Below are the top features of the leading commercial key management solutions:

  • Leading commercial key management solutions have dedicated hardware/software appliances for key storage that can be on the Cloud or on-premises. This key storage is only accessible by the customer, and allows the customer to inject the key into any CSP.
  • Commercial key management solutions are up to FIPS 140-2 Level 4 compliant and support symmetric and asymmetric keys. They also support AES – 128, 192, or 256 bit keys, RSA keys with  SHA-1, SHA-256, SHA-384, SHA-512, SSL3, Blake2b (256, 384, 512), or Blake2s-256 between 1024 and 8192 bits, DES keys of 56bits, 3DES keys of 168bits, and HMAC keys between 128 and 512 bits.
  • Commercial key management solutions are capable of key management, storage, and auditing, encryption, and tokenization.

About the Author

Search any posts

A collection of Encryption related products and resources that every organization should have!

Cyber security experts conference 2022

Free Downloads

Datasheet of Encryption Consulting Services

Encryption Consulting is a customer focused cybersecurity firm that provides a multitude of services in all aspects of encryption for our clients.

Download

Table of Contents

Companies in every sector must comply with standards and regulations, and one of the best ways to do this is to utilize encryption. Encryption takes data that can be clearly read, also known as plaintext, and runs it through an encryption algorithm. An encryption algorithm uses a key and mathematics to convert the plaintext into ciphertext, which is an undecipherable collection of letters and symbols. The process of encryption can be reversed using the same key, or the other key in a key pair, in a process called decryption. There are two different types of encryption: asymmetric and symmetric encryption.

Asymmetric vs Symmetric Encryption

Symmetric encryption involves the use of one key for both encryption and decryption. The plaintext is read into an encryption algorithm along with a key. The key works with the algorithm to turn the plaintext into ciphertext, thus encrypting the original sensitive data. This works well for data that is being stored and needs to be decrypted at a later date. The use of just one key for both encryption and decryption reveals an issue, as the compromise of the key would lead to a compromise of any data the key has encrypted. This also does not work for data-in-motion, which is where asymmetric encryption comes in.

Asymmetric encryption works with a pair of keys. The beginning of asymmetric encryption involves the creation of a pair of keys, one of which is a public key, and the other which is a private key. The public key is accessible by anyone, while the private key must be kept a secret from everyone but the creator of the key. This is because encryption occurs with the public key, while decryption occurs with the private key. The recipient of the sensitive data will provide the sender with their public key, which will be used to encrypt the data. This ensures that only the recipient can decrypt the data, with their own private key.

Uses for Asymmetric and Symmetric Encryption

Asymmetric and symmetric encryption are each better used for different situations. Symmetric encryption, with its use of a single key, is better used for data-at-rest. Data stored in databases needs to be encrypted to ensure it is not compromised or stolen. This data does not require two keys, just the one provided by symmetric encryption, as it only needs to be safe until it needs to be accessed in the future. Asymmetric encryption, on the other hand, should be used on data sent in emails to other people. If only symmetric encryption were used on data in emails, the attacker could take the key used for encryption and decryption and steal or compromise the data. With asymmetric encryption, the sender and recipient ensure only the recipient of the data can decrypt the data, because their public key was used to encrypt the data. Both types of encryption are used with other processes, like digital signing or compression, to provide even more security to the data.

Common Asymmetric and Symmetric Encryption Algorithms

Symmetric Encryption Algorithms:

Asymmetric Encryption Algorithms:

 

