]> Fastly Inc.

fde@00f.net

Individual Contributor

samuel-lucas6@pm.me

Crypto Forum Internet-Draft This document describes the AEGIS-128L, AEGIS-256, AEGIS-128X, and AEGIS-256X AES-based authenticated encryption algorithms designed for high-performance applications. The document is a product of the Crypto Forum Research Group (CFRG). It is not an IETF product and is not a standard. Discussion Venues Source for this draft and an issue tracker can be found at .

Introduction This document describes the AEGIS family of authenticated encryption with associated data (AEAD) algorithms , which were chosen as additional finalists for high-performance applications in the Competition for Authenticated Encryption: Security, Applicability, and Robustness (CAESAR). Whilst AEGIS-128 was selected as a winner for this use case, AEGIS-128L has a better security margin alongside improved performance and AEGIS-256 uses a 256-bit key . All variants of AEGIS are inverse-free and constructed from the AES encryption round function . This document specifies:

• AEGIS-128L, which has a 128-bit key, a 128-bit nonce, a 1024-bit state, a 128- or 256-bit authentication tag, and processes 256-bit input blocks.
• AEGIS-256, which has a 256-bit key, a 256-bit nonce, a 768-bit state, a 128- or 256-bit authentication tag, and processes 128-bit input blocks.
• AEGIS-128X, which is a mode based on AEGIS-128L, specialized for CPUs with large vector registers and vector AES instructions.
• AEGIS-256X, which is a mode based on AEGIS-256, specialized for CPUs with large vector registers and vector AES instructions.

The AEGIS cipher family offers performance that significantly exceeds that of AES-GCM with hardware support for parallelizable AES block encryption . Similarly, software implementations can also be faster, although to a lesser extent. Unlike with AES-GCM, nonces can be safely chosen at random with no practical limit when using AEGIS-256 and AEGIS-256X. AEGIS-128L and AEGIS-128X also allow for more messages to be safely encrypted when using random nonces. With some existing AEAD schemes, such as AES-GCM, an attacker can generate a ciphertext that successfully decrypts under multiple different keys (a partitioning oracle attack) . This ability to craft a (ciphertext, authentication tag) pair that verifies under multiple keys significantly reduces the number of required interactions with the oracle to perform an exhaustive search, making it practical if the key space is small. For example, with password-based encryption, an attacker can guess a large number of passwords at a time by recursively submitting such a ciphertext to an oracle, which speeds up a password search by reducing it to a binary search. In AEGIS, finding distinct (key, nonce) pairs that successfully decrypt a given (associated data, ciphertext, authentication tag) tuple is believed to have a complexity that depends on the tag size. A 128-bit tag provides 64-bit committing security, which is generally acceptable for interactive protocols. With a 256-bit tag, finding a collision becomes impractical. Unlike most other AES-based AEAD constructions, leaking a state does not leak the key or previous states. Finally, an AEGIS key is not required after the setup phase, and there is no key schedule. Thus, ephemeral keys can be erased from memory before any data has been encrypted or decrypted, mitigating cold boot attacks. Note that an earlier version of Hongjun Wu and Bart Preneel’s paper introducing AEGIS specified AEGIS-128L and AEGIS-256 sporting differences with regards to the computation of the authentication tag and the number of rounds in the Finalize() function. We follow the specification of , which can be found in the References section of this document.

Conventions and Definitions The key words “MUST”, “MUST NOT”, “REQUIRED”, “SHALL”, “SHALL NOT”, “SHOULD”, “SHOULD NOT”, “RECOMMENDED”, “NOT RECOMMENDED”, “MAY”, and “OPTIONAL” in this document are to be interpreted as described in BCP 14 when, and only when, they appear in all capitals, as shown here. Throughout this document, “byte” is used interchangeably with “octet” and refers to an 8-bit sequence. Primitives:

