NIST SP 800-57 Part 1 Revision 5, "Recommendation for Key Management: Part 1 – General," defines 19 distinct cryptographic key types and the guidance for managing them through their lifecycle — the protection each type requires, cryptoperiod length, and algorithm/key-size selection. Per its own executive summary, "the proper management of cryptographic keys is essential to the effective use of cryptography for security. Poor key management may easily compromise strong algorithms." Authored by Elaine Barker, Revision 5 was published in May 2020, superseding Revision 4 of January 2016.
Key takeaways
- Part 1 is the one a device or firmware team works from day to day.
- Over a dozen factors inform the choice, including the embodiment of the mechanism (e.g., a FIPS 140 hardware module vs.
- Section 8 groups the key-management functions into four phases.
- The quantum caveat in Rev.
How is SP 800-57 split into three parts, and which part fits a product team?
Part 1 is the one a device or firmware team works from day to day. It covers the security services cryptography provides (Section 3), the approved algorithm classes (Section 4), the key types, their protection and their lifecycle.
Part 2 Revision 1, "Best Practices for Key Management Organizations" (May 2019), is written for U.S. government agencies: key management policies, practice statements and planning. One part of it is aimed squarely at product developers. Appendix C is a "Key Management Specification Checklist for Cryptographic Product Development", with yes/no questions such as whether the cryptographic module implements a software download capability for importing updated cryptographic functions, and whether its design precludes the use of any approved algorithm.
Part 3 Revision 1 (January 2015) gives application-specific guidance for PKI, IPsec, S/MIME, Kerberos, over-the-air rekeying (OTAR), DNSSEC, encrypted file systems and SSH. Its TLS section now only points to SP 800-52, where NIST moved its TLS guidance.
Which 19 key types does Part 1 define?
Section 5.1.1 classifies cryptographic keys into 19 distinct types, distinguished by whether they're public, private, or symmetric (secret), with key-agreement keys further split into static (long-term) or ephemeral (single-transaction) variants:
- Private signature key
- Public signature-verification key
- Symmetric authentication key
- Private authentication key
- Public authentication key
- Symmetric data-encryption key
- Symmetric key-wrapping key
- Symmetric random-number-generation key
- Symmetric master key/key-derivation key
- Private key-transport key
- Public key-transport key
- Symmetric key-agreement key
- Private static key-agreement key
- Public static key-agreement key
- Private ephemeral key-agreement key
- Public ephemeral key-agreement key
- Symmetric authorization key
- Private authorization key
- Public authorization key
Each type is defined by its use — a key-wrapping key "encrypts other keys." Section 6.1.1's Table 5 ties each of the 19 types to the specific protection requirements it needs (e.g., confidentiality, integrity, source/destination association), the reference point for how a key is stored, transmitted and sanitized; see our SP 800-88 media sanitization guide for cryptographic erase as a purge technique.
How long should each key live?
Section 5.3 defines a cryptoperiod as "the time span during which a specific key is authorized for use by legitimate entities or the keys for a given system will remain in effect." A suitable cryptoperiod limits the data available for cryptanalysis, the exposure from one key's compromise, and the time available to attack the mechanisms protecting a key.
Over a dozen factors inform the choice, including the embodiment of the mechanism (e.g., a FIPS 140 hardware module vs. software on a PC), the operating environment, data volume, the security life of the protected data, the number of nodes sharing a key, and "the threat to the information from new and disruptive technologies (e.g., quantum computers)." For asymmetric key pairs, each key gets its own cryptoperiod: the key that applies protection (e.g., generates a signature) has an originator-usage period (OUP), while the key that processes protected information (e.g., verifies that signature) has a recipient-usage period, which can extend beyond the OUP.
Section 5.3.6 closes with Table 1, the suggested cryptoperiods. An extract of the rows that matter most on devices, from the Part 1 Rev. 5 PDF:
| Key type (Table 1 row) | Originator-usage period | Recipient-usage period |
|---|---|---|
| 1. Private signature key | 1 to 3 years | — |
| 2. Public signature-verification key | Several years (depends on key size) | — |
| 4./5. Private and public authentication key | 1 to 2 years | — |
| 6. Symmetric data-encryption key | < 2 years | < OUP + 3 years |
| 7. Symmetric key-wrapping key | < 2 years | < OUP + 3 years |
| 9. Symmetric master key/key-derivation key | About 1 year | — |
| 13./14. Private and public static key-agreement key | 1 to 2 years | — |
| 15./16. Ephemeral key-agreement keys | One key-agreement transaction | — |
The table is a suggestion, not a ceiling: "Longer or shorter cryptoperiods may be warranted, depending on the application and environment," but "when assigning a longer cryptoperiod than suggested below, serious consideration should be given to the risks associated with doing so." And if a key is compromised, "its cryptoperiod shall no longer be considered valid."