Comparison Table

 Asymmetric EncryptionSymmetric Encryption
DefinitionA two-way function that takes in plaintext data, and turns it into undecipherable ciphertext. This process utilizes a public key for encryption and a private key for decryption.A two-way function that takes in plaintext data, and turns it into undecipherable ciphertext. This process uses the same key for both encryption and decryption.
Use Cases
  • Digital Signing: Asymmetric encryption is much better for digital signing, compared to symmetric encryption. The use of both a public and private key means the identity of the signer of the data can easily be known. The signer uses their private key for encryption, while the recipient verifies their identity with their public key. As only the public key of the signer can decrypt data encrypted with the signer’s private key, the identity of the signer is verified when the data is decrypted.
  • Blockchain: Again, the identification of the user during cryptocurrency transactions is much easier done with asymmetric encryption.
  • Public Key Infrastructure (PKI): The identity of key owners is proven with certificates in PKI, and thus asymmetric encryption is the better choice in PKIs.
  • Banking: Encrypting sensitive customer data in banks is extremely important, as is decrypting that information as quickly as possible. For this reason, symmetric encryption is the preferred method of encryption in banks, as one key encryption is much swifter than two key encryption.
  • Data Storage: As with banking, data storage services and products tend to use symmetric encryption. This method is much swifter to encrypt and decrypt data needed in a timely manner.
Advantages
  • The loss of the public key does not result in the compromise of data
  • More secure than symmetric encryption
  • Only the owner of the private key can decrypt the data sent to them
  • Simpler to implement
  • Faster than asymmetric encryption
  • Protects data from compromise
Disadvantages
  • Slower than symmetric encryption
  • More complicated to implement than symmetric encryption
  • Loss of a key means any data encrypted with that key can be compromised
  • Less secure than asymmetric encryption
Common AlgorithmsECDSA, RSA, PGPAES, Blowfish, Twofish, RC4

About the Author

Search any posts

A collection of Encryption related products and resources that every organization should have!

Cyber security experts conference 2022

Free Downloads

Datasheet of Encryption Consulting Services

Encryption Consulting is a customer focused cybersecurity firm that provides a multitude of services in all aspects of encryption for our clients.

Download

Table of Contents

Threats loom ever present in our digital world, which is why methods of securing data are constantly advancing. Tactics like encryption are used every day on sensitive data-at-rest or in-motion. Encryption is the process of putting data in the form of plaintext into an encryption algorithm, and producing a ciphertext. Ciphertext is a form of data where all the patterns of letters that create words in the plaintext are scrambled into a new text that cannot be read without decrypting the data. Encryption uses a key to ensure the ciphertext cannot be deciphered by anyone but the authorized recipient.

Signing of data works to authenticate the sender of the data and tends to implement a form of encryption in its process. The process of signing emails, sensitive data, and other information has become necessary, as it verifies the identity of the sender and ensures the data has not been altered in transit. If a Man in the Middle attack occurred and the data was altered or compromised by the attacker, the recipient of the information would know that this has occurred. The attacker could alter the data, but as they do not have the key used by the sender to sign the data, the recipient of the data will know not to trust the sent data when analyzing the key and data.

How does digital signing work?

The process of digital signing works similarly to encryption. Encryption comes in two types, asymmetric and symmetric encryption. The process of asymmetric encryption works by creating a key pair with a public and private key. The private key is kept secret from everyone but the creator of the key, while the public key is available to everyone. The data is encrypted with the private key, and decrypted when needed with the public key. Symmetric encryption only uses one key for both encryption and decryption. As asymmetric encryption is more secure than symmetric encryption, it tends to be used more often. When sending data to a recipient, the correct method of encryption is to encrypt the data with the recipient’s public key, as this means only the owner of the key pair can decrypt that data.

Digital signing works oppositely. The data is signed by hashing the message with a hashing algorithm and the sender’s private key. This produces a hash digest, which can only be recreated through use of one of the keys in the key pair created by the sender. The recipient then receives the message, the hash digest, and the public key, if they did not already have it. The recipient then uses the sender’s public key to hash the message they have received. If the resulting hash digest matches the hash digest that has been sent along with the message, then the identity of the sender has been confirmed. This also confirms that the data has not been changed in transit. However, signing alone does not ensure the data has not been intercepted and read.

Encryption and Signing

To protect data from compromise and authenticate the sender at the same time, encryption and digital signing are used together. They are also both used in tandem to fulfill compliance standards for companies. Standards, like the Federal Information Processing Standards (FIPS) or the General Data Protection Regulation (GDPR), require companies to protect data as securely as possible along with authenticating data received from others. Encryption and digital signing ensures these standards are reached, and that users can be secure in the knowledge that data that is sent to and from them will not be compromised.

Confidential or sensitive data should always be encrypted and signed for its own safety. The use of encryption and signing together ensures that the main goals of cryptography, Confidentiality, Integrity, Authenticity, and Non-Repudiation are all met. Confidentiality and integrity are reached when data is encrypted asymmetrically, as only the intended recipient can decrypt the message. Non-repudiation and authenticity occur due to digital signing. Non-repudiation means that using the technique of digital signing, the sender of any information cannot, in the future, say they did not send the data, as the use of their private key confirms that they sent the data.

