Hi there. So the first
building block of cryptography that we'd
like to talk about are block ciphers and block
ciphers are basically an algorithm where you use it to take a
plaintext message. So something that
you want to keep secret and you put it through the block cipher and get
out the cipher text output. So cipher-text is the plaintext
message all jumbled up so no one but you
can read it if you have the private key and
the decryption algorithm, and that will be the focus
for this lesson. After this lesson, you will be able to explain
what a block cipher is and the types of block cipher modes of
operation. Let's dive in. Block ciphers come in
specific bit length sizes. So for example, a block cipher of a 128 bits or 256 bits size. We can encrypt a message having at most the same bit length
as a block cipher. When we use the term
'block' we mean a fixed length group of bits. This is what is called a block, a fixed length group of bits. With that said,
the ciphertext output, given a private key and
a plaintext message should have no relationship to the plaintext message input. This means that you should not
be able to see any sort of relational or a pattern that maps back to
the plaintext message. Now, with the
introduction completed, we can think about a first stab
at using a block cipher and the way I'm going
to show you now how one would start off with
using a block cipher. This method is just a way
for you to get a feel for what it is that
the block cipher is doing. What I'm talking about here, this method right now, is called ECB mode. This is a mode of operation
for using a block cipher, but don't encrypt your data in this way that I'm
about to show you. This is because you
will be able to see a pattern in the ciphertext if the original message has a regular pattern and this
can leak information. With that said, to get an instinct for how
block ciphers are used, we can look at this example. Say you had a message
and a message is a set of bits
of ones and zeros. Let's assume that
that length of this message, that set of bits, its length is divisible by
one length of a block cipher. So what we go on and do is we
chop up this total size of message length into little pieces with a length of the
block cipher size. So here we have the message
chopped up into those little pieces
with one piece each being equal
to the block size. Now, given a private key, what we could do
for one operation is to take one block of
the message and feed it into the block cipher given a private key and have as output some ciphertext which
has the same number of bits as the block cipher
and the message block. Now we do the same operation
with the same private key. We take the next block
of that message and we use that as input
into the block cipher and we get out another
ciphertext block with the same number of bits as the block cipher and
the message block. We continue this
operation until we no longer have any message
blocks to encrypt and we take the set of
cipher text blocks and we concatenate them into one giant ciphertext
block as seen here. Now again this is
called ECB mode. Do not do this. There is a better way to encrypt and we'll
talk about that next. But with this
explanation you might be wondering what happens if the message that we are
wanting to encrypt is not divisible evenly by
one block cipher size? What we do is, we
use block ciphers in different block cipher modes and I'm going to talk
about that next. Block cipher modes are what we use when we need to encrypt a message that isn't evenly
divisible by a block size. What you need to do is, you need to pad the total
message with some zero bits. In other words, you pad it
with some extra bits such that the total message length
then becomes evenly divisible by
one block cipher size. So effectively, you pad it with a set of zero bits so it
doesn't change the meaning. Then you perform
the encryption in CBC mode for example.
This is what's done. Now you can use a block cipher as a stream cipher
so that you don't have to pad it with
extra bits and you can do this using the mode
called CTR mode. CTR mode requires the use of
something called a nonce. A nonce is a number
that you use only once, and it adds a randomization
to the resulting ciphertext. But according to Ferguson,
Schneier and Kohno, CTR is only good if you can guarantee that that nonce
is always unique. Now, I mentioned CBC
cipher mode earlier. We're going to take a look at that in a little bit more detail. As I mentioned earlier, there are several
different ways or block cipher modes that
you can use in order to use a block cipher and we
touched upon the ECB mode, which is just going
and encrypting each divisible
block-by-block cipher and getting the ciphertext out. Remember, do not use ECB mode. Now you can also use
OFB or CTR and also, CBC is another way of
using a block cipher. We're going to concentrate
on the CBC mode in this lecture because it's the one that's
currently recommended. In general, the way CBC mode works is that when
you have plaintext, you XOR it with the previous
ciphertext result and use that XOR result as input to the block cipher
algorithm given a private key, and you get the resulting
ciphertext out. You keep doing
this process until you have no more message
blocks to encrypt. So if we look at this
a little further, the name of the game is to randomize the plaintext message using the previous
ciphertext block as I had mentioned earlier. You would do this by
XORing or performing an XOR operation with a plaintext message and
a previous ciphertext result. So in a little bit more detail, what ends up happening is, if you imagine
the plaintext message as one giant set of bits, then we chop up
that set of bits into smaller bit groups that are the same size as
the block cipher block size. As you see here the plaintext here has already been chopped up into the same size as
the block cipher block size. Now again, what happens is, the ciphertexts from
the previous operation is used as input to the XOR
operation between itself and the current plaintext
block that we want to encrypt and we take
that result and we put it into the block cipher
encryption algorithm. Then given the private key, we run that block cipher
encryption algorithm and out comes ciphertext. Then we use that
ciphertext as input into the XOR operation for the next block that
we want to encrypt. But now you must be asking, what happens to
the very first block? There is no previous
ciphertext to XOR it with. What we would need to do is use something called an
initialization vector. The initialization vector
or also known as IV, is a set of bits that has the same bit length as the block cipher
block size that we use to XOR the very first
plaintext block with, and that is what
we're able to get in order to randomize
the first plaintext block. So that's what initialization
vector does and what's suggested is to use
a random initialization vector. According to Ferguson,
Schneier and Kohno, the block cipher mode to use is CBC block cipher mode
with a random IV. Now, to take this a step further, let's look at this from
a higher level perspective. In summary, before you start
an encryption session, you need a private key,
a plaintext message, and some encryption
algorithm that is set to a specific cipher mode and
the resulting output of this encryption algorithm is a ciphertext message
that you need to send to whomever you want to send this message to, so the receiver. What's recommended
by Ferguson et al is the encryption
algorithm called AES with a 256 bit block size in CBC mode with a random IV.