TLS 1.0 and 1.1, SHA-1 signatures and 3DES are formally deprecated, yet they survive in load balancers, mainframe links, payment terminals, appliances and partner integrations because nobody wants to be the one who breaks a connection. This template is a repeatable playbook for finding legacy cryptography, deciding what to do about each instance, removing it without an outage and handling the partners and systems that cannot move yet. It treats each discovery as a managed case with triggers, owners, containment, removal, rollback and review. It is written for security engineering, network and platform teams and the service owners they work with. For the wider programme of replacing algorithms over time, see the Crypto-Agility Guide.
How to use this template
- Replace text in square brackets, such as [Organisation], [Crypto Owner] and [Change Advisory Board], with your own names and thresholds.
- "Must" is mandatory; "should" is recommended practice where a documented exception is acceptable.
- Each legacy crypto finding is a case. Sections 1 to 8 describe how a case moves from discovery to closure. Implementation guidance covers the programme around it.
- Where legacy crypto is found during an active security incident, the incident response plan takes precedence.
- The standards cited reflect their status on 27 October 2026. NIST SP 800-52 Rev. 2 is under review and SP 800-131A Rev. 3 was released as a draft, so re-check both before each annual review.
1. Scope and triggers
1.1 This playbook covers protocols, cipher suites, signature and hash algorithms and key sizes used by [Organisation] systems, including servers, clients, appliances, managed services and connections to third parties.
1.2 The baseline the organisation retires against:
| Item | Status in the source | Source |
|---|---|---|
| TLS 1.0, TLS 1.1, DTLS 1.0 | Deprecated; "MUST NOT be used"; TLS 1.0 and 1.1 moved to Historic | RFC 8996 (BCP 195), March 2021 |
| SSL 2.0, SSL 3.0 | MUST NOT be negotiated | RFC 9325 (BCP 195), November 2022 |
| TLS 1.2 and 1.3 | Implementations MUST support TLS 1.2 and SHOULD support TLS 1.3, preferring it when supported. US government servers and clients must support TLS 1.2 with FIPS-based suites, and TLS 1.3 from 1 January 2024 | RFC 9325; NIST SP 800-52 Rev. 2 |
| NULL, RC4 and export cipher suites, anything under 112 bits | MUST NOT be negotiated | RFC 9325 |
| 3DES (TDEA) suites | Effective strength 112 bits, below the 128-bit level under which suites SHOULD NOT be negotiated. NIST withdrew SP 800-67 Rev. 2, its TDEA recommendation, on 1 January 2024 | RFC 9325; NIST SP 800-67 Rev. 2 page |
| Static RSA key transport, static DH and ECDH, CBC suites without encrypt-then-MAC | SHOULD NOT be negotiated | RFC 9325 |
| MD5 and SHA-1 in TLS 1.2 signatures | Clients MUST NOT offer them; servers MUST NOT use them in ServerKeyExchange | RFC 9155, December 2021 |
| SHA-1 generally | NIST will stop using SHA-1 in its last remaining specified protocols by 31 December 2030; modules using SHA-1 after 2030 will not be permitted for federal purchase | NIST announcement, December 2022 |
1.3 A case must be opened when any of the following is detected: a listener or client negotiating a baseline-prohibited protocol or suite; a certificate or code signature using SHA-1; a vendor advisory or scan finding for weak cryptography; a new supplier integration that requires legacy settings; or an audit finding.
2. Roles and contacts
| Role | Responsibility | Name and contact |
|---|---|---|
| [Crypto Owner] | Owns the baseline, this playbook and the exception register | [Name] |
| Service owner | Accepts the risk until fixed; approves the change window | From the CMDB |
| Platform engineer | Makes configuration changes on servers, load balancers, proxies and appliances | [Team] |
| Supplier or partner manager | Contacts external parties that connect with legacy settings | [Name] |
| PKI team | Reissues certificates and signing keys | [Team] |
| Identity team | Handles service accounts, API clients and machine certificates affected by the change | [Team] |
| [Change Advisory Board] | Approves production changes and rollbacks | [Contact] |
3. Detection and triage
3.1 Discovery sources must include:
- External and internal TLS scans of every listening service, recording protocol versions and suites offered, not only the one negotiated.