Common Encryption and Signing Algorithms

Symmetric Encryption Algorithms:

Asymmetric Encryption Algorithms:

Signing Algorithms:

  • RSA
  • ElGamal Encryption System
  • Digital Signing Algorithm (DSA)
  • ECDSA

About the Author

Search any posts

A collection of Encryption related products and resources that every organization should have!

Cyber security experts conference 2022

Free Downloads

Datasheet of Encryption Consulting Services

Encryption Consulting is a customer focused cybersecurity firm that provides a multitude of services in all aspects of encryption for our clients.

Download

Table of Contents

The Digital Age we live in requires more and more data to be sent at one time, but this can quickly become an issue for the hardware sending the data. If too much data is sent across a connection, it could take hours or days to transfer, when the recipient needs it in minutes. This is where compression comes in. Compression uses patterns in data to shorten the size of the data, saving bandwidth and storage space in the process. This works well, but what if the compressed data is stolen in transit? Compression is not complicated to reverse so retrieving the original would be easy. This is why many people use encryption and compression together when sending messages.

Encryption and Compression

Encryption can be either symmetric or asymmetric. Symmetric encryption is the less complicated of the two, using one key to encrypt and decrypt data. Symmetric encryption is much faster than asymmetric encryption, but is not as secure. Only one key needs to be compromised to compromise the original data. Asymmetric encryption uses a public and private key pair to encrypt plaintext data. The plaintext is run through the encryption algorithm along with a private key, which only the sender knows. The data is then sent to the recipient, who uses the public key to decrypt the data. This ensures the identity of the sender of the data, as the public key only decrypts data encrypted with the private key in that pair.

Using encryption and compression, larger portions of data can be sent across the Internet securely. If only compression is used, the security of the data is at risk. A Man in the Middle attack could retrieve the data, decompress it, and read the original message. On the other hand, if only encryption were used, there is a limit to how big a message that is encrypted can be. This is why the majority of messages are both encrypted and compressed. The next question to answer is do you compress the data and then encrypt the data, or vice versa?

Encryption or Compression first?

The choice of encrypting or compressing data first depends on the types of attacks you need to circumnavigate. Initially, it was thought that the only way to combine encryption and compression was to compress the data before encryption. This is due to the fact that compression relies on patterns in data to function, while encryption attempts to destroy these patterns in data. The method of compression then encryption has been found to be weak to side channel attacks. These types of attacks use JavaScript code to perform a brute force attack on the victim’s browser, allowing the attacker to infer what the transmitted data is based on the size of that compressed data.

Though this issue was found with the method of compression then encryption, encryption then compression has many more issues comparatively. This method is more difficult than the previous method, because compression needs patterns in data to work correctly. This means the value of compressing data will not exist, as there should be no patterns in encrypted data. Due to this, differential cryptanalysis can be used against this method to determine the sensitive data’s value. Compression then encryption is much more resilient to differential cryptanalysis.

Common Encryption and Compression Algorithms

Symmetric Encryption Algorithms:

Asymmetric Encryption Algorithms:

Hashing Algorithms:

  • Deflate
  • Huffman Coding
  • LZ77

About the Author

Search any posts

A collection of Encryption related products and resources that every organization should have!

Cyber security experts conference 2022

Free Downloads

Datasheet of Encryption Consulting Services

Encryption Consulting is a customer focused cybersecurity firm that provides a multitude of services in all aspects of encryption for our clients.

Download

Table of Contents

In the data security field, encryption and hashing are commonly compared, but why is this the case. Encryption is a two-way function where data is passed in as plaintext and comes out as ciphertext, which is unreadable. Since encryption is two-way, the data can be decrypted so it is readable again. Hashing, on the other hand, is one-way, meaning the plaintext is scrambled into a unique digest, through the use of a salt, that cannot be decrypted. Technically, hashing can be reversed, but the computational power needed to decrypt it makes decryption infeasible.

The way hashing works is with a hashing algorithm. This algorithm is most effective when it collision resistant. Collision resistance means that all the digests are unique and do not overlap with each other. This means that the hashing algorithm must be complex enough to not have overlapping hashes, but not so complex as to take too long to compute hashes. Encryption comes in two different types, and both encryption and hashing have several common types of algorithms.