• {}: an empty bit array.
• |x|: the length of x in bits.
• a ^ b: the bitwise exclusive OR operation between a and b.
• a & b: the bitwise AND operation between a and b.
• a || b: the concatenation of a and b.
• a mod b: the remainder of the Euclidean division between a as the dividend and b as the divisor.
• LE64(x): the little-endian encoding of unsigned 64-bit integer x.
• ZeroPad(x, n): padding operation. Trailing zeros are concatenated to x until the total length is a multiple of n bits.
• Truncate(x, n): truncation operation. The first n bits of x are kept.
• Split(x, n): splitting operation. x is split into n-bit blocks, ignoring partial blocks.
• Tail(x, n): returns the last n bits of x.
• AESRound(in, rk): a single round of the AES encryption round function, which is the composition of the SubBytes, ShiftRows, MixColums and AddRoundKey transformations, as defined in section 5 of . Here, in is the 128-bit AES input state, and rk is the 128-bit round key.
• Repeat(n, F): n sequential evaluations of the function F.
• CtEq(a, b): compares a and b in constant-time, returning True for an exact match, False otherwise.

AEGIS internal functions:

• Update(M0, M1) or Update(M): the state update function.
• Init(key, nonce): the initialization function.
• Absorb(ai): the input block absorption function.
• Enc(xi): the input block encryption function.
• Dec(ci): the input block decryption function.
• DecPartial(cn): the input block decryption function for the last ciphertext bits when they do not fill an entire block.
• Finalize(ad_len_bits, msg_len_bits): the authentication tag generation function.

Input blocks are 256 bits for AEGIS-128L and 128 bits for AEGIS-256. AES blocks:

• Si: the i-th AES block of the current state.
• S'i: the i-th AES block of the next state.
• {Si, ...Sj}: the vector of the i-th AES block of the current state to the j-th block of the current state.
• C0: an AES block built from the following bytes in hexadecimal format: { 0x00, 0x01, 0x01, 0x02, 0x03, 0x05, 0x08, 0x0d, 0x15, 0x22, 0x37, 0x59, 0x90, 0xe9, 0x79, 0x62 }.
• C1: an AES block built from the following bytes in hexadecimal format: { 0xdb, 0x3d, 0x18, 0x55, 0x6d, 0xc2, 0x2f, 0xf1, 0x20, 0x11, 0x31, 0x42, 0x73, 0xb5, 0x28, 0xdd }.

AES blocks are always 128 bits in length. Input and output values:

• key: the encryption key (128 bits for AEGIS-128L, 256 bits for AEGIS-256).
• nonce: the public nonce (128 bits for AEGIS-128L, 256 bits for AEGIS-256).
• ad: the associated data.
• msg: the plaintext.
• ct: the ciphertext.
• tag: the authentication tag (128 or 256 bits).

The AEGIS-128L Algorithm AEGIS-128L has a 1024-bit state, made of eight 128-bit blocks {S0, ...S7}. The parameters for this algorithm, whose meaning is defined in are:

• K_LEN (key length) is 16 bytes (128 bits).
• P_MAX (maximum length of the plaintext) is 2⁶¹ bytes (2⁶⁴ bits).
• A_MAX (maximum length of the associated data) is 2⁶¹ bytes (2⁶⁴ bits).
• N_MIN (minimum nonce length) = N_MAX (maximum nonce length) = 16 bytes (128 bits).
• C_MAX (maximum ciphertext length) = P_MAX + tag length = 2⁶¹ + 16 or 32 bytes (2⁶⁴ + 128 or 256 bits).

Distinct associated data inputs, as described in shall be unambiguously encoded as a single input. It is up to the application to create a structure in the associated data input if needed.

Authenticated Encryption The Encrypt function encrypts a message and returns the ciphertext along with an authentication tag that verifies the authenticity of the message and associated data, if provided. Security:

• For a given key, the nonce MUST NOT be reused under any circumstances; doing so allows an attacker to recover the internal state.
• The key MUST be randomly chosen from a uniform distribution.

