 00:02:48,480 POODLE tried to exploit this by forcing its victims 83 00:02:48,480 --> 00:02:51,240 to revert to this older, more insecure protocol, 84 00:02:51,240 --> 00:02:52,290 and this would allow attackers 85 00:02:52,290 --> 00:02:55,140 to extract sensitive data from these communications. 86 00:02:55,140 --> 00:02:58,290 So, why are downgrade attacks so dangerous? 87 00:02:58,290 --> 00:03:00,030 Well, these downgrade attacks are dangerous 88 00:03:00,030 --> 00:03:02,850 because they turn the very nature of evolving security, 89 00:03:02,850 --> 00:03:04,440 such as the development of a stronger, 90 00:03:04,440 --> 00:03:07,740 more robust cryptographic protocol, against itself. 91 00:03:07,740 --> 00:03:10,290 In trying to be more inclusive and comprehensive, 92 00:03:10,290 --> 00:03:12,570 our systems that maintain this backwards compatibility 93 00:03:12,570 --> 00:03:15,540 with older standards are unknowingly leaving themselves open 94 00:03:15,540 --> 00:03:17,580 to these types of downgrade attacks. 95 00:03:17,580 --> 00:03:19,530 But it's not all doom and gloom. 96 00:03:19,530 --> 00:03:20,850 Awareness of downgrade attacks 97 00:03:20,850 --> 00:03:22,380 has led to the development of countermeasures 98 00:03:22,380 --> 00:03:23,550 to help prevent them. 99 00:03:23,550 --> 00:03:26,070 For instance, many systems have phased out support 100 00:03:26,070 --> 00:03:29,310 for legacy protocols that are known to be truly insecure, 101 00:03:29,310 --> 00:03:32,850 even if it means sacrificing backward compatibility. 102 00:03:32,850 --> 00:03:35,340 In other scenarios, version intolerance checks 103 00:03:35,340 --> 00:03:38,250 can also be employed where a system tests the waters 104 00:03:38,250 --> 00:03:39,930 by initially claiming to only support 105 00:03:39,930 --> 00:03:41,940 the most recent protocol version. 106 00:03:41,940 --> 00:03:42,960 If the other end user 107 00:03:42,960 --> 00:03:44,940 truly doesn't support the latest version, 108 00:03:44,940 --> 00:03:46,950 they will naturally respond accordingly, 109 00:03:46,950 --> 00:03:49,530 but if there's any interference forcing a downgrade, 110 00:03:49,530 --> 00:03:51,630 this will quickly become apparent. 111 00:03:51,630 --> 00:03:53,970 Second, we have collision attacks. 112 00:03:53,970 --> 00:03:56,730 Now, a collision attack aims to find two different inputs 113 00:03:56,730 --> 00:03:58,890 that produce the same hash output. 114 00:03:58,890 --> 00:04:00,000 This can be a vulnerability 115 00:04:00,000 --> 00:04:03,060 because hashing is commonly used for data verification. 116 00:04:03,060 --> 00:04:05,070 For example, when you're downloading a file 117 00:04:05,070 --> 00:04:07,620 from the internet, the website may provide a hash value 118 00:04:07,620 --> 00:04:08,880 for that file. 119 00:04:08,880 --> 00:04:10,560 Once you finish downloading the file, 120 00:04:10,560 --> 00:04:12,270 you can run it through a hash algorithm 121 00:04:12,270 --> 00:04:14,370 and calculate a new hash digest. 122 00:04:14,370 --> 00:04:15,720 Then you compare that 123 00:04:15,720 --> 00:04:17,670 to the one that's listed on the website. 124 00:04:17,670 --> 00:04:19,560 If they match, it's assumed that the file 125 00:04:19,560 --> 00:04:21,000 has not been tampered with. 126 00:04:21,000 --> 00:04:23,670 However, if an attacker can produce a malicious file 127 00:04:23,670 --> 00:04:26,310 with the same hash value by exploiting a collision, 128 00:04:26,310 --> 00:04:27,930 the user could be tricked into thinking 129 00:04:27,930 --> 00:04:29,220 they have a legitimate file 130 00:04:29,220 --> 00:04:32,190 when in fact they have a file that has the same hash value, 131 00:04:32,190 --> 00:04:34,470 which means it's subject to a collision. 