Common Encryption and Hashing Algorithms

Encryption comes in two types: Asymmetric and Symmetric. Asymmetric encryption uses two different keys, a public and private key, for encryption and decryption. The private key is used to encrypt data, and is kept a secret from everyone but the person encrypting the data. The public key is available for anyone, and is used for decryption. Using asymmetric encryption, the authenticity of the data can be verified, because if the data was modified in transit, it would not be able to be re-encrypted with the private key. Symmetric encryption uses the same key for both encryption and decryption. This type of encryption uses less processing power and is faster, but is less secure as only one key is used.

Symmetric Encryption Algorithms:

Asymmetric Encryption Algorithms:

Hashing Algorithms:

  • Message Digest Algorithm (MD5)
  • Secure Hashing Algorithm (SHA-1, SHA-2, SHA-3)
  • WHIRLPOOL
  • TIGER
  • Cyclical Reduction Check (CRC32)

Hashing and Encryption Use Cases

Though they are similar, encryption and hashing are utilized for different purposes. One of the uses for hashing is to compare large amounts of data. Hash values are much easier to compare than large chunks of data, as they are more concise. Hashing is also used for mapping data, as finding values using hashes is quick, and good hashes do not overlap. Hashes are used in digital signatures and to create random strings to avoid duplication of data in databases too. As hashing is extremely infeasible to reverse, hashing algorithms are used on passwords. This makes the password shorter and undiscoverable by attackers.

Encryption, on the other hand, tends to be used for encrypting data that is in transit. Data being transmitted is data that needs to be read by the recipient only, thus it must be sent so that an attacker cannot read it. Encryption hides the data from anyone taking it in the middle of transit, and allows only the decryption key owner to read the data. Other times encryption would be used over hashing is for storing and retrieving data in databases, authentication methods, and other cases where data must be hidden at rest, but retrieved later.

 EncryptionHashing
DefinitionA two-way function that takes in plaintext data, and turns it into undecipherable ciphertext.A one-way method of hiding sensitive data. Using a hashing algorithm, hashing turns a plaintext into a unique hash digest that cannot be reverted to the original plaintext, without considerable effort.
Reversible or Irreversible?ReversibleIrreversible
Variable or Fixed Length Output?Variable LengthFixed Length
TypesAsymmetric and SymmetricHashing
Common AlgorithmsAES, RC4, DES, RSA, ECDSASHA-1, SHA-2, MD5, CRC32, WHIRLPOOL

About the Author

Search any posts

A collection of Encryption related products and resources that every organization should have!

Cyber security experts conference 2022

Free Downloads

Datasheet of Encryption Consulting Services

Encryption Consulting is a customer focused cybersecurity firm that provides a multitude of services in all aspects of encryption for our clients.

Download

Table of Contents

The Advanced Encryption Standard, or AES, is an encryption algorithm created by the National Institute of Science and Technology (NIST) in 2001. The cipher utilized in AES is a block cipher from the Rjindael cipher family. When AES was created, three different Rjindael block ciphers were selected for use, to make AES even more secure. All three ciphers used were 128 bits, but the keys they each used were of different sizes: 128, 192, and 256 bits. This is considered a symmetric block cipher, as only one key is used in the encryption process.

Symmetric encryption is a form of encryption that uses a single key for both encryption and decryption. Its counterpart, asymmetric encryption, uses two keys during the encryption and decryption process. One key is kept secret from everyone but the key’s creator, while the other key is a public key that can be viewed and utilized by anyone. Initially, AES was only used by the United States, but it has now been adopted worldwide as one of the most secure encryption algorithms.

Why was AES developed?

The Advanced Encryption Standard was created as a replacement for the Data Encryption Standard, or DES. DES was found to be increasingly more vulnerable to brute-force attackers, and thus needed to be phased out. AES’ original creation was to protect sensitive government information, but the security and ease of implementation provided by AES caused the majority of organizations to utilize AES in their encryption processes. Both public and private sector companies use AES now, as it protects against cyber-attacks, like brute force. AES does present an issue when exporting products encrypted with this encryption algorithm.