Inputs:

• msg: the message to be encrypted (length MUST be less than P_MAX).
• ad: the associated data to authenticate (length MUST be less than A_MAX).
• key: the encryption key.
• nonce: the public nonce.

Outputs:

• ct: the ciphertext.
• tag: the authentication tag.

Steps:

Authenticated Decryption The Decrypt function decrypts a ciphertext, verifies that the authentication tag is correct, and returns the message on success or an error if tag verification failed. Security:

• If tag verification fails, the decrypted message and wrong message authentication tag MUST NOT be given as output. The decrypted message MUST be overwritten with zeros.
• The comparison of the input tag with the expected_tag MUST be done in constant time.

Inputs:

• ct: the ciphertext to be decrypted (length MUST be less than C_MAX).
• tag: the authentication tag.
• ad: the associated data to authenticate (length MUST be less than A_MAX).
• key: the encryption key.
• nonce: the public nonce.

Outputs:

• Either the decrypted message msg or an error indicating that the authentication tag is invalid for the given inputs.

Steps:

The Init Function The Init function constructs the initial state {S0, ...S7} using the given key and nonce. Inputs:

• key: the encryption key.
• nonce: the public nonce.

Defines:

• {S0, ...S7}: the initial state.

Steps:

The Update Function The Update function is the core of the AEGIS-128L algorithm. It updates the state {S0, ...S7} using two 128-bit values. Inputs:

• M0: the first 128-bit block to be absorbed.
• M1: the second 128-bit block to be absorbed.

Modifies:

• {S0, ...S7}: the state.

Steps:

The Absorb Function The Absorb function absorbs a 256-bit input block ai into the state {S0, ...S7}. Inputs:

• ai: the 256-bit input block.

Steps:

The Enc Function The Enc function encrypts a 256-bit input block xi using the state {S0, ...S7}. Inputs:

• xi: the 256-bit input block.

Outputs:

• ci: the 256-bit encrypted block.

Steps:

The Dec Function The Dec function decrypts a 256-bit input block ci using the state {S0, ...S7}. Inputs:

• ci: the 256-bit encrypted block.

Outputs:

• xi: the 256-bit decrypted block.

Steps:

The DecPartial Function The DecPartial function decrypts the last ciphertext bits cn using the state {S0, ...S7} when they do not fill an entire block. Inputs:

• cn: the encrypted input.

Outputs:

• xn: the decryption of cn.

Steps:

The Finalize Function The Finalize function computes a 128- or 256-bit tag that authenticates the message and associated data. Inputs:

• ad_len_bits: the length of the associated data in bits.
• msg_len_bits: the length of the message in bits.

Outputs:

• tag: the authentication tag.

Steps:

The AEGIS-256 Algorithm AEGIS-256 has a 768-bit state, made of six 128-bit blocks {S0, ...S5}. The parameters for this algorithm, whose meaning is defined in are:

• K_LEN (key length) is 32 bytes (256 bits).
• P_MAX (maximum length of the plaintext) is 2⁶¹ bytes (2⁶⁴ bits).
• A_MAX (maximum length of the associated data) is 2⁶¹ bytes (2⁶⁴ bits).
• N_MIN (minimum nonce length) = N_MAX (maximum nonce length) = 32 bytes (256 bits).
• C_MAX (maximum ciphertext length) = P_MAX + tag length = 2⁶¹ + 16 or 32 bytes (2⁶⁴ + 128 or 256 bits).

Distinct associated data inputs, as described in shall be unambiguously encoded as a single input. It is up to the application to create a structure in the associated data input if needed.

Authenticated Encryption The Encrypt function encrypts a message and returns the ciphertext along with an authentication tag that verifies the authenticity of the message and associated data, if provided. Security:

• For a given key, the nonce MUST NOT be reused under any circumstances; doing so allows an attacker to recover the internal state.
• The key MUST be randomly chosen from a uniform distribution.

Inputs:

• msg: the message to be encrypted (length MUST be less than P_MAX).
• ad: the associated data to authenticate (length MUST be less than A_MAX).
• key: the encryption key.
• nonce: the public nonce.