- Handshake telemetry: load balancer, reverse proxy and web server logs that record the negotiated protocol and cipher per connection. On Windows, Schannel logging is controlled by the
EventLoggingDWORD underHKLM\SYSTEM\CurrentControlSet\Control\SecurityProviders\SCHANNEL(values 0 to 7, reboot required). - Certificate inventory: signature algorithm and key size of every certificate, including internal CAs, device certificates and machine identities. See the Cryptographic Inventory and CBOM Guide.
- Code and configuration scanning for hard-coded protocol versions, cipher strings and hash functions.
- Supplier questionnaires for managed services and inbound integrations.
3.2 For each finding the triager must record: system and owner; direction (inbound, outbound, internal); which clients actually use the legacy setting, from logs over at least [30] days; whether the data or credentials carried are sensitive; whether the endpoint is internet-facing; and whether a modern alternative is already enabled alongside it.
3.3 Severity:
| Severity | Example | Target to fix |
|---|---|---|
| Critical | SSL 2.0/3.0, NULL, RC4 or export suites on an internet-facing service | [7] days |
| High | TLS 1.0/1.1 or 3DES on an internet-facing service or carrying credentials | [30] days |
| Medium | TLS 1.0/1.1 internally; SHA-1 certificates on internal services | [90] days |
| Low | Legacy option enabled but unused for [90] days | Next change window |
4. Containment
4.1 Containment limits exposure while the full fix is prepared. It must not break active, legitimate connections without service owner approval.
- Prefer order changes first: put TLS 1.3 and TLS 1.2 ECDHE with AES-GCM suites first so modern clients never negotiate legacy options. RFC 9325 recommends TLS_ECDHE_RSA_WITH_AES_128_GCM_SHA256, TLS_ECDHE_RSA_WITH_AES_256_GCM_SHA384 and the ECDSA equivalents for TLS 1.2.
- Restrict who can use legacy settings: move the legacy listener to a separate endpoint or port that only named partner addresses can reach.
- Remove unused options immediately: where logs show no client used a legacy protocol or suite in the observation window, disable it under a standard change.
- Protect credentials crossing legacy links: rotate service account passwords and API keys used over connections that ran with weak cryptography, and move them to stronger authentication where possible. See the Service Account Security Guide.
5. Eradication
5.1 Remove the legacy setting at every layer that can negotiate it: application, web server, load balancer, TLS-terminating proxy, operating system crypto library and appliance firmware.
5.2 Platform notes (examples, verify against current vendor documentation):
- Windows Schannel: protocol keys live under
HKLM\SYSTEM\CurrentControlSet\Control\SecurityProviders\SCHANNEL\Protocols\<protocol>\Client|Server, with the DWORDEnabledset to 0 to disable. Changes apply to new credential handles and may need a service restart. Microsoft recommends controlling cipher suites through cipher suite order (Group Policy, MDM or PowerShell) rather than undocumented registry values. - Linux: use system-wide crypto policies where the distribution supports them. See the Linux System-Wide Crypto Policies and FIPS Mode Guide.
- SSH: remove CBC and 3DES ciphers, SHA-1 key exchange and MD5 MACs. See the OpenSSH Server Hardening Guide.
- Storage and appliances: many have their own TLS and SSH settings, for example ONTAP's
security config modify -supported-protocols. See the NetApp ONTAP Security Hardening Guide. - Certificates: reissue SHA-1 signed certificates with SHA-256 or stronger, and retire SHA-1 intermediate and root CAs on a published schedule.
5.3 Fix the source of the setting, such as configuration management templates, container base images and infrastructure as code, so the legacy option does not return on the next build. Add a policy check to the pipeline.
6. Recovery
6.1 Every change must have a tested rollback: the previous configuration saved and the exact command or setting to restore it, approved by the [Change Advisory Board] in advance.
6.2 After the change, monitor handshake failures and application errors for [72] hours. Failures from a known partner trigger section 7, not an automatic rollback.
6.3 Roll back only when a business-critical connection fails and no compensating path exists. A rollback must open an exception (section 7.4) with an expiry date, and the case stays open.
