Cryptography and Ciphers
|
Student: I really understood the explanation of modular arithmetic. Mentor: Good. Did you know that this kind of mathematics is used in defending the United States and other countries during wars and other times when there is information that has to be kept confidential? Student: How would someone use modular arithmetic during a war? Mentor: It was used during the U.S. Civil War in the 1860's, and even thousands of years before that, in Caesar's Roman Empire. People who wanted to communicate with allies but not their enemies would send encrypted messages back and forth. Student: What is an encrypted message? Mentor: Encrypted messages take the letters and numbers of a message and transform them into a different series of letters and numbers that do not make sense unless you know the code to unscramble them. Student: And if you do know the code? Mentor: Then you get to read the secret message. Amazingly complex scrambling procedures can stump highly trained people and even computers. Right now I would like to talk about some simple ciphers, so you can practice the basic idea. Student: What is a cipher? Mentor: A cipher is the method by which you encrypt a message. The ones that I want to teach you have to do with numerical operations: multiplication, division, addition and subtraction, so we must first create a numerical alphabet by assigning numbers, beginning with zero, to each letter of our English alphabet. Student: So A would be 0, B would be 1 ..... like this:
Mentor: Good. Most ciphers use modular arithmetic during some steps of the encryption and decryption process. We have used the numbers 0 through 25 to represent 26 English letters, so we will use mod 26 in all of our cipher examples. Let's begin with what is called a shift cipher. First we must translate our message into our numerical alphabet. For example, the code name 'James' looks like this:
Now choose the number you want to shift by. Student: Let's choose 7. Mentor: We need to remember this number, which we will call the 'B-shift' value, to both cipher and later *de*cipher the message. We now 'shift' our cipher by adding B to each of the numbers from our code word like this:
Now you have :
We can now translate the numbers back into letters using our numerical alphabet. The name 'James' encrypted with a B-shift of 7 is:
Now try another example, the phrase 'is a spy' using the same B-shift. Student: So first we translate the code words into the numerical alphabet like this:
I S A S P Y
8 18 0 18 15 24
Then we add the B-shift to each number,
and then I translate back to letters, so 15 is P, 25 is Z, 7 is H, 25 is Z again, and 22 is W, but what is 31? Mentor: Maybe mod 26 might be useful? Student: Right! 31 mod 26 is 5, which corresponds to F, so 'S P Y' becomes 'Z W F'! Mentor: Now if you wanted to send the message 'James is a spy.' to someone, but didn't want everyone (like James!) to be able to read the message, you could send the message 'Qhtlz pz h zwf.'. A good cryptographer, however, might use the size of the words and your punctuation to help them figure out your cipher and your secret message! For example, what English words have only one letter? Student: 'I', like 'I went to the store' and 'a' as in 'for a loaf of bread'. But 'I' would be capitalized, so 'h' in our code would have to be an 'a'! Mentor: Good! That's exactly how a cryptographer would do it as well. We don't want just anyone to figure out our cipher, so we make it harder by removing punctuation and grouping the letters in 'words' that are usually five letters long, so our message might look something like this:
QHTLZ PZHZW F
In order for someone to translate the message, they would have to use our cipher backwards. Student: So first they would translate the letters into numbers again, so QHTLZ would again be
16 7 19 11 25
Mentor: But in order to solve the cipher they would also need to know that we used a B-shift of 7... Student: ...which, working backwards we would need to subtract! So 16 - 7 = 9, which is 'J', 7 - 7 = 0, which is 'A', and so on... Neat! Mentor: What about the last letter? Student: F becomes 5, 5 minus 7 is negative 2 ! What happened? Mentor: You forgot to reverse the 'mod 26' calculation. If you remember, the mod 26 operation made us subtract a multiple of 26, in this case just 26*1 = 26. Try adding it back. Student: Negative 2 plus 26 is 24, which translates to Y. Good, that makes sense. Mentor: Shift ciphers can also work in the opposite order, where you subtract the B-shift first when you are encrypting and then add it back when you are decrypting. Let's try another kind of cipher. It is called a multiplication cipher. It is similar to the shift cipher, except that you multiply and divide instead of add and subtract. Choose an easy example word for us to use. Student: How about S I M P L E which in the numerical alphabet is
18 8 12 15 11 4
Mentor: Good. We'll use the number 7 again, but this time we will call it 'A' so as not confuse it with the shift. Now multiply the numeric alphabet letters each by the 'A-multiplier':
The numbers are too big for our numeric alphabet, so we need to use mod 26 again. Student: All right.
126 mod 26: 126/26 = 4 with remainder 22, so 22
56 mod 26: 56/26 = 2 with remainder 4, so 4
84 mod 26: 3 with remainder 6, so 6
105 mod 26: 4 with remainder 1, so 1
77 mod 26: 2 with remainder 25, so 25
28 mod 26: 1 with remainder 2, so 2
So the encrypted message is W E G B Z C. That was 'simple'! Mentor: Perhaps, but now how do you decrypt it? Student: Hmmm... Working backwards, W is 22. Divide by our 'A-multiplier' is 22 over 7 which is three remainder one... What do I do with that!?! Mentor: There's the hard part! Before we divide we need to reverse the 'mod 26' operation again. As before, let's add 26 and then see if the division works. Student: Okay. 22 + 26 = 48, which is not evenly divisible by 7. Mentor: Try a few more times. Student:
That works, because 18 is our original letter 'S', but that's not really so simple anymore. Mentor: There is another method. Try this: multiply each of your numerical alphabet code numbers by fifteen and then mod 26. Student:
That worked, but where did the 15 come from? Mentor: Modular arithmetic is not quite the same as regular arithmetic. In regular arithmetic the inverse of a multiplier A would be 1/A, for example 7 and the fraction 1/7th. In modular arithmetic a multiplier will only have an inverse if the multiplier (for example 7), and the modular base (in our case, 26) have no common factors greater than one! The modular inverse - we will call it A' (A-prime) - will be a whole number that also has no common factors with the modular base. 7 and 15 are modular inverses in mod 26. Student: Is there a way to figure out what the modular inverse of a number is? Mentor: Just as in real arithmetic, a number and it's inverse will multiply in modular arithmetic to give you 1, so (A * A') mod 26 = 1. Start by making a list of multiples of 26, and then add one to each of them. Student:
26 * 1 = 26, plus one is 27
26 * 2 = 52, plus one is 53
26 * 3 = 78, plus one is 79
26 * 4 = 104, plus one is 105
26 * 5 = 130, plus one is 131
26 * 6 = 156, plus one is 157
26 * 7 = 182, plus one is 183 ...
