WEBVTT 00:00.710 --> 00:07.700 Symmetric block mode encryption is the cornerstone of pretty much everything we do in terms of encryption 00:07.970 --> 00:08.920 on the Internet. 00:08.960 --> 00:15.380 And it works really well and it's very very robust but block encryption has a problem. 00:15.380 --> 00:20.120 Now what I'm going to use is a E-S encryption as an example here. 00:20.120 --> 00:26.420 But what we're going to do is set up a a little diagram and I want to show you one little problem with 00:26.420 --> 00:27.610 block encryption. 00:27.860 --> 00:32.540 So let's start off by setting up a little encryption machine. 00:32.540 --> 00:36.170 So I want to start off our diagram here is in the upper right hand corner. 00:36.200 --> 00:41.140 Let's make like a little conveyor belt of ones and zeros that are coming out to be encrypted. 00:41.150 --> 00:47.660 Remember the whole idea of block mode is that it's going to take a chunk of ones and zeros and it's 00:47.660 --> 00:52.090 going to drop it into this encryption block where we apply the key. 00:52.340 --> 00:55.600 And then it is output as encrypted text. 00:55.670 --> 00:59.290 And I'm going to make another little conveyor belt here as we show each block. 00:59.300 --> 01:01.110 Now I want you to look very closely here. 01:01.220 --> 01:07.850 So as this runs you're going to see that we're only grabbing eight ones and zeros at a time. 01:07.850 --> 01:14.000 Now if you've been watching in other episodes you know that we don't do anything a little 8 bit chunks. 01:14.090 --> 01:20.210 Today's powerful symmetric block encryptions like abs are going to be working in an absolute minimum 01:20.270 --> 01:22.190 at 64 bit chunks. 01:22.190 --> 01:26.780 The reason I'm doing little 8 bit chunks is just to make the diagram a little bit easier for you to 01:26.780 --> 01:28.100 understand. 01:28.100 --> 01:34.220 So the bottom line is is that with any form of symmetric block mode if you grab that chunk and you drop 01:34.220 --> 01:41.060 it through your encryptor with the key and you output it every time you grab a 64 bit chunk or whatever 01:41.060 --> 01:48.820 the chunk size is and you keep encrypting it with the same key you're always going to get the same output. 01:48.830 --> 01:56.480 So this mode which we call electronic codebook is a bit of a problem in that we run into situations 01:56.570 --> 01:58.280 where patterns can appear. 01:58.280 --> 02:00.400 Let me give you a great example. 02:00.500 --> 02:04.000 This cute kid right here is my little grand niece Mattie. 02:04.100 --> 02:08.250 Now what I want to do is take this image and I want to encrypt it. 02:08.450 --> 02:15.410 So the problem we run into is that the individual pixels of most images are going to be of a particular 02:15.410 --> 02:20.770 size and it depends on the type of file format jpeg or whatever you might have. 02:20.870 --> 02:26.810 But what will start happening is if we encrypt this in ECB mode it's going to end up looking something 02:26.810 --> 02:27.870 like this. 02:29.120 --> 02:29.690 Yeah. 02:29.750 --> 02:34.460 All the individual pixels are encrypted but because they're encrypted with the same block size with 02:34.460 --> 02:40.640 the same key they keep coming out absolutely identical so it may not be a black pixel it might be a 02:40.640 --> 02:45.340 yellow pixel but the bottom line is I can still make out Miss Mattie there. 02:45.710 --> 02:49.280 So we've got a problem with symmetric block encryption. 02:49.280 --> 02:54.920 The problem is is that we can generate patterns that can actually give away our data. 02:54.920 --> 02:59.880 Now in that example I just showed you we're using a photograph that's pretty easy to make that out. 02:59.960 --> 03:05.110 But we can run into the same types of problems with voice with databases with all kinds of stuff. 03:05.150 --> 03:10.010 The patterns aren't as easy to show in a video but the same problems exist. 03:10.010 --> 03:15.750 So what we do is we don't use ECB mode with our symmetric encryptions. 03:15.830 --> 03:23.000 What we do instead is we use different types of what are known as block modes to obfuscate the data 03:23.000 --> 03:23.800 better. 03:23.840 --> 03:26.960 So there are a number of block modes that are on the exam. 03:26.960 --> 03:29.600 So what I want to do is go through each one of them. 03:29.660 --> 03:36.290 The basic trick to all of these different block mood's is that we're going to encrypt something and 03:36.290 --> 03:39.780 then we're going to use that encryption to help encrypt the next one. 03:39.890 --> 03:44.420 So we end up there really kind of like a chain and they're actually very interesting they're on the 03:44.420 --> 03:45.080 exam. 03:45.140 --> 03:49.860 So let's go through each and every one of these because you're going to see it you're ready. 