Outputs:

• ct: the ciphertext.
• tag: the authentication tag.

Steps:

Authenticated Decryption The Decrypt function decrypts a ciphertext, verifies that the authentication tag is correct, and returns the message on success or an error if tag verification failed. Security:

• If tag verification fails, the decrypted message and wrong message authentication tag MUST NOT be given as output. The decrypted message MUST be overwritten with zeros.
• The comparison of the input tag with the expected_tag MUST be done in constant time.

Inputs:

• ct: the ciphertext to be decrypted (length MUST be less than C_MAX).
• tag: the authentication tag.
• ad: the associated data to authenticate (length MUST be less than A_MAX).
• key: the encryption key.
• nonce: the public nonce.

Outputs:

• Either the decrypted message msg or an error indicating that the authentication tag is invalid for the given inputs.

Steps:

The Init Function The Init function constructs the initial state {S0, ...S5} using the given key and nonce. Inputs:

• key: the encryption key.
• nonce: the public nonce.

Defines:

• {S0, ...S5}: the initial state.

Steps:

The Update Function The Update function is the core of the AEGIS-256 algorithm. It updates the state {S0, ...S5} using a 128-bit value. Inputs:

• msg: the block to be absorbed.

Modifies:

• {S0, ...S5}: the state.

Steps:

The Absorb Function The Absorb function absorbs a 128-bit input block ai into the state {S0, ...S5}. Inputs:

• ai: the input block.

Steps:

The Enc Function The Enc function encrypts a 128-bit input block xi using the state {S0, ...S5}. Inputs:

• xi: the input block.

Outputs:

• ci: the encrypted input block.

Steps:

The Dec Function The Dec function decrypts a 128-bit input block ci using the state {S0, ...S5}. Inputs:

• ci: the encrypted input block.

Outputs:

• xi: the decrypted block.

Steps:

The DecPartial Function The DecPartial function decrypts the last ciphertext bits cn using the state {S0, ...S5} when they do not fill an entire block. Inputs:

• cn: the encrypted input.

Outputs:

• xn: the decryption of cn.

Steps:

The Finalize Function The Finalize function computes a 128- or 256-bit tag that authenticates the message and associated data. Inputs:

• ad_len_bits: the length of the associated data in bits.
• msg_len_bits: the length of the message in bits.

Outputs:

• tag: the authentication tag.

Steps:

Parallel Modes Some CPUs, such as Intel and Intel-compatible CPUs with the VAES extensions, include instructions to efficiently apply the AES round function to a vector of AES blocks. AEGIS-128X and AEGIS-256X are optional, specialized modes designed to take advantage of these instructions. They share the same properties as the ciphers they are based on but can be significantly faster on these platforms, even for short messages. AEGIS-128X and AEGIS-256X are parallel evaluations of multiple AEGIS-128L and AEGIS-256 instances respectively, with distinct initial states. On CPUs with wide vector registers, different states can be stored in different 128-bit lanes of the same vector register, allowing parallel updates using vector instructions. The modes are parameterized by the parallelism degree. With 256-bit registers, 2 parallel operations can be applied to 128-bit AES blocks. With 512-bit registers, the number of instances can be raised to 4. The state of a parallel mode is represented as a vector of AEGIS-128L or AEGIS-256 states.

Additional Conventions and Definitions

• D: the degree of parallelism.
• R: the absorption and output rate of the mode. With AEGIS-128X, the rate is 2 * 128 * D bits. With AEGIS-256X, the rate is 128 * D bits.
• V[j,i]: the j-th AES block of the i-th state. i is in the [0..D) range. For AEGIS-128X, j is in the [0..8) range, while for AEGIS-256, j is in the [0..6) range.
• V'[j,i]: the j-th AES block of the next i-th state.
• ctx[i]: the i-th context separator. This is a 128-bit mask, made of a byte representing the state index, followed by a byte representing the highest index and 112 all-zero bits.
• Byte(x): the value x encoded as 8 bits.

