Advanced Encryption Standard Algorithm

In the next days I will be writing about the AES algorithm, so I am going to start with a brief historical introduction and basic description of the algorithm before I move into more technical stuff.

In January 1997, the US National Institute for Standards and Technology (NIST) announced a competition to develop a new encryption algorithm to replace the old Data Encryption Standard DES. This latter algorithm was urged to be replaced due to computational power improvements making of DES a weak cipher. Some computers at the time were capable to break DES inspecting its key space of 2^56 keys by brute force.

The new encryption standard developed, was the Advanced Encryption Standard or AES becoming the United States encryption standard for sensitive information defined in Federal Information Processing Standard (FIPS 197). Unlike the design of DES, AES was designed as an open and public competition and many companies and researchers around the world participated in the contest . In total fifteen new algorithms were submitted and a second round narrow the choice to five of these. Among them was the Rijndael algorithm which in 2001 was announced to be the winner cipher. Rijndael was designed by two Belgium cryptographers; Joan Daemen and Vincent Rijmen from which the algorithm receives its name.

AES is a symmetric algorithm based on the original Rijndael algorithm. It encrypts in blocks of 128 bits of data called plaintext, and transforms it into an encrypted new block of the same length called ciphertext. The difference between Rijnadael and AES is that Rijndael encrypts blocks of data of 128, 192 and also 256 bits and AES only encrypts blocks of 128 bits. In order to encrypt AES makes usage of a key which is going to be secret and shared between the encryption/decryption parties. The key was designed to have sizes of 128-bits, 192-bits or 256-bits length.

Image taken from a very cool AES animation found in: http://www.formaestudio.com/rijndaelinspector

The algorithm performs a number of rounds transforming the plaintext depending on the key length. For example, when using a 128 bit key the algorithm executes 10 rounds, when using 192 bit keys 12 rounds and finally when using a 256 bit key 14 rounds are executed. Through the process the 128 bit block of plaintext will go to what we are going to call as the AES encryption process illustrated in the image above, and the key will go through a process called the KeyExpansion or sometimes referred to as the KeySchedule (Generate expanded keys for the different AES rounds). The 128 bit block of plaintext will be processed in an intermediate state which “can” be seen as a 4x4 matrix holding 16 bytes (is not always a real matrix and can be a vector).

If we look back to the AES process illustrated this refers particularly to AES-128 bit key which executes a total of 10 rounds. As we can see, AES consists of only 4 operations: SubBytes, ShiftRows, MixColumns and AddRoundKey. It is important to note that the very last round differs from the rest of rounds by skipping the MixColumns process. Before the plaintext is transformed to ciphertext it has to pass through all these operations transforming gradually the plaintext stored on the state matrix into encrypted text which will result into the final cipher text.

This is an introductory and general explanation of the AES algorithm. In future posts I will be explaining how the different AES operations work. However, if in the meantime you feel interested in getting to know more about the AES steps I really recommend checking out this AES encryption animation (http://www.formaestudio.com/rijndaelinspector) it gives a very simple understanding of how AES works without digging into complex mathemtical explanations. Though, if that is what you need “The Design of Rijndael” or the “AES specification” might be more useful.