132 00:04:34,470 --> 00:04:38,430 Historically, MD5, also known as Message Digest Algorithm 5, 133 00:04:38,430 --> 00:04:40,230 was a popular hashing function. 134 00:04:40,230 --> 00:04:41,850 However, over the years, 135 00:04:41,850 --> 00:04:43,680 many vulnerabilities in its structure 136 00:04:43,680 --> 00:04:46,170 made collision attacks much more feasible. 137 00:04:46,170 --> 00:04:48,270 Cybersecurity researchers were eventually able 138 00:04:48,270 --> 00:04:49,800 to create two different sequences 139 00:04:49,800 --> 00:04:52,320 that hashed out to the same MD5 hash, 140 00:04:52,320 --> 00:04:54,720 rendering MD5 unsuitable for further use 141 00:04:54,720 --> 00:04:57,900 in security certificates and encryption technologies. 142 00:04:57,900 --> 00:04:59,970 These collisions do undermine the trust 143 00:04:59,970 --> 00:05:02,340 and reliability placed on cryptographic tools, 144 00:05:02,340 --> 00:05:04,260 and they could potentially allow malicious actors 145 00:05:04,260 --> 00:05:07,320 to impersonate trusted entities, forge digital signatures, 146 00:05:07,320 --> 00:05:10,470 or distribute tampered data while appearing genuine. 147 00:05:10,470 --> 00:05:12,930 Now, the "Birthday Paradox" or "Birthday Attack" 148 00:05:12,930 --> 00:05:15,240 is a counterintuitive probability theory 149 00:05:15,240 --> 00:05:16,410 that plays a significant role 150 00:05:16,410 --> 00:05:18,300 in understanding collision attacks. 151 00:05:18,300 --> 00:05:22,200 The paradox itself posits that in a group of just 23 people, 152 00:05:22,200 --> 00:05:23,700 there's a better than even chance 153 00:05:23,700 --> 00:05:26,190 that two of them share the same birthday. 154 00:05:26,190 --> 00:05:27,900 Translating this to cryptography, 155 00:05:27,900 --> 00:05:30,060 the "Birthday Attack" refers to the probability 156 00:05:30,060 --> 00:05:32,370 that two distinct inputs or files, 157 00:05:32,370 --> 00:05:34,770 when processed through the same hashing function, 158 00:05:34,770 --> 00:05:38,010 will produce the same output, which we call a collision. 159 00:05:38,010 --> 00:05:39,180 Due to this principle, 160 00:05:39,180 --> 00:05:41,070 finding collisions in a hashing function 161 00:05:41,070 --> 00:05:43,680 might be more probable than initially perceived. 162 00:05:43,680 --> 00:05:45,090 For cryptographic systems, 163 00:05:45,090 --> 00:05:47,850 especially hashing algorithms with a fixed size output, 164 00:05:47,850 --> 00:05:51,690 like MD5, SHA-1, SHA-256, and SHA-3, 165 00:05:51,690 --> 00:05:53,760 this means that vulnerabilities to collisions 166 00:05:53,760 --> 00:05:55,590 might arise faster than expected, 167 00:05:55,590 --> 00:05:57,240 and this underscores the necessity 168 00:05:57,240 --> 00:05:58,800 for robust cryptographic designs 169 00:05:58,800 --> 00:06:00,870 and frequent algorithm updates. 170 00:06:00,870 --> 00:06:03,150 Third, we need to talk about quantum computing 171 00:06:03,150 --> 00:06:04,020 and the threat it poses 172 00:06:04,020 --> 00:06:07,200 to our current cryptographic implementations and algorithms. 173 00:06:07,200 --> 00:06:09,420 Before we dive into the threat of quantum computing, 174 00:06:09,420 --> 00:06:11,400 we need to take a moment to define it. 175 00:06:11,400 --> 00:06:13,320 Now, quantum computing involves a computer 176 00:06:13,320 --> 00:06:14,640 that uses quantum mechanics 177 00:06:14,640 --> 00:06:17,610 to generate and manipulate quantum bits, known as qubits, 178 00:06:17,610 --> 00:06:20,340 in order to access enormous processing power. 179 00:06:20,340 --> 00:06:21,870 Now, I know this is a weird definition 180 00:06:21,870 --> 00:06:23,520 because in the definition itself, 181 00:06:23,520 --> 00:06:25,440 I'm using two terms inside of it, 182 00:06:25,440 --> 00:06:27,450 both quantum and computing. 