Work

The Key Distribution Problem: Prior Advances and Future Challenges

Year: 2020
Type: article
Abstract: Introduction and BackgroundIn order to ensure secure communications, known information must be shared among both communicating parties.This piece of information, called a cryptographic key, allows the sending party to successfully encrypt, and the receiving party to successfully decrypt the transmitted information.However, because both parties must have possession of this same key in order to successfully communicate, cryptographers throughout history have struggled with finding a method to transmit this key from the sending to the receiving party without it being intercepted along the way.For a time, until the advent of electronic cryptography, keys had to be distributed in person, via mail or by courier, as these were the only relatively secure methods for sending information from one person to another.Later in the computer age, however, computer scientists and cryptographers alike worked to devise schemes to successfully and securely transmit key material over insecure channels.As a result, there today exist solutions to the Key Distribution Problem for both symmetric, as well as public key cryptosystems.On the symmetric side, the two most promising solutions are to either use a Key Derivation Function or a Key Wrapping scheme.Both of these methods provide a secure means for a symmetric cryptographic key to be transmitted from sender to receiver without the threat of outside intrusion.With public key systems, both Merkle's Puzzles and the Diffie-Hellman Key Exchange are viable, efficient solutions.While these methods differ by the type of cryptosystem in which they are employed, there also exists disparity in their respective methods of action.Both methods used in symmetric systems, Key Derivation Functions and Key Wrapping, are protocols which directly transmit the key material.In other words, it is the key itself (though successfully shielded) that is being sent over the insecure channel.Conversely, both protocols used in public key systems are key agreement protocols.With these schemes, trivial information is sent over the insecure channel, and this information is used by both parties to construct the same shared key governed by a predetermined set of rules.With all four of these protocols, cryptographic keys are able to be successfully and securely transmitted from one party to another.Trouble lies ahead, though, when considering the future of computing and its effect on the cryptographic landscape.Quantum computing is the future.In fact, today, multiple companies such as Google and IBM already have working quantum computers which are able to complete tasks at speeds exponentially faster than their classical computer counterparts.When quantum computing becomes a reality, all cryptographic key distribution as well as encryption/decryption protocols will be rendered null and void.This is because quantum computers are able to withstand an exponentially larger computational load than common classical computers.Because of this, they would be easily able to break all current encryption/decryption and key distribution.Due to this conundrum, many "quantum safe" cryptographic algorithms have been developed, in which the computational effort required to break these ciphers is too great even for a quantum computer to realistically solve.The Key Distribution Problem carries over to the quantum space as well, and as such various "quantum safe" key distribution protocols have been developed as well, such as the well-known BB84 protocol.The work being done now to create standards of cryptography that can remain in a quantum future will ensure that people's data and privacy remain safeguarded for the years to come. Key Distribution in Symmetric CryptosystemsFor the majority of the history of cryptography, symmetric ciphers were the gold standard for secure communication.Until the advent of more advanced computing technology beginning in the late 1930s, manual, symmetric ciphers were for the most part, the only type of encryption system that had been developed and was the only one which was widely used.Looking at these ciphers through a lens of the Key Distribution Problem, just as the ciphers themselves were manual, so were key distribution solutions.These included physically transporting the key, either over mail, by in person meeting or via courier, to the desired recipient of the future message.It was not until computerized encryption algorithms were created (such as DES) that solutions for distributing keys over networks, instead of over physical distances were needed.As such, multiple protocols for key distribution in symmetric cryptosystems were devised. Key Derivation FunctionsThe first solution to the Key Distribution Problem for symmetric cryptosystems is the use of a Key Derivation Function.A Key Derivation Function is a mathematical function which takes a fixed input and produces an output which is used by both the sender and receiver as the encryption/decryption key (Sönmez, 2009).The information that is transmitted over the insecure channel is called the Key Derivation Key and is consists of a random string of bits.The sender will transmit this Key Derivation Key to the receiver.Both parties will then use the Key Derivation Key as the input to the agreed upon Key Derivation Function, and the result will be used as the encryption/decryption key.Mathematical operations called pseudorandom functions are used to generate the key material (Sönmez, 2009).These functions are designed to behave as close to random as possible, since it is surprisingly difficult to successfully mathematically model random behavior.The random behavior of these functions is critical in a scheme like this, as it ensures that the creation of the key is one-way.That is, it would be extremely difficult for an adversary to construct the correct key following recovery of the Key Derivation Key.This is due to the relative random behavior of strong Key Derivation Functions.The more random the function behaves, the more difficult it is for an adversary to reconstruct the key from intercepted material.When considering the security of a Key Derivation Function, it is also important to look at the length of the Key Derivation Key being used.This depends on the pseudorandom function being used in the specific key derivation.The two main pseudorandom functions used in most key derivations, Hash Message Authenticated Code (HMAC) and Cipher Message Authenticated Code (CMAC), require different key lengths to ensure a secure key derivation (Sönmez, 2009).When using an HMAC function, the Key Derivation Key is allowed to be any length (Sönmez, 2009).On the other hand, when using CMAC as the pseudorandom function, the length of the Key Derivation Key is determined by the length of the blocks being used in the CMAC, which is itself a form of block cipher (Sönmez, 2009).The security of a Key Derivation Function also depends on the strength of the pseudorandom function being used.In this case, the strength of the pseudorandom function is defined as the amount of effort it would take an adversary to correctly map a random input string of bits to the correct output string, in other words, the key (Chen, 2008).Therefore, the greater the amount of work it would take an adversary to recover correct output material, the greater
Author: Lee, Benjamin
Institution: Trinity College
Language: English
Open Access status: green
OpenAlex ID: W7029538434
FWCI: 0.098
Cites: 11
Cited by: 1
Related to: 10