Mentor: That should be enough. Now we know that all of those numbers, mod 26 is equal to 1. Are any of those numbers evenly divisible by seven? Student: Yes, 105 divided by 7 is ... 15! Mentor: What if we wanted to use the multiplier 9 instead? Does 9 have an inverse? Student: Hmmm. 27 divided by 9 is 3, so 9 and 3 are modular inverses, right? Mentor: In mod 26, yes. If we used a different modulus, different numbers would have different inverses, and some would not have any inverses at all. Student: Like when they share a common factor with 26, right? Why is that? Mentor: Let's see why by using the number 6, which shares the factor 2 with 26. Here are some letters encoded by multiplying by 6:
A is 0, times 6 = 0, mod 26 = 0
B is 1, times 6 = 6, mod 26 = 6
C is 2, times 6 = 12, mod 26 = 12
D is 3, times 6 = 18, mod 26 = 18
E is 4, times 6 = 24, mod 26 = 24
F is 5, times 6 = 30, mod 26 = 4
...
M is 12, times 6 = 72, mod 26 = 20
N is 13, times 6 = 78, mod 26 = 0
0 is 14, times 6 = 84, mod 26 = 6
P is 15, times 6 = 90, mod 26 = 12
Student: But A and N are coded the same, as are B and O, and C and P. Does this pattern continue? Mentor: Yes. Because 26 and 6 share the factor of two, each number will represent two letters. We cannot reverse the process because we would not know which of the two letters to select. This happens when the number we try to use as a multiplier shares a factor with 26. Student: So it won't happen with 5, but it will happen with 13? Mentor: Try it! Student:
Student: For a common factor of 13 there are thirteen letter possibilities for each number, like you said! So 13 doesn't have an inverse, but 5 does. Mentor: Look back at your 'multiples of 26 plus one' list and see if you can find it. Student: Okay... 27... 53... 79... 105! 105 divided by 5 is 21 so 5 and 21 are inverses in mod 26. So for mod 26 we have 7 with 15, 5 with 21, and 3 with 9, are there any others? Mentor: Yes. These pairs all work for mod 26, but remember they won't apply to another system with a different number of letters!
Student: How hard can ciphers get? Mentor: There are ciphers that stump even the best programmers with the newest computers. It has to be that way in order to be able to stay ahead of the technology of the opposition. I am sure that as technology gets better and faster, ciphers do the same. The ciphers we have been covering were some of the first, dating to way before we had computers. In fact, Caesar liked to use the shift cipher with 3 as the shift. Student: I bet the newer ones have a lot of really hard math in them. Mentor: Yes they do. Let's move on to the last cipher we are going to talk about. It is called an affine cipher. Affine means linear, so this cipher takes on the same form as a line: (A*x + B) mod 26 With this, you must have both the A-multiplier and B-shift to decode the message. Note that if A = 1 you have a normal shift cipher, and when B = 0 you have a multiplication cipher. Again here the greatest common factor for A and 26 has to be 1, in order to have a way to decode the message. Student: Can we use 7 again? Mentor: Okay, let's use that for our A and we can use 5 for our B. Student: So if I let someone know my A and B and what type of cipher then they can decrypt the message. Mentor: Right. Let's practice with the word 'chair'. Student: Let me try. The first step is C H A I R ---- 2 7 0 8 17 Now we have (7*x +5) mod 26, so:
7*2 + 5 = 19, mod 26 = 19 which is T
7*7 + 5 = 54, mod 26 = 2 which is C
7*0 + 5 = 5, mod 26 = 5 which is F
7*8 + 5 = 61, mod 26 = 9 which is J
7*17 +5 = 124, mod 26 = 20 which is U
The encrypted message is T C F J U. Mentor: Very good. Now to work in reverse, so we should subtract the 5 and then multiply by our modular inverse, which we remember for 7 in mod 26 is 15. Student: So,
T = 19: 15*(19-5) mod 26 = 210 mod 26 = 2
C = 2: 15*( 2-5) mod 26 = -45 mod 26 = ??
F = 5: 15*( 5-5) mod 26 = 0 mod 26 = 0
J = 9: 15*( 9-5) mod 26 = 60 mod 26 = 8
U = 20: 15*(20-5) mod 26 = 225 mod 26 = 17
Wait! How do I mod a negative number like -45? Mentor: Just like a positive number, so that -45 mod 26 = -19, but then you add 26 to make the result positive, so that -45 mod 26 = -19 + 26 = 7. Student: So the final decoded answer is 2 7 0 8 17, which gives: CHAIR. Cool! |