183 00:06:27,450 --> 00:06:30,000 And in this case, I really don't like that definition, 184 00:06:30,000 --> 00:06:32,760 but there really is no better way to explain this. 185 00:06:32,760 --> 00:06:34,830 Now, when you think about a classic computer 186 00:06:34,830 --> 00:06:36,960 like the one you're watching this course on right now, 187 00:06:36,960 --> 00:06:39,900 it's using ones and zeros to process information. 188 00:06:39,900 --> 00:06:41,640 The faster you can flip those bits 189 00:06:41,640 --> 00:06:44,220 and turn a one to a zero, or a zero to a one, 190 00:06:44,220 --> 00:06:46,890 that means the faster you can get information done 191 00:06:46,890 --> 00:06:49,410 and the faster computer you have overall. 192 00:06:49,410 --> 00:06:50,520 Well, at a certain point, 193 00:06:50,520 --> 00:06:52,440 we can't make our computers any faster 194 00:06:52,440 --> 00:06:54,090 by simply flipping bits. 195 00:06:54,090 --> 00:06:56,460 This means we've run out of computing capability. 196 00:06:56,460 --> 00:07:00,210 So to overcome that, we started taking single processors 197 00:07:00,210 --> 00:07:01,950 and putting in dual processors. 198 00:07:01,950 --> 00:07:04,680 And then we made quad processors, which had four cores, 199 00:07:04,680 --> 00:07:07,530 and then we made octa-core processors that had eight cores, 200 00:07:07,530 --> 00:07:10,560 and things like that as we kept speeding up our computers. 201 00:07:10,560 --> 00:07:11,880 Well, with quantum computing, 202 00:07:11,880 --> 00:07:13,860 it is a completely different ball game. 203 00:07:13,860 --> 00:07:15,510 Instead of using ones and zeros, 204 00:07:15,510 --> 00:07:18,180 we use things known as quantum bits, or qubits, 205 00:07:18,180 --> 00:07:21,450 and this can be done in computing or in communications. 206 00:07:21,450 --> 00:07:23,010 Now, when we deal with communications, 207 00:07:23,010 --> 00:07:25,080 we're talking about quantum communications 208 00:07:25,080 --> 00:07:26,550 being a communications network 209 00:07:26,550 --> 00:07:29,100 that relies on using qubits made of photons, 210 00:07:29,100 --> 00:07:30,210 or in our case, light, 211 00:07:30,210 --> 00:07:31,650 to send multiple combinations 212 00:07:31,650 --> 00:07:33,480 of ones and zeros simultaneously, 213 00:07:33,480 --> 00:07:35,160 which will result in tamper-resistant 214 00:07:35,160 --> 00:07:37,230 and extremely fast communications. 215 00:07:37,230 --> 00:07:39,150 Again, I know this is really vague, 216 00:07:39,150 --> 00:07:41,370 but hang in with me here for just a second. 217 00:07:41,370 --> 00:07:42,690 When we talk about quantum, 218 00:07:42,690 --> 00:07:44,910 I keep talking about these things called qubits, 219 00:07:44,910 --> 00:07:46,380 and I talk about the fact that there's a difference 220 00:07:46,380 --> 00:07:48,030 between our traditional electrons, 221 00:07:48,030 --> 00:07:50,340 which are either on or off, ones or zeros, 222 00:07:50,340 --> 00:07:52,080 and this thing called a qubit. 223 00:07:52,080 --> 00:07:54,480 So let's talk about a qubit for a moment. 224 00:07:54,480 --> 00:07:55,950 What exactly is it? 225 00:07:55,950 --> 00:07:58,680 Well, a qubit is really just a quantum bit. 226 00:07:58,680 --> 00:08:01,110 It's composed of either electrons or photons, 227 00:08:01,110 --> 00:08:03,750 so it can either be electrical or made by light, 228 00:08:03,750 --> 00:08:06,870 and it can represent numerous combinations of ones and zeros 229 00:08:06,870 --> 00:08:10,560 at the same time using something known as superpositioning. 