Contents
StandardNIST

FIPS 202 (SHA-3)

SHA-3 Standard: Permutation-Based Hash and Extendable-Output Functions

FIPS 202 is NIST's public definition of the SHA-3 hash functions and the SHAKE extendable-output functions, which turn a message into a cryptographic fingerprint.

What it is

FIPS 202 is the NIST standard for SHA-3, a family of cryptographic hash functions, and for two extendable-output functions, SHAKE128 and SHAKE256. A hash function maps a message of any length to a short digest that is hard to invert or to fake. SHA-3 does this with the Keccak permutation inside a sponge: the message is absorbed into an internal state, then the digest is squeezed out.

SHA3-224, SHA3-256, SHA3-384, and SHA3-512 each have a fixed digest length. SHAKE128 and SHAKE256 let the caller choose how many bytes to squeeze. These are not renamed SHA-2 functions. SHA-2 uses a different construction, specified in a different NIST standard, and a SHA-3 digest of a message is not the SHA-2 digest of that message. Inside FIPS 202 the fixed hashes and the SHAKE functions are separated as well, so truncating a SHAKE output does not produce a SHA-3 digest.

Who it applies to

FIPS 202 applies when a product claims to compute SHA-3 or SHAKE. In a vehicle program the cybersecurity engineer puts the algorithm in the crypto concept, software and basic software invoke it, and the supplier of a library or hardware security module must match the standard byte for byte.

The standard does not require every ECU to use SHA-3. People who write the security concept still need to recognize the names so they do not treat SHA3-256 and SHA-256 as the same requirement.

What it covers

The document defines the permutation-based algorithms, how input is padded and absorbed, how output is produced, and the approved names of each function. It does not define vehicle protocols, certificate layouts, key storage, or where a hash is mandatory.

Function Output Role in an E/E design
SHA3-224, SHA3-256, SHA3-384, SHA3-512 Fixed length Fingerprint of a message, image, or configuration blob
SHAKE128, SHAKE256 Length chosen by the caller Digest, mask, or keying material sized by the protocol

A hash is not a message authentication code. Anyone who can see the input can recompute the digest. Keyed authentication, such as KMAC, is specified in a companion publication that uses SHAKE as a building block. If the goal is to stop forgery by someone without the key, naming a SHA-3 hash does not meet that goal.

How it fits with other documents

FIPS 180 specifies SHA-2 (and the older SHA-1). FIPS 202 does not redefine SHA-2. Whether an ECU uses SHA-2, SHA-3, or both is a design choice about interoperability and about what the crypto hardware accelerates.

NIST Special Publication 800-185 defines cSHAKE, KMAC, and related tuple and parallel functions on top of SHAKE. A requirement that says KMAC or cSHAKE still depends on FIPS 202 for the underlying sponge, and on SP 800-185 for the keyed or customizable wrapper. HMAC is yet another construction; an HMAC tag and a KMAC tag are not substitutes.

ISO/SAE 21434 and UN Regulation No. 155 ask for appropriate cryptographic controls. They do not name SHA-3. AUTOSAR crypto services are where application software requests a hash or an extendable output. Which names the stack implements depends on the crypto driver; the definition of SHA-3 and SHAKE remains FIPS 202.

In practice

Write the full function name and, for SHAKE, the output length into the requirement and the test plan. SHA3-256 and SHAKE256 differ, and two ends that squeeze a different number of SHAKE bytes will disagree even when both are conformant.

Use a library or hardware security module checked with known-answer tests. The sponge state is wide, and byte order is a common cause of a digest that matches nothing else. Record which driver command maps to which FIPS name, and repeat those tests after a library or secure-element firmware change.

On review, check that a shared field is not SHA-3 on one ECU and SHA-2 on another, that a forgery threat uses a MAC rather than a raw digest, and that any domain-separation string is specified for every sender and receiver. For software updates, store the digest algorithm and length with the image format so the backend, the update client, and the boot verifier compute the same fingerprint.

What it means for your discipline

  • CybersecurityCore

    You name SHA-3 or SHAKE in the crypto concept, including the output length and whether a keyed MAC is required instead of a bare hash. Reviews check that every ECU agrees on the function and that an integrity claim is not left as an unkeyed digest.

  • HPC & application softwarePartial

    You call the crypto API with the exact function name and, for SHAKE, the byte length the protocol demands. Mixing this up with SHA-256, or squeezing the wrong length, shows up as a failed check rather than a compile error.

  • Base softwarePartial

    Crypto service modules and drivers map application requests onto SHA-3 or SHAKE commands in the library or hardware security module. You keep those algorithm identifiers, and the known-answer tests, aligned with the firmware you integrate.

  • Hardware & siliconPartial

    A security module or accelerator must produce the same digest as a conformant implementation, including the byte order of the Keccak state. The command set has to distinguish the fixed SHA-3 hashes from SHAKE.

  • Software update & OTAPartial

    Image checks in an update path depend on a named digest and a fixed length. The backend, the update client, and the boot verifier have to use that same choice or a valid image will fail the check.

  • E/E architecturePartial

    You decide where SHA-3 is allowed beside SHA-2 so zones and suppliers do not mix families in one shared field. The choice follows interoperability and what the crypto hardware supports.

No direct stake: Functional safety · Networks & communication · Diagnostics & lifecycle