The Bureau of Industry and Security (BIS) has a number of controls and regulations in place that make it difficult export encryption products encrypted with AES. Commercial encryption products are required by the BIS to gain a license for their product that allows the organization to export their product to several destinations, without needing to acquire a separate license for each destination. Certain embargoed countries cannot receive commercial encryption products from the United States at all. These countries are: Cuba, Iran, Iraq, Libya, North Korea, Sudan and Syria.

Choosing the Rjindael cipher

To create the AES algorithm, a competition was held, which initially had 15 different encryption algorithms in the running. It was eventually narrowed down to just 5 algorithms:

  • MARS
  • Rivest Cipher 6 (RC6)
  • Rjindael
  • Serpent
  • Twofish 

These encryption algorithms were extensively analyzed by both the NIST and the National Security Agency (NSA), to determine the most secure one to use in the Advanced Encryption Standard. After rigorous testing of these algorithms, the Rjindael cipher was selected to be used in AES. The use of a 256 bit key gives the Rjindael cipher strong security, while maintaining its interoperability with existing hardware and software. Stronger ciphers exist, but they do not have the ability to be implemented into existing systems easily, like the Rjindael cipher can.

Understanding AES key size differences

The way a block cipher works is the plaintext of the data being encrypted is broken down into blocks of equal size, which for AES is 128 bits. Using a series of bitwise operations, the blocks of data are encrypted using keys of a specific length as well. AES allows 128, 192, and 256 bit keys for use, and the bigger the key size, the more secure the encryption. If a 128 bit key is used, the encryption on the block is done 10 times. With 192, the encryption is done 12 times, and with 256, 14 times. Thus, 256 bit keys are the most secure, but for most encryption cases, 128 bit keys are sufficient. The higher the security level of the data, however, the higher the size of the key should be.

To give an example on the security of AES, let’s take a look at how long it would take someone to crack one password encrypted with an AES-256 bit key. To break one 16-byte section of data encrypted with an AES-256 bit key, it would take centuries using a brute force method. The total amount of permutations that are possible with a 256 bit key are 2256, which makes cracking an AES-256 encrypted message virtually impossible. Even using a 128 bit key, the smallest size, there are still 2128 different permutations available, which would still take decades to brute force.

Attacks on AES

Researchers continually attempt to break AES with methods that are viable. The reason researchers are attempting to crack AES is to be one step ahead of attackers. If an attacker were to crack AES, and keep it a secret, then the world would continue to use AES believing it is completely secure. So far, a few different, theoretical attacks have been proposed, including:

  • Related-key attack: A related-key attack involves identifying how a cipher works under different keys. This cryptanalysis technique involves feeding a cipher, used to encrypt data, several different keys with the same plaintext. The process that occurs between the key and cipher can help identify a mathematical relationship between the cipher and key, thus helping identify the actual key’s value. This attack method is, however, not considered a big threat to AES, as it is useless as long as the protocols were implemented correctly.
  • Distinguished key attack: An attack that used a known key to find out the inner workings of an 8 round AES-128 algorithm was successfully used. As this was done on an 8 round algorithm, as opposed to the official 10 round algorithm, this is an attack that should not cause issues with any official AES algorithms.
  • Side channel attack: A side channel attack involves the leaking of information from an organization’s infrastructure. The data is leaked through locations, and the attacker listens at in to the sound, timing information, electromagnetic information or the power consumption in order to gather inferences from the algorithm which can then be used to break it. This can be stopped, however, by fixing the source of the leak or ensuring no pattern exists in the leaking information.
  • Key compromise: Though not a direct attack on the AES algorithm, the compromise of the key used for encryption cripples the entire AES algorithm. This is why proper key management and security are vital to the IT infrastructure of any organization.
  • Quantum computing: Quantum computing is the successor to classical computing, which we do now, that is still in the process of being created and understood. Though it has not been fully realized yet, the creation of quantum computers will make all classical computing cryptography irrelevant, as quantum computing could crack any classical cryptography algorithm in potentially seconds.

 

Who and what uses AES?

The majority of products, services, and organizations using symmetric encryption utilize AES. Most agencies and organizations in the United States government, including the NSA, use AES as well. The proven strength of AES and the inability to crack it mean the majority of companies looking for an encryption algorithm will use AES. A number of file transfer methods use AES for encryption as well. HTTPS is just one example of this.

About the Author

Search any posts

A collection of Encryption related products and resources that every organization should have!