Authenticated Encryption The Encrypt function of AEGIS-128X resembles that of AEGIS-128L, and similarly, the Encrypt function of AEGIS-256X mirrors that of AEGIS-256, but processes R-bit input blocks per update. Steps:

Authenticated Decryption The Decrypt function of AEGIS-128X resembles that of AEGIS-128L, and similarly, the Decrypt function of AEGIS-256X mirrors that of AEGIS-256, but processes R-bit input blocks per update. Steps:

AEGIS-128X

The Init Function The Init function initializes a vector of D AEGIS-128L states with the same key and nonce but a different context ctx[i]. The context is added to the state before every update. Steps:

The Update Function The AEGIS-128X Update function is similar to the AEGIS-128L Update function, but absorbs R (2 * 128 * D) bits at once. M0 and M1 are 128 * D bits instead of 128 bits but are split into 128-bit blocks, each of them updating a different AEGIS-128L state. Steps:

The Absorb Function The Absorb function is similar to the AEGIS-128L Absorb function, but absorbs R bits instead of 256 bits. Steps:

The Enc Function The Enc function is similar to the AEGIS-128L Enc function, but encrypts R bits instead of 256 bits. Steps:

The Dec Function The Dec function is similar to the AEGIS-128L Dec function, but decrypts R bits instead of 256 bits. Steps:

The DecPartial Function The DecPartial function is similar to the AEGIS-128L DecPartial function, but decrypts up to R bits instead of 256 bits. Steps:

The Finalize Function The Finalize function finalizes every AEGIS-128L instance and combines the resulting authentication tags using the bitwise exclusive OR operation. Steps:

AEGIS-256X

The Init Function The Init function initializes a vector of D AEGIS-256 states with the same key and nonce but a different context ctx[i]. The context is added to the state before every update. Steps:

The Update Function The AEGIS-256X Update function is similar to the AEGIS-256 Update function, but absorbs R (128 * D) bits at once. M is 128 * D bits instead of 128 bits and is split into 128-bit blocks, each of them updating a different AEGIS-256 state. Steps:

The Absorb Function The Absorb function is similar to the AEGIS-256 Absorb function, but absorbs R bits instead of 128 bits. Steps:

The Enc Function The Enc function is similar to the AEGIS-256 Enc function, but encrypts R bits instead of 128 bits. Steps:

The Dec Function The Dec function is similar to the AEGIS-256 Dec function, but decrypts R bits instead of 128 bits. Steps:

The DecPartial Function The DecPartial function is similar to the AEGIS-256 DecPartial function, but decrypts up to R bits instead of 128 bits. Steps:

The Finalize Function The Finalize function finalizes every AEGIS-256 instance and combines the resulting authentication tags using the bitwise exclusive OR operation. Steps:

Implementation Considerations AEGIS-128X and AEGIS-256X with a degree of 1 are identical to AEGIS-128L and AEGIS-256. This property can be used to reduce the code size of a generic implementation. In AEGIS-128X, V can be represented as eight 256-bit registers (when D = 2) or eight 512-bit registers (when D = 4). In AEGIS-256X, V can be represented as six 256-bit registers (when D = 2) or six 512-bit registers (when D = 4). With this representation, loops over 0..D in the above pseudocode can be replaced by vector instructions.

Operational Considerations The AEGIS parallel modes are specialized and can only improve performance on specific CPUs. The degrees of parallelism implementations are encouraged to support are 2 (for CPUs with 256-bit registers) and 4 (for CPUs with 512-bit registers). The resulting algorithms are called AEGIS-128X2, AEGIS-128X4, AEGIS-256X2, and AEGIS-256X4. The following table summarizes how many bits are processed in parallel (rate), the memory requirements (state size), and the minimum vector register sizes a CPU should support for optimal performance.