230 00:08:10,560 --> 00:08:12,180 Now, this is really the main benefit 231 00:08:12,180 --> 00:08:15,300 of using a quantum computer over a traditional computer 232 00:08:15,300 --> 00:08:17,880 because we're not just going to have a single one or zero. 233 00:08:17,880 --> 00:08:20,310 You can have multiple combinations of ones or zeros 234 00:08:20,310 --> 00:08:23,610 at the same time, and this would all be one qubit. 235 00:08:23,610 --> 00:08:24,630 So you can actually crunch through 236 00:08:24,630 --> 00:08:27,810 a wide variety of potential outcomes simultaneously. 237 00:08:27,810 --> 00:08:30,420 And so it's really great when you're trying to do a bunch 238 00:08:30,420 --> 00:08:32,370 of really complex math problems, 239 00:08:32,370 --> 00:08:33,480 and that's really the whole reason 240 00:08:33,480 --> 00:08:35,429 that we have quantum computing. 241 00:08:35,429 --> 00:08:36,929 When you think about quantum computing, 242 00:08:36,929 --> 00:08:39,390 it's never going to replace the computer on your desktop. 243 00:08:39,390 --> 00:08:41,220 That is not what it's designed for. 244 00:08:41,220 --> 00:08:44,159 Instead, it's designed for very specific use cases, 245 00:08:44,159 --> 00:08:45,900 such as very complex math problems 246 00:08:45,900 --> 00:08:48,240 or trying to do something like modeling of an atom 247 00:08:48,240 --> 00:08:51,120 or some kind of atomic structure, or something like that. 248 00:08:51,120 --> 00:08:52,980 Now, why should I care about quantum 249 00:08:52,980 --> 00:08:54,090 if it's never going to be something 250 00:08:54,090 --> 00:08:55,620 that's sitting on my desktop? 251 00:08:55,620 --> 00:08:57,420 Well, because there's one thing that quantum 252 00:08:57,420 --> 00:08:59,130 is really, really good at. 253 00:08:59,130 --> 00:09:01,380 And what is that? It's math problems. 254 00:09:01,380 --> 00:09:03,870 And if you think about cryptography, all cryptography is 255 00:09:03,870 --> 00:09:06,390 is a bunch of really complex math problems. 256 00:09:06,390 --> 00:09:08,370 And so because cryptography is used 257 00:09:08,370 --> 00:09:10,050 to secure our communications and our data 258 00:09:10,050 --> 00:09:12,630 by relying on how difficult a math problem is to compute, 259 00:09:12,630 --> 00:09:15,300 with a traditional computer, this is actually something 260 00:09:15,300 --> 00:09:17,340 that gives us strength in cryptography. 261 00:09:17,340 --> 00:09:19,020 But if quantum computers are really good 262 00:09:19,020 --> 00:09:20,100 at doing math problems, 263 00:09:20,100 --> 00:09:22,320 and they can do it much, much faster, 264 00:09:22,320 --> 00:09:25,050 that means our cryptography can actually be defeated 265 00:09:25,050 --> 00:09:27,630 and taken down by a quantum computer. 266 00:09:27,630 --> 00:09:29,760 And this is why we have to start thinking about this, 267 00:09:29,760 --> 00:09:30,720 because it's one of the things 268 00:09:30,720 --> 00:09:33,900 we rely so heavily on in cryptography is the key exchange 269 00:09:33,900 --> 00:09:36,000 using asymmetric communications. 270 00:09:36,000 --> 00:09:37,110 Now, one way we do that 271 00:09:37,110 --> 00:09:38,940 is with the public key infrastructure. 272 00:09:38,940 --> 00:09:41,310 Well, that whole concept is that we have to have a way 273 00:09:41,310 --> 00:09:44,670 to do a simple math problem to create a very complex thing 274 00:09:44,670 --> 00:09:47,370 that has a difficult time of being broken back apart. 275 00:09:47,370 --> 00:09:50,010 So for example, if I take two prime numbers 276 00:09:50,010 --> 00:09:52,470 and I multiply them together and I give you the result, 277 00:09:52,470 --> 00:09:54,810 that result is really easy to calculate 278 00:09:54,810 --> 00:09:56,760 if you know what the two prime numbers are 279 00:09:56,760 --> 00:09:58,530 to be able to multiply them together. 