6.4 Close the case when scans and [30] days of logs show no legacy negotiation and the exception register has no open item for the system.
7. Communication and notification
7.1 Internal: announce planned removals to service owners [30] days ahead, with the date, what will change and how to test.
7.2 External partners and customers: give at least [90] days' notice for inbound connections, with the new minimum (TLS 1.2 or later, approved suites), a test endpoint and a contact. Repeat at [30] and [7] days.
7.3 Where a supplier's service only offers legacy cryptography, the supplier manager must record it in third-party risk and request a remediation date.
7.4 Exceptions: every exception must name the system, the legacy setting, the business reason, compensating controls (network restriction, separate endpoint, monitoring), the owner and an expiry no later than [6] months. Exceptions are reviewed monthly by the [Crypto Owner].
8. Post-incident review
8.1 After each completed wave, or any outage caused by a removal, review: what was found, how long it had been present, what broke and why it was not found in testing, and which template or pipeline change prevents recurrence.
8.2 Update the baseline table in section 1.2 when the cited standards change, and record the date checked.
8.3 Feed lessons into the crypto-agility programme: every system that was hard to change is a candidate for configuration that can be updated centrally. See the Post-Quantum Cryptography Migration Guide, which reuses the same inventory.
Implementation guidance
Rollout waves
- Inventory (weeks 1 to 4): scan everything, collect handshake telemetry, build the certificate inventory, and tag owners.
- Quick wins: disable legacy options with zero observed use. This usually removes most findings with little risk.
- Internet-facing services: TLS 1.2 minimum, approved suites, then TLS 1.3 enabled.
- Internal services and clients: start with a pilot ring (one site or business unit), monitor for two weeks, then extend.
- Partners and legacy platforms: work through the notice periods in section 7; isolate what cannot move.
- Prevent regression: pipeline checks, configuration templates and periodic scans.
Decision guidance
| Situation | Recommended action |
|---|---|
| Legacy option enabled, no observed use | Disable under standard change |
| Used by internal clients that can be upgraded | Upgrade clients first, then disable server-side |
| Used by a partner who commits to a date | Separate restricted endpoint until the date, then disable |
| Used by an unsupported device or system | TLS-terminating proxy or gateway in front, network isolation, exception with replacement plan |
| Required by a regulated counterpart | Escalate; record the counterpart's own requirement and timeline |
Small organisations can usually complete this in one or two waves using managed scanners and provider defaults. Large estates should run it as a programme with monthly reporting.
Metrics
- Number of endpoints offering TLS 1.0 or 1.1, trended to zero.
- Percentage of connections negotiated with TLS 1.3.
- Number of certificates signed with SHA-1.
- Number of open exceptions and the oldest exception's age.
- Changes rolled back because of breakage, per wave.
Framework alignment
| Activity | NIST CSF 2.0 | NIST SP 800-53 Rev. 5 |
|---|---|---|
| Protect data in transit | PR.DS-02 | SC-8, SC-13 |
| Configuration baseline and least functionality | PR.PS-01 | CM-6, CM-7 |
| Inventory of cryptographic assets | ID.AM-02 | CM-8 |
| Supplier requirements | GV.SC-05 | SA-4 |
Standards and references
- IETF: RFC 8996, Deprecating TLS 1.0 and TLS 1.1
- IETF: RFC 9325, Recommendations for Secure Use of TLS and DTLS
- IETF: RFC 9155, Deprecating MD5 and SHA-1 Signature Hashes in TLS 1.2 and DTLS 1.2
- NIST: SP 800-52 Rev. 2, Guidelines for TLS Implementations
- NIST: SP 800-131A Rev. 2, Transitioning the Use of Cryptographic Algorithms and Key Lengths
- NIST: NIST Retires SHA-1 Cryptographic Algorithm
- NIST: SP 800-67 Rev. 2, TDEA (withdrawn)
- Microsoft: TLS registry settings
Related NHI Mgmt Group resources: Crypto-Agility Guide · TLS Server Configuration Baseline Guide · NIST SP 800-57 and 800-131A Cryptographic Key Management and Algorithm Transition Guide · How TLS Works Guide