Cyber security experts conference 2022

Free Downloads

Datasheet of Encryption Consulting Services

Encryption Consulting is a customer focused cybersecurity firm that provides a multitude of services in all aspects of encryption for our clients.

Download

Table of Contents

Cryptography is the study of securing communications from outside observers. Encryption algorithms take the original message, or plaintext, and converts it into ciphertext, which is not understandable. The key allows the user to decrypt the message, thus ensuring on they can read the message. The strength of the randomness of an encryption is also studied, which makes it harder for anyone to guess the key or input of the algorithm. Cryptography is how we can achieve more secure and robust connections to elevate our privacy. Advancements in cryptography makes it harder to break encryptions so that encrypted files, folders, or network connections are only accessible to authorized users.

Cryptography focuses on four different objectives:

  1. Confidentiality: Confidentiality ensures that only the intended recipient can decrypt the message and read its contents.
  2. Non-repudiation: Non-repudiation means the sender of the message cannot backtrack in the future and deny their reasons for sending or creating the message.
  3. Integrity: Integrity focuses on the ability to be certain that the information contained within the message cannot be modified while in storage or transit.
  4. Authenticity: Authenticity ensures the sender and recipient can verify each other’s identities and the destination of the message.

These objectives help ensure a secure and authentic transfer of information.

History of Cryptography

Cryptography began with ciphers, the first of which was the Caesar Cipher. Ciphers were a lot easier to unravel compared to modern cryptographic algorithms, but they both used keys and plaintext. Though simple, ciphers from the past were the earliest forms of encryption. Today’s algorithms and cryptosystems are much more advanced. They use multiple rounds of ciphers and encrypting the ciphertext of messages to ensure the most secure transit and storage of data. There are also methods of cryptography used now that are irreversible, maintaining the security of the message forever.

The reason for more advanced cryptography methods is due to the need for data to be protected more and more securely. Most of the ciphers and algorithms used in the early days of cryptography have been deciphered, making them useless for data protection. Today’s algorithms can be deciphered, but it would require years and sometimes decades to decipher the meaning of just one message. Thus, the race to create newer and more advanced cryptography techniques continues.


Types of Cryptography

Cryptography can be broken down into three different types:

  • Secret Key Cryptography
  • Public Key Cryptography
  • Hash Functions

Secret Key Cryptography, or symmetric cryptography, uses a single key to encrypt data. Both encryption and decryption in symmetric cryptography use the same key, making this the easiest form of cryptography. The cryptographic algorithm utilizes the key in a cipher to encrypt the data, and when the data must be accessed again, a person entrusted with the secret key can decrypt the data. Secret Key Cryptography can be used on both in-transit and at-rest data, but is commonly only used on at-rest data, as sending the secret to the recipient of the message can lead to compromise.

Examples:

  • AES
  • DES
  • Caesar Cipher

Public Key Cryptography, or asymmetric cryptography, uses two keys to encrypt data. One is used for encryption, while the other key can decrypts the message. Unlike symmetric cryptography, if one key is used to encrypt, that same key cannot decrypt the message, rather the other key shall be used.

One key is kept private, and is called the “private key”, while the other is shared publicly and can be used by anyone, hence it is known as the “public key”. The mathematical relation of the keys is such that the private key cannot be derived from the public key, but the public key can be derived from the private. The private key should not be distributed and should remain with the owner only. The public key can be given to any other entity.

Examples:

  • ECC
  • Diffie-Hellman
  • DSS

Hash functions are irreversible, one-way functions which protect the data, at the cost of not being able to recover the original message. Hashing is a way to transform a given string into a fixed length string. A good hashing algorithm will produce unique outputs for each input given. The only way to crack a hash is by trying every input possible, until you get the exact same hash. A hash can be used for hashing data (such as passwords) and in certificates.

Some of the most famous hashing algorithms are:

  • MD5
  • SHA-1
  • SHA-2 family which includes SHA-224, SHA-256, SHA-384, and SHA-512
  • SHA-3
  • Whirlpool
  • Blake 2
  • Blake 3

About the Author

Search any posts

A collection of Encryption related products and resources that every organization should have!

Cyber security experts conference 2022

Free Downloads

Datasheet of Encryption Consulting Services

Encryption Consulting is a customer focused cybersecurity firm that provides a multitude of services in all aspects of encryption for our clients.

Download

Let's talk