280 00:09:58,530 --> 00:10:01,470 This would be, in our case, a public and private key 281 00:10:01,470 --> 00:10:04,350 that really represent two really large prime numbers. 282 00:10:04,350 --> 00:10:06,330 Now, once you have that result, though, 283 00:10:06,330 --> 00:10:07,890 it might be a really large number, 284 00:10:07,890 --> 00:10:10,410 but there's only two things that can be factored into it, 285 00:10:10,410 --> 00:10:12,720 those two keys, the public and private key 286 00:10:12,720 --> 00:10:13,980 that I just gave you. 287 00:10:13,980 --> 00:10:16,170 Those are our two prime numbers. 288 00:10:16,170 --> 00:10:17,760 Well, with quantum computing, 289 00:10:17,760 --> 00:10:20,610 they can make very quick work of this type of math problem, 290 00:10:20,610 --> 00:10:22,200 whereas our traditional computers can't 291 00:10:22,200 --> 00:10:24,150 because it'd involve taking that big number 292 00:10:24,150 --> 00:10:26,280 and dividing it by one and then dividing by two 293 00:10:26,280 --> 00:10:27,300 and then dividing by three 294 00:10:27,300 --> 00:10:29,520 and then dividing by five, then dividing by seven, 295 00:10:29,520 --> 00:10:31,770 and keep going until we get all the prime numbers 296 00:10:31,770 --> 00:10:33,360 and find the two factors. 297 00:10:33,360 --> 00:10:34,770 Well, with quantum computing, 298 00:10:34,770 --> 00:10:36,060 we don't have to do it that way. 299 00:10:36,060 --> 00:10:38,940 Instead, this becomes a relatively easy math problem, 300 00:10:38,940 --> 00:10:40,890 and we can crack it much, much quicker 301 00:10:40,890 --> 00:10:42,360 with a quantum computer. 302 00:10:42,360 --> 00:10:44,640 In fact, asymmetric encryption algorithms, 303 00:10:44,640 --> 00:10:46,920 which are those that rely on this hard math problem, 304 00:10:46,920 --> 00:10:48,330 have been mathematically proven 305 00:10:48,330 --> 00:10:50,430 to be broken by quantum computers. 306 00:10:50,430 --> 00:10:51,660 Now, the good news here 307 00:10:51,660 --> 00:10:53,550 is that at the time of me saying this, 308 00:10:53,550 --> 00:10:56,700 there are no real operational quantum computers in use. 309 00:10:56,700 --> 00:10:58,200 In fact, the only quantum computers 310 00:10:58,200 --> 00:10:59,610 at the time of me saying this 311 00:10:59,610 --> 00:11:01,860 are ones being used for prototyping and development 312 00:11:01,860 --> 00:11:03,540 on a very small scale, 313 00:11:03,540 --> 00:11:05,100 and governments and large companies 314 00:11:05,100 --> 00:11:07,410 are the ones who are spending enormous sums of money 315 00:11:07,410 --> 00:11:10,170 to create these small-scale quantum computers. 316 00:11:10,170 --> 00:11:11,460 Right now, the estimate 317 00:11:11,460 --> 00:11:13,440 is that we won't have a working quantum computer 318 00:11:13,440 --> 00:11:16,770 until at least 2030 for government and large corporate use. 319 00:11:16,770 --> 00:11:18,330 But consumers like you and I 320 00:11:18,330 --> 00:11:19,860 won't see quantum computers available 321 00:11:19,860 --> 00:11:23,550 until sometime in the late 2030s or early 2040s. 322 00:11:23,550 --> 00:11:25,440 But it is something we need to be aware of 323 00:11:25,440 --> 00:11:27,060 because our current encryption algorithms 324 00:11:27,060 --> 00:11:30,300 are not made to withstand the power of quantum computing. 325 00:11:30,300 --> 00:11:32,370 So, that brings us to this idea 326 00:11:32,370 --> 00:11:34,740 of how do you overcome quantum computers 327 00:11:34,740 --> 00:11:36,420 when you're dealing with cryptography? 328 00:11:36,420 --> 00:11:39,120 And that is post-quantum cryptography. 