03:49.940 --> 03:54.990 The first black vote I want to show you is called Sipher block chaining or CBC. 03:55.010 --> 04:01.520 So we've got our encryption methodology here we've got our plain text at the top we've got our output 04:01.520 --> 04:05.120 text which is going to be we'll start loading there at the bottom. 04:05.120 --> 04:08.590 And here's our encryption but we're going to do something different this time. 04:08.600 --> 04:12.920 First of all we're going to add something called an initialization vector. 04:12.950 --> 04:14.720 So here's an initialization vector. 04:14.720 --> 04:17.630 It is the same size as the blocks themselves. 04:17.660 --> 04:23.030 And what we're going to do is before we encrypted we're going to take that first block and we're going 04:23.030 --> 04:28.850 to do an x or against this initialization vector then we're going to go ahead and encrypt it and put 04:28.850 --> 04:30.510 the encryption output down there. 04:30.530 --> 04:34.210 But we're also going to keep another copy. 04:34.430 --> 04:38.960 And this is going to be used for the next block in that way. 04:38.960 --> 04:44.570 What's happening is that we're not going to get the exact same data for every different key because 04:44.570 --> 04:47.240 we're actually changing it a little bit every time. 04:47.240 --> 04:53.680 Next is Sipher feedback now with Sipher feedback or CFB. 04:53.720 --> 04:58.040 You're going to see it's actually pretty similar to what we just looked at in this case though we're 04:58.040 --> 05:04.400 going to take that initialization vector and we're going to go ahead and encrypt we encrypt the initialization 05:04.400 --> 05:11.500 vector and then we take the output of that encryption and we x or it to the first block. 05:11.510 --> 05:12.760 So here we go ahead. 05:12.860 --> 05:17.710 Outputs and and it just gets X ored to that first block. 05:17.710 --> 05:18.770 So it comes in. 05:18.830 --> 05:26.150 We put one copy of the output down on to the ciphertext and then we go Hinze cycle that output and use 05:26.150 --> 05:26.530 that. 05:26.540 --> 05:29.250 And we go ahead and encrypt that for the next one. 05:29.300 --> 05:32.030 Third is output feedback. 05:32.120 --> 05:38.210 Now with output feedback what we're gonna do is we're going to take one initialization vector and we're 05:38.210 --> 05:43.120 going to go Edmond crypt that we're going to take the output of that encryption and we're going to X 05:43.120 --> 05:47.170 or it to that first block then we go ahead and put that output there. 05:47.180 --> 05:52.160 Now in this case what's actually kind of interesting is we just keep using that same initialization 05:52.160 --> 05:57.000 vector C-T our counter is a little bit different than the others. 05:57.140 --> 06:01.220 In this particular case what we're going to have is a non-SS Dahlia. 06:01.220 --> 06:08.000 Now in this particular example it's only four bits long but again in the real McCoys it's much bigger 06:08.000 --> 06:08.660 than this. 06:08.720 --> 06:12.670 And then the counter value that continues to increment in binary. 06:12.890 --> 06:17.780 So in this case what we're going to do is we're going to combine the non-SS and the counter we go ahead 06:17.780 --> 06:25.610 and encrypt that and then we take the first block of the plain text an X or it to create the first block 06:25.850 --> 06:33.140 of ciphertext the next time we do this all we do is we increment the counter so the counter gets incremented. 06:33.140 --> 06:38.000 We go ahead and concatenate that to the non-SS we encrypt all that. 06:38.180 --> 06:42.740 And then we go ahead and take that and output that to the second block. 06:42.950 --> 06:49.580 The power of these different forms of block modes is that they get around this limitation of symmetric 06:49.580 --> 06:51.030 block encryptions. 06:51.230 --> 06:56.090 So by using anything other than ECB we should be able to do something like this. 06:56.090 --> 07:01.240 So here's my cute little niece Mattie one more time and applying a good block mode to it. 07:01.280 --> 07:03.170 It should look something like this. 07:03.530 --> 07:05.480 Totally randomized data. 07:06.410 --> 07:07.590 For the exam. 07:07.640 --> 07:15.420 Make sure you're aware of the fact that number one nobody uses ECB anymore it's simply not done there 07:15.420 --> 07:21.120 are a number of different block moods out there and different types of cryptosystems will take advantage 07:21.120 --> 07:25.120 of them depending on what they need for their particular application. 07:25.920 --> 07:45.710 In.