Algorithm	Rate (bits)	Optimal Register Size	State Size (bits)
AEGIS-128L	256	128 bits	1024
AEGIS-128X2	512	256 bits	2048
AEGIS-128X4	1024	512 bits	4096
AEGIS-256	128	128 bits	768
AEGIS-256X2	256	256 bits	1536
AEGIS-256X4	512	512 bits	3072

Note that architectures with smaller vector registers but with many registers and large pipelines may still benefit from the parallel modes. Protocols SHOULD opt for a parallel mode only when all the involved parties agree on a specific variant. AEGIS-128L and AEGIS-256 SHOULD remain the default choices. Implementations MAY choose not to include the parallel AEGIS modes.

Encoding (ct, tag) Tuples Applications MAY keep the ciphertext and the authentication tag in distinct structures or encode both as a single string. In the latter case, the tag MUST immediately follow the ciphertext:

AEGIS as a Stream Cipher All AEGIS variants can also be used as stream ciphers. The Stream function expands a key and an optional nonce into a variable-length, secure keystream. Inputs:

• len: the length of the keystream to generate in bits.
• key: the AEGIS key.
• nonce: the nonce. If unspecified, it is set to N_MAX zero bytes.

Outputs:

• stream: the keystream.

Steps: This is equivalent to encrypting a len all-zero bits message without associated data, and discarding the authentication tag. Instead of relying on the generic Encrypt function, implementations can skip the finalization step. After initialization, the Update function is called with constant parameters, allowing further optimizations.

Implementation Status This note is to be removed before publishing as an RFC. Multiple implementations of the schemes described in this document have been developed and verified for interoperability. A comprehensive list of known implementations and integrations can be found at , which includes reference implementations closely aligned with the pseudocode provided in this document.

Security Considerations

Usage Guidelines

Key and Nonce Selection All AEGIS variants MUST be used in a nonce-respecting setting: for a given key, a nonce MUST only be used once, even with different tag lengths. Failure to do so would immediately reveal the bitwise difference between two messages. Every key MUST be randomly chosen from a uniform distribution. The nonce MAY be public or predictable. It can be a counter, the output of a permutation, or a generator with a long period. With AEGIS-128L and AEGIS-128X, random nonces can safely encrypt up to 2⁴⁸ messages using the same key with negligible (~ 2^-33, to align with NIST guidelines) collision probability. With AEGIS-256 and AEGIS-256X, random nonces can be used with no practical limits.

Key Commitment An authentication tag may verify under multiple keys, nonces, or associated data, but AEGIS is assumed to be key committing in the receiver-binding game, preventing common attacks when used with low-entropy keys such as passwords. Finding distinct keys and/or nonces that successfully verify the same (ad, ct, tag) tuple is expected to require ~2⁶⁴ attempts with a 128-bit authentication tag and ~2¹²⁸ attempts with a 256-bit tag. AEGIS is fully committing in the restricted setting where an adversary cannot control the associated data. As shown in , with the ability to alter the associated data, it is possible to efficiently find multiple keys that will verify the same authenticated ciphertext. Protocols mandating a fully committing scheme without that restriction can provide the associated data as input to a cryptographic hash function and use the output as the ad parameter of the Encrypt and Decrypt functions. The selected hash function must ensure a minimum of 128-bit preimage resistance. An instance of such a function is SHA-256 . Alternatively, the associated data can be fed into a collision-resistant KDF, such as HKDF , via the info input to derive the key parameter. The ad parameter can then be left empty. Note that the salt input MUST NOT be used since large salts get hashed, which affects commitment. Furthermore, this requires values concatenated to form the info input to be unambiguously encoded, like by appending their lengths.

Multi-User Security AEGIS nonces match the size of the key. AEGIS-128L and AEGIS-128X feature 128-bit nonces, offering an extra 32 bits compared to the commonly used AEADs in IETF protocols. The AEGIS-256 and AEGIS-256X variants provide even larger nonces. With 192 random bits, 64 bits remain available to optionally encode additional information. In all these variants, unused nonce bits can encode a key identifier, enhancing multi-user security. If every key has a unique identifier, multi-target attacks don’t provide any advantage over single-target attacks.