329 00:11:39,120 --> 00:11:40,590 Now, post-quantum cryptography 330 00:11:40,590 --> 00:11:42,720 is a new kind of cryptographic algorithm 331 00:11:42,720 --> 00:11:45,540 that can be implemented using today's classic computers 332 00:11:45,540 --> 00:11:47,760 but would still be impervious to attacks 333 00:11:47,760 --> 00:11:49,350 from future quantum computers 334 00:11:49,350 --> 00:11:52,140 once they're in production and they are widely available. 335 00:11:52,140 --> 00:11:53,550 Now, there's really two methods that we use 336 00:11:53,550 --> 00:11:56,370 to try to create this post-quantum cryptography world. 337 00:11:56,370 --> 00:11:58,800 The first method is to just increase our key size 338 00:11:58,800 --> 00:12:00,810 and to increase the number of permutations that are needed 339 00:12:00,810 --> 00:12:02,040 to be brute-forced. 340 00:12:02,040 --> 00:12:03,150 This works well when you're dealing 341 00:12:03,150 --> 00:12:05,730 with symmetric encryption algorithms like AES. 342 00:12:05,730 --> 00:12:07,680 If I take AES-128 343 00:12:07,680 --> 00:12:10,350 and I increase it to AES-256, for example, 344 00:12:10,350 --> 00:12:14,790 I have doubled my key length from 128 bits to 256 bits. 345 00:12:14,790 --> 00:12:16,470 But I have actually squared 346 00:12:16,470 --> 00:12:18,030 the number of possible combinations 347 00:12:18,030 --> 00:12:19,260 that are going to have to be figured out 348 00:12:19,260 --> 00:12:21,810 by that quantum computer, and this extends the time 349 00:12:21,810 --> 00:12:24,270 and makes it much stronger and harder to crack. 350 00:12:24,270 --> 00:12:27,300 The other way we do this is by working on other approaches, 351 00:12:27,300 --> 00:12:29,700 and researchers are doing that right now. 352 00:12:29,700 --> 00:12:32,760 These are called post-quantum resistant algorithms. 353 00:12:32,760 --> 00:12:34,620 Now, these things are going to be looking at using things 354 00:12:34,620 --> 00:12:36,540 like lattice-based cryptography 355 00:12:36,540 --> 00:12:39,210 and supersingular isogeny key exchanges. 356 00:12:39,210 --> 00:12:40,650 Now again, both of these terms 357 00:12:40,650 --> 00:12:42,180 are not something you need to understand 358 00:12:42,180 --> 00:12:43,530 or go in depth on. 359 00:12:43,530 --> 00:12:45,510 In fact, when we talk about quantum computing 360 00:12:45,510 --> 00:12:46,890 and quantum communications, 361 00:12:46,890 --> 00:12:48,330 we really need to think about qubits 362 00:12:48,330 --> 00:12:50,880 and all the stuff we're talking about with the word quantum. 363 00:12:50,880 --> 00:12:52,200 And I just want you to associate this 364 00:12:52,200 --> 00:12:54,540 with something that can be really easy to break apart 365 00:12:54,540 --> 00:12:57,120 our asymmetric encryption algorithms pretty quickly 366 00:12:57,120 --> 00:13:00,480 once it's in production, but it doesn't exist today yet. 367 00:13:00,480 --> 00:13:02,700 Remember, asymmetric encryption algorithms 368 00:13:02,700 --> 00:13:04,350 are the things that we use for key exchanges 369 00:13:04,350 --> 00:13:07,320 and digital signatures in a public key infrastructure, 370 00:13:07,320 --> 00:13:08,610 and that's really where the threat is 371 00:13:08,610 --> 00:13:10,980 when we start talking about quantum computers. 372 00:13:10,980 --> 00:13:12,990 Again, right now, we do not have 373 00:13:12,990 --> 00:13:15,750 any dedicated quantum resistant cryptography yet, 374 00:13:15,750 --> 00:13:17,700 but in 2022, 375 00:13:17,700 --> 00:13:19,710 the National Institute of Science and Technology 376 00:13:19,710 --> 00:13:21,720 did complete a six-year competition 377 00:13:21,720 --> 00:13:24,030 to select post-quantum cryptographic standards, 378 00:13:24,030 --> 00:13:26,970 and they ended up selecting four viable algorithms. 