Which states and phases does a key pass through?
Section 7 defines six states: pre-activation (generated, awaiting authorization for use), active (authorized to apply and/or process cryptographic protection), suspended (temporarily not to be used, e.g., pending investigation of a suspected compromise), deactivated (no longer used to apply new protection, though it may still process previously protected information), compromised, and destroyed.
Section 8 groups the key-management functions into four phases. In the pre-operational phase keys are in the pre-activation state (registration, initialization, key establishment). In the operational phase keys are active or suspended (normal storage, backup, key change). In the post-operational phase keys are deactivated or compromised and may sit in an archive. In the destroyed phase the keys are gone, though metadata such as the key name, type and cryptoperiod may be retained. Section 8 is explicit that "a key shall not be able to transfer back to any previous phase", so a key pulled from service is never quietly reactivated.
Section 8.2.3 separates two ways of replacing a key. Re-keying generates the new key independently of the old one and is the method for a compromised key. Key update derives the new key from the old one, and the publication warns that this "could result in a security exposure if an adversary obtains a key in the chain of derived keys."
Which algorithm and key sizes are still adequate?
Section 5.6.1 gives estimated maximum security strengths for the levels of 80, 112, 128, 192 and 256 bits and states that "a security strength of 80 bits is no longer considered adequate." Table 2 maps each level to comparable algorithms and key sizes:
| Security strength | Symmetric algorithm | FFC (DSA/DH/MQV) | IFC / RSA (k) | ECC (f) |
|---|---|---|---|---|
| ≤ 80 bits | 2TDEA | L=1024, N=160 | k=1024 | f=160–223 |
| 112 bits | 3TDEA (deprecated through 2023) | L=2048, N=224 | k=2048 | f=224–255 |
| 128 bits | AES-128 | L=3072, N=256 | k=3072 | f=256–383 |
| 192 bits | AES-192 | L=7680, N=384 | k=7680 | f=384–511 |
| 256 bits | AES-256 | L=15360, N=512 | k=15360 | f=512+ |
Footnote 68 says 3TDEA's use "has been deprecated (see SP 800-131A) through 2023, after which it will be disallowed for applying cryptographic protection." Table 4 then sets the time frames: 112-bit strength is acceptable for applying protection through 2030 and disallowed from 2031, with processing already-protected data allowed as legacy use. 128 bits and above stay acceptable.
The quantum caveat in Rev. 5 now has dates. NIST IR 8547 (final, October 2026) replaces the 2031 cut-off for public-key schemes: NIST "will deprecate quantum-vulnerable public-key cryptography after 2030 and disallow its use under NIST standards and guidelines after 2035", covering RSA, ECDSA and EdDSA signatures and RSA/ECC key establishment. NIST's 112-bit symmetric standards "will be disallowed in 2030." The replacements are FIPS 203 (ML-KEM), FIPS 204 (ML-DSA) and FIPS 205 (SLH-DSA); see our FIPS 203, 204 and 205 guide.
How do the cryptoperiods map to keys inside a device?
Most connected products carry five or six of the 19 types. The mapping below is SoftWSP's reading of Table 1 for a typical embedded product, not a NIST table:
| Key in the product | SP 800-57 type | Table 1 suggestion | Where a rotation path has to exist |
|---|---|---|---|
| Firmware-signing key (in the build HSM) | Private signature key | OUP 1–3 years | Build pipeline; old images stay verifiable |
| Firmware-verification key or its hash (in ROM, fuses or the bootloader) | Public signature-verification key | Several years | Spare key slots or a signed key-update image |
| Device identity key behind a TLS client certificate | Private authentication key | 1–2 years | Certificate re-enrolment over the network |
| Storage or configuration encryption key | Symmetric data-encryption key | OUP < 2 years, recipient < OUP + 3 years | Re-encryption on update |
| Key that wraps keys in a secure element | Symmetric key-wrapping key | OUP < 2 years | Secure-element command set |
| Factory master key used to derive per-device keys | Symmetric master key/key-derivation key | About 1 year | Factory provisioning station |
| ECDHE session keys | Ephemeral key-agreement keys | One transaction | TLS stack, automatic |
Two Part 1 rules shape the hardware side. Section 8.2.1.1 says keying material stored in a cryptographic module "shall be consistent with Section 6.2.2 as well as with FIPS 140", which is the link to FIPS 140-3 module validation. And Table 7 says a private signature key should in general not be backed up, but that backup "may be warranted in some cases, such as a CA's private signing key". A firmware-signing key behaves like one: losing it means no further signed updates for the installed base.