Other Uses of AEGIS All variants can be used as a MAC by calling the Encrypt() function with the message as the ad and leaving msg empty, resulting in just a tag. However, they MUST NOT be used as a hash function; if the key is known, inputs generating state collisions can easily be crafted. Similarly, as opposed to hash-based MACs, tags MUST NOT be used for key derivation as there is no proof they are uniformly random.

Implementation Security If tag verification fails, the unverified plaintext and the computed message authentication tag MUST NOT be released. As shown in , even a partial leak of the plaintext without verification would facilitate chosen ciphertext attacks. The security of AEGIS against timing and physical attacks is limited by the implementation of the underlying AESRound() function. Failure to implement AESRound() in a fashion safe against timing and physical attacks, such as differential power analysis, timing analysis, or fault injection attacks, may lead to leakage of secret key material or state information. The exact mitigations required for timing and physical attacks also depend on the threat model in question. Regardless of the variant, the key and nonce are only required by the Init function; other functions only depend on the resulting state. Therefore, implementations can overwrite ephemeral keys with zeros right after the last Update call of the initialization function.

Security Guarantees AEGIS-256 offers 256-bit message security against plaintext and state recovery, whereas AEGIS-128L offers 128-bit security. Under the assumption that the secret key is unknown to the attacker, all AEGIS variants target 128-bit security against forgery attacks regardless of the tag size. Encrypting the same message with the same key and nonce but different associated data generates distinct ciphertexts that do not reveal any additional information about the message. AEGIS has been shown to have reforgeability resilience in . Without the ability to set the associated data, a successful forgery does not increase the probability of subsequent forgeries. AEGIS-128X and AEGIS-256X share the same security properties and requirements as AEGIS-128L and AEGIS-256 respectively. In particular, the security level and usage limits remain the same . AEGIS is considered secure against guess-and-determine attacks aimed at recovering the state from observed ciphertexts. This resilience extends to quantum adversaries in the Q1 model, wherein quantum attacks do not confer any practical advantage for decrypting previously recorded ciphertexts or achieving key recovery. Security analyses of AEGIS can be found in , , , , , , , , and .

IANA Considerations IANA has assigned the following identifiers in the AEAD Algorithms Registry: AEGIS entries in the AEAD Algorithms Registry

Algorithm Name	ID
AEAD_AEGIS128L	32
AEAD_AEGIS256	33

IANA is requested to update the references of these entries to refer to the final version of this document. IANA is also requested to register the following identifiers in the AEAD Algorithms Registry:

