Adopt
Quantum-ready metadata
Updated 1 section on this page is new or updated since 1.2.0-rc.1 (2026-11-06). Everything that changed
Quantum computers able to break today's public-key cryptography do not exist yet, but data recorded now can be decrypted later. The NCSC asks organisations to plan now. Metadata helps in two ways: it can record the cryptography that protects data in transit, so catalogues become part of the inventory, and it can carry integrity checks that stay trustworthy for decades.
DCAT-AP-UK does not make data or services quantum-safe. It records what protects them, so owners can find and fix what is not. "Quantum-ready" here means ready for the migration, not finished with it.
The NCSC timelineUpdated
- By 2028: Define migration goals, complete discovery of where cryptography is used, and build an initial plan.
- By 2031: Complete the highest-priority migration activities.
- By 2035: Complete migration to post-quantum cryptography across all systems, services and products.
Source: Timelines for migration to post-quantum cryptographyS. The post-quantum standards: FIPS 203, 204 and 205V.
What is at risk
| Cryptography | The risk | What to do |
|---|---|---|
| Public-key key exchange (RSA, elliptic-curve Diffie-Hellman) | Broken by a large quantum computer. Data in transit recorded today could be decrypted later. | Move to ML-KEM (FIPS 203), first as a hybrid with a classical key exchange. Record it with connection security. |
| Digital signatures (RSA, ECDSA) | Broken by a large quantum computer. Signatures could be forged. | Plan for ML-DSA (FIPS 204) or SLH-DSA (FIPS 205) where metadata or data is signed. |
| Hash functions (SHA-2, SHA-3) | Weakened only modestly. SHA-256 stays acceptable for integrity checks. | Use SHA-384, SHA-512 or SHA3 for files that must be trusted for decades. Stop using MD5 and SHA-1, which are broken by classical means already. |
| Symmetric encryption (AES) | Weakened only modestly. | AES-256 for long-lived secrets. Not recorded in metadata. |
Checksums that stay trustworthy
Every checksum names its algorithm from this list. The validator warns about weak algorithms (rule Q-01) but still accepts them, so old records conform.
| Algorithm | Advice | Identifier | Why |
|---|---|---|---|
| SHA-256 | Acceptable | spdx:checksumAlgorithm_sha256 | Acceptable today and through the quantum transition for integrity checks. Prefer longer outputs for files that must be trusted for decades. |
| SHA-384 | Recommended | spdx:checksumAlgorithm_sha384 | Recommended for long-lived records. |
| SHA-512 | Recommended | spdx:checksumAlgorithm_sha512 | Recommended for long-lived records. |
| SHA3-256 | Acceptable | spdx:checksumAlgorithm_sha3_256 | A different design from SHA-2; acceptable, and useful for crypto-agility. |
| SHA3-384 | Recommended | spdx:checksumAlgorithm_sha3_384 | Recommended for long-lived records. |
| SHA3-512 | Recommended | spdx:checksumAlgorithm_sha3_512 | Recommended for long-lived records. |
| BLAKE2b-256 | Acceptable | spdx:checksumAlgorithm_blake2b256 | Acceptable. |
| BLAKE2b-512 | Acceptable | spdx:checksumAlgorithm_blake2b512 | Acceptable. |
| BLAKE3 | Acceptable | spdx:checksumAlgorithm_blake3 | Acceptable. |
| SHA-224 | Weak: warned | spdx:checksumAlgorithm_sha224 | Too short for new records. Use SHA-256 or longer. |
| SHA-1 | Weak: warned | spdx:checksumAlgorithm_sha1 | Broken against deliberate collisions. Kept so old records still conform; the validator warns (rule Q-01). |
| MD5 | Weak: warned | spdx:checksumAlgorithm_md5 | Broken. Kept so old records still conform; the validator warns (rule Q-01). |
The catalogue as a cryptographic inventory
Data services can record connection security: the key exchange their endpoint uses. The NCSC asks for discovery by 2028; a catalogue that records this answers “which of our APIs still use classical-only key exchange?” with one query.
| Value | Code | What it means |
|---|---|---|
| Hybrid post-quantum | ukcl:transport-security/pqc-hybrid | Negotiates a post-quantum key exchange (ML-KEM, FIPS 203) combined with a classical one, such as X25519MLKEM768 in TLS 1.3. |
| Post-quantum only | ukcl:transport-security/pqc-only | Negotiates a post-quantum key exchange alone. |
| TLS 1.3 with classical key exchange | ukcl:transport-security/tls13-classical | Secure against today's attackers, but data recorded now could be decrypted later by a quantum computer (harvest now, decrypt later). Plan the migration. |
| TLS 1.2 with classical key exchange | ukcl:transport-security/tls12-classical | Upgrade to TLS 1.3 first; post-quantum key exchange needs it. |
| Not yet assessed | ukcl:transport-security/not-assessed | Not yet checked. Checking is part of the discovery work the NCSC asks for by 2028. |
PREFIX dcat: <http://www.w3.org/ns/dcat#>
PREFIX dcatuk: <https://w3id.org/dcat-ap-uk/ns#>
PREFIX ukcl: <https://w3id.org/dcat-ap-uk/codelist/>
# Services still on classical-only key exchange, or not yet assessed
SELECT ?service ?endpoint WHERE {
?service a dcat:DataService ; dcat:endpointURL ?endpoint .
OPTIONAL { ?service dcatuk:transportSecurity ?ts }
FILTER (!BOUND(?ts) || ?ts NOT IN (ukcl:transport-security/pqc-hybrid, ukcl:transport-security/pqc-only))
}Crypto-agility in metadata
- Record algorithms as identifiers from a list, never as free text, so they can be found and changed.
- Keep the algorithm with every checksum. A checksum without its algorithm cannot be checked.
- Prefer standards that let algorithms change without changing the format.
- Treat the catalogue as part of the cryptographic inventory: services with classical-only key exchange are findable with one filter.