A key compiled into every unit of a firmware image breaks the model entirely: it has no per-device identity, no cryptoperiod that can be enforced and no destroyed state. MITRE tracks it as CWE-321, Use of Hard-coded Cryptographic Key; the fix is covered in our hard-coded credentials guide.
Worked example: what does a key plan for one connected controller look like?
Synthetic example, for illustration only. A maker ships a building controller with a 12-year service life. It boots through a ROM bootloader that checks an ECDSA P-256 signature against a key hash in one of four fuse slots, and it talks to a cloud service with a TLS client certificate.
InventorySix keys: firmware-signing key, its verification-key hash, device identity key, configuration encryption key, a factory derivation key, ECDHE session keys. CryptoperiodsFirmware-signing key: 2-year originator-usage period, so up to six signing keys over the service life. The four fuse slots cover only four, so the plan adds a signed key-update image that can carry the next verification key, and the gap goes in the risk file. StrengthP-256 gives 128-bit classical strength (Table 2). IR 8547 deprecates ECDSA after 2030 and disallows it after 2035, inside the service life, so one fuse slot is reserved for a hash of a post-quantum or hybrid verification key. IdentityEach unit generates its own identity key on first boot. Certificates last 1 year and renew online; a unit offline at expiry falls back to a re-enrolment path. StatesThe build HSM logs every transition (pre-activation, active, deactivated, compromised, destroyed) with date and reason, which is the audit trail Sections 7 and 8 assume.
What should a device team do first?
Name the key type before choosing its protectionTable 5's protection requirements (and therefore storage, access-control and sanitization decisions) differ by type, so misidentifying a key-wrapping key as a data-encryption key, for instance, can lead to the wrong safeguard. Set the cryptoperiod from risk, using Table 1 as the starting pointSection 5.3 notes that cryptoperiods are sometimes defined by an arbitrary time period or data volume, and asks that the risk and consequences of exposure be weighed instead. Any period longer than Table 1 gets a written reason. Track asymmetric key pairs' two cryptoperiods separatelyA signature-verification key's recipient-usage period legitimately outlives its signing key's originator-usage period — conflating the two risks destroying a verification key while old signatures still need checking. Retire 80-bit and 3TDEA combinations now, and plan the 112-bit and quantum exits3TDEA is disallowed for applying protection; 112-bit strength is disallowed for applying protection from 2031 (Table 4); quantum-vulnerable public-key algorithms follow IR 8547's 2030/2035 dates. Record every state transitionThe six-state model is only useful as an audit trail if transitions are actually logged when they happen. Use Part 2 Appendix C as a design reviewAnswer its checklist for every product line and keep the answers with the technical file.
What are the common pitfalls?
- One key for the whole fleet. A symmetric key or private key shared by every unit turns one extracted device into a fleet-wide compromise (CWE-321).
- A single fused verification key. Without a second slot or a signed key-update path, the signing key can never be rotated, even after a compromise.
- Rotating the signing key without keeping the old verification key. Rollback or recovery images signed with the earlier key stop booting.
- Treating the recipient-usage period as unlimited. Table 1 caps it at OUP + 3 years for symmetric data-encryption and key-wrapping keys.
- Choosing RSA-2048 or P-256 for a product supported past 2035 without a post-quantum migration path.
- No inventory of crypto libraries. Key management fails as surely through an unpatched TLS library as through a weak key; a cryptographic bill of materials and a firmware SBOM show which versions ship. The weak cryptography fix guide covers the code side, and the secure boot and hardware root of trust guide covers the boot chain.
What does Revision 6 change?
NIST published the initial public draft of Part 1 Revision 6 on 5 December 2025; comments closed on 5 February 2026, and Revision 5 remains the final version. Per NIST's announcement, the draft adds Ascon (SP 800-232) and the quantum-resistant algorithms in FIPS 203, 204 and 205, discusses keys for key establishment and keys for storage separately, adds the PQC security categories, adds a section on keying material storage and mechanisms, and replaces the algorithm time frames with references to SP 800-131A.
Where does SoftWSP fit?
SoftWSP does not find, classify or rotate keys. The platform extracts a firmware SBOM to CycloneDX 1.6 JSON, matches the components it finds (including TLS and crypto libraries) against NVD, CISA KEV, FIRST EPSS and CISA CSAF ICS advisories, and drafts a cited report that an engineer reviews. That gives a key-management review an accurate list of the crypto code that ships in each image. The key inventory, cryptoperiods and state records stay with the device team; see Device & Firmware Security, and the OWASP firmware security testing methodology for where key extraction tests sit in a firmware assessment.