379 00:13:26,970 --> 00:13:28,560 For your general encryption needs, 380 00:13:28,560 --> 00:13:31,560 NIST recommends that you use the CRYSTALS-Kyber algorithm, 381 00:13:31,560 --> 00:13:33,900 which uses a relatively small encryption key 382 00:13:33,900 --> 00:13:36,030 and is focused on the difficulty of solving learning 383 00:13:36,030 --> 00:13:38,730 with error problems over modular lattices 384 00:13:38,730 --> 00:13:40,770 to provide a post-quantum encryption algorithm 385 00:13:40,770 --> 00:13:41,730 that is roughly equivalent 386 00:13:41,730 --> 00:13:45,300 to the strength of our current AES 256-based algorithms 387 00:13:45,300 --> 00:13:47,190 used in symmetric encryption. 388 00:13:47,190 --> 00:13:48,390 For digital signatures, 389 00:13:48,390 --> 00:13:51,030 NIST recommends you select one of three algorithms, 390 00:13:51,030 --> 00:13:54,930 either the CRYSTALS-Dilithium, FALCON, or SPHINCS+, 391 00:13:54,930 --> 00:13:57,630 although NIST does recommend CRYSTALS-Dilithium 392 00:13:57,630 --> 00:14:00,840 as the primary post-quantum digital signature algorithm. 393 00:14:00,840 --> 00:14:03,120 All these algorithms do focus on the difficulty 394 00:14:03,120 --> 00:14:06,180 of solving structured lattices except for SPHINCS+, 395 00:14:06,180 --> 00:14:09,120 which instead focuses on the use of hashing functions. 396 00:14:09,120 --> 00:14:11,730 So remember, when it comes to cryptographic attacks, 397 00:14:11,730 --> 00:14:13,860 there are three main types that you need to consider. 398 00:14:13,860 --> 00:14:15,990 These are downgrade attacks, collision attacks, 399 00:14:15,990 --> 00:14:17,910 and the threat of quantum computing. 400 00:14:17,910 --> 00:14:19,860 A cryptographic downgrade attack is used 401 00:14:19,860 --> 00:14:21,540 to force a system into using a weaker 402 00:14:21,540 --> 00:14:23,820 or older cryptographic standard or protocol 403 00:14:23,820 --> 00:14:25,470 than what it's currently using. 404 00:14:25,470 --> 00:14:27,330 A collision attack is going to be used 405 00:14:27,330 --> 00:14:28,590 to find two different inputs 406 00:14:28,590 --> 00:14:30,450 that produce the same hash output 407 00:14:30,450 --> 00:14:31,830 by using the principles demonstrated 408 00:14:31,830 --> 00:14:33,540 by the Birthday Paradox. 409 00:14:33,540 --> 00:14:35,100 Quantum computing, on the other hand, 410 00:14:35,100 --> 00:14:36,900 is more at the theoretical stage today 411 00:14:36,900 --> 00:14:38,070 than the practical stage, 412 00:14:38,070 --> 00:14:40,440 and it harnesses the principles of quantum mechanics 413 00:14:40,440 --> 00:14:43,350 to process vast amounts of information simultaneously 414 00:14:43,350 --> 00:14:44,340 in order to enable it 415 00:14:44,340 --> 00:14:46,860 to potentially break traditional encryption algorithms 416 00:14:46,860 --> 00:14:50,580 like RSA and ECC in just a few seconds 417 00:14:50,580 --> 00:14:52,860 by rapidly factoring large prime numbers 418 00:14:52,860 --> 00:14:55,290 or solving discreet logarithmic problems. 419 00:14:55,290 --> 00:14:57,630 Since this capability threatens the foundational security 420 00:14:57,630 --> 00:14:58,950 of our current digital communication 421 00:14:58,950 --> 00:15:00,420 and data storage systems, 422 00:15:00,420 --> 00:15:02,400 it is one that you should continually read up on 423 00:15:02,400 --> 00:15:04,200 and study over the next few years 424 00:15:04,200 --> 00:15:06,720 as quantum computing and post-quantum algorithms 425 00:15:06,720 --> 00:15:08,100 begin to become a larger part 426 00:15:08,100 --> 00:15:10,143 of our enterprise systems and networks.