• AEAD_AEGIS128X2
• AEAD_AEGIS128X4
• AEAD_AEGIS256X2
• AEAD_AEGIS256X4

References Normative References National Institute of Standards and Technology

US Gaithersburg

In many standards track documents several words are used to signify the requirements in the specification. These words are often capitalized. This document defines these words as they should be interpreted in IETF documents. This document specifies an Internet Best Current Practices for the Internet Community, and requests discussion and suggestions for improvements. RFC 2119 specifies common key words that may be used in protocol specifications. This document aims to reduce the ambiguity by clarifying that only UPPERCASE usage of the key words have the defined special meanings. This document defines algorithms for Authenticated Encryption with Associated Data (AEAD), and defines a uniform interface and a registry for such algorithms. The interface and registry can be used as an application-independent set of cryptoalgorithm suites. This approach provides advantages in efficiency and security, and promotes the reuse of crypto implementations. [STANDARDS-TRACK] Federal Information Processing Standard, FIPS This document specifies a simple Hashed Message Authentication Code (HMAC)-based key derivation function (HKDF), which can be used as a building block in various protocols and applications. The key derivation function (KDF) is intended to support a wide range of applications and requirements, and is conservative in its use of cryptographic hash functions. This document is not an Internet Standards Track specification; it is published for informational purposes. Informative References Nanyang Technological University KU Leuven Université de Lorraine, CNRS, Inria, LORIA Université de Rennes, CNRS, Inria, IRISA Cryptology ePrint Archive, Paper 2023/1306 Fastly Inc. Cryptology ePrint Archive, Paper 2023/523 Graz University of Technology Graz University of Technology Graz University of Technology IACR Transactions on Symmetric Cryptology, 2019(4), pp. 348–368 Beuth Hochschule für Technik Berlin Bauhaus-Universität Weimar Bauhaus-Universität Weimar Bauhaus-Universität Weimar Cryptology ePrint Archive, Paper 2017/332 University of Hyogo University of Hyogo Cryptology ePrint Archive, Paper 2023/1495 State Key Laboratory of Cryptology State Key Laboratory of Information Security, Institute of Information Engineering, Chinese Academy of Sciences; School of Cyber Security, University of Chinese Academy of Sciences State Key Laboratory of Cryptology The Computer Journal, vol 65, 2022(8), pp. 2221–2230 Cornell Tech Cornell Tech Cornell Tech 30th USENIX Security Symposium (USENIX Security 21), pp. 195–212 East China Normal University; University of Hyogo University of Hyogo; National Institute of Information and Communications Technology; PRESTO, Japan Science and Technology Agency University of Applied Sciences and Arts Northwestern Switzerland University of Hyogo IACR Transactions on Symmetric Cryptology, 2021(2), pp. 104–139 ANSSI Selected Areas in Cryptography. SAC 2014. Lecture Notes in Computer Science, vol 8781, pp. 290–305 University of Hyogo University of Hyogo University of Hyogo University of Hyogo; National Institute of Information and Communications Technology IET Information Security, vol 17, 2023(3), pp. 458-467 EPFL EPFL Applied Cryptography and Network Security. ACNS 2018. Lecture Notes in Computer Science, vol 10892, pp. 476–494

Test Vectors

AESRound Test Vector

AEGIS-128L Test Vectors

Update Test Vector

Test Vector 1

Test Vector 2

Test Vector 3

Test Vector 4

Test Vector 5

Test Vector 6 This test MUST return a “verification failed” error.

Test Vector 7 This test MUST return a “verification failed” error.

Test Vector 8 This test MUST return a “verification failed” error.

Test Vector 9 This test MUST return a “verification failed” error.

AEGIS-256 Test Vectors

Update Test Vector

Test Vector 1

Test Vector 2

Test Vector 3

Test Vector 4

Test Vector 5

Test Vector 6 This test MUST return a “verification failed” error.

Test Vector 7 This test MUST return a “verification failed” error.

Test Vector 8 This test MUST return a “verification failed” error.

Test Vector 9 This test MUST return a “verification failed” error.

AEGIS-128X2 Test Vectors

Initial State After initialization:

Test Vector 1

Test Vector 2

AEGIS-128X4 Test Vectors

Initial State After initialization:

Test Vector 1

Test Vector 2

AEGIS-256X2 Test Vectors

Initial State After initialization:

Test Vector 1

Test Vector 2

AEGIS-256X4 Test Vectors

Initial State After initialization:

Test Vector 1

Test Vector 2

Acknowledgments The AEGIS authenticated encryption algorithm was invented by Hongjun Wu and Bart Preneel. The round function leverages the AES permutation invented by Joan Daemen and Vincent Rijmen. They also authored the Pelican MAC, which partly motivated the design of the AEGIS MAC. We would like to thank the following individuals for their contributions:

• Eric Lagergren and Daniel Bleichenbacher for catching a broken test vector and Daniel Bleichenbacher for many helpful suggestions.
• John Preuß Mattsson for his review of the draft, and for suggesting how AEGIS should be used in the context of DTLS and QUIC.
• Bart Mennink and Charlotte Lefevre as well as Takanori Isobe and Mostafizar Rahman for investigating the commitment security of the schemes specified in this document.