This guide summarises NIST SP 800-57 Parts 1 (Rev. 5, May 2020), 2 (Rev. 1, May 2019) and 3 (Rev. 1, January 2015), the Rev. 6 draft announcement and NIST IR 8547; it is not a substitute for the full texts and does not constitute compliance or legal advice.
Put this guide to work
Frequently asked questions
What is NIST SP 800-57, and how are its three parts divided?
"Recommendation for Key Management" — a NIST publication providing cryptographic key-management guidance, split into three parts. Per its own abstract: **Part 1** "provides general guidance and best practices for the management of cryptographic keying material," including the security services cryptography provides, the key types involved, the protection each requires, and the functions and issues involved in key management; **Part 2**, "Best Practices for Key Management Organizations," addresses policy and security-planning requirements, primarily for U.S. government agencies; and **Part 3**, "Application-Specific Key Management Guidance," addresses key-management issues for specific, currently available cryptographic implementations. Part 1 Revision 5, authored by Elaine Barker (NIST), was published in May 2020, superseding Revision 4 of January 2016.
What types of cryptographic keys does Part 1 define?
Section 5.1.1 defines **19 distinct key types**, classified as public, private, or symmetric (secret), with key-agreement keys further split into static or ephemeral. They include: private signature key and public signature-verification key; symmetric authentication key and private/public authentication key; symmetric data-encryption key; symmetric key-wrapping key; symmetric random-number-generation key; symmetric master key/key-derivation key; private/public key-transport key; symmetric key-agreement key; private/public static key-agreement key; private/public ephemeral key-agreement key; and symmetric/private/public authorization key. Each type is defined by its function (e.g., a key-wrapping key encrypts other keys; a key-transport key establishes a symmetric key via a public-key algorithm) — Section 6.1.1's Table 5 ties each type to the specific protection requirements (confidentiality, integrity, association, etc.) it needs.
What is a cryptoperiod, and what determines how long one should be?
Section 5.3 defines a cryptoperiod as "the time span during which a specific key is authorized for use by legitimate entities or the keys for a given system will remain in effect." A well-chosen cryptoperiod limits, among other things, the amount of data available for cryptanalysis, the exposure from a single key compromise, the window for an attacker to penetrate key-protection mechanisms, and the time available for computationally intensive cryptanalysis. The publication lists over a dozen factors affecting the right cryptoperiod length — the embodiment of the cryptographic mechanism, the operating environment, personnel turnover, data volume, the security life of the data, and the threat posed by both current adversaries and future technologies such as quantum computers among them — and notes that for asymmetric key pairs, the two keys of a pair have separate cryptoperiods: an **originator-usage period** (for the key that applies protection) and a **recipient-usage period** (for the key that processes protected information), which can extend beyond the originator-usage period.
What security strengths does NIST recommend, and which are no longer adequate?
Section 5.6.1 gives estimated maximum security strengths, in bits, for approved symmetric and asymmetric algorithms at commonly used levels — 80, 112, 128, 192 and 256 bits — stating plainly that "a security strength of 80 bits is no longer considered adequate." Table 2 maps each strength to comparable algorithms and key sizes: 2TDEA at ≤80 bits; 3TDEA at 112 bits (deprecated through 2023, per SP 800-131A, after which it is disallowed for applying new cryptographic protection); AES-128 at 128 bits; AES-192 at 192 bits; and AES-256 at 256 bits — with corresponding finite-field (FFC), integer-factorization (IFC/RSA) and elliptic-curve (ECC) key-size equivalents given for each row. The publication separately flags that large-scale quantum computers, once available, will threaten the approved public-key (asymmetric) algorithms specifically, and that NIST was mid-process on standardizing post-quantum replacements at the time of this revision's publication.
Sources
- NIST: SP 800-57 Part 1 Rev. 5, Recommendation for Key Management: Part 1 – General (May 2020) — publication page
- NIST SP 800-57 Part 1 Rev. 5 (Barker), full text PDF
- NIST: SP 800-57 Part 2 Rev. 1, Best Practices for Key Management Organizations (May 2019)
- NIST SP 800-57 Part 2 Rev. 1, full text PDF (Appendix C, Key Management Specification Checklist for Cryptographic Product Development)
- NIST: SP 800-57 Part 3 Rev. 1, Application-Specific Key Management Guidance (January 2015)
- NIST: SP 800-57 Part 1 Rev. 6, initial public draft (5 December 2025)
- NIST IR 8547, Transition to Post-Quantum Cryptography Standards (final, October 2026)
- MITRE CWE-321: Use of Hard-coded Cryptographic Key
About this guide
SoftWSP Firmware Security Desk
Covers the Firmware vulnerabilities & fixes cluster: 131 guides.

In this cluster
Questions about this page?
