idnits 2.17.1 draft-ietf-uta-rfc7525bis-02.txt: -(1251): Line appears to be too long, but this could be caused by non-ascii characters in UTF-8 encoding Checking boilerplate required by RFC 5378 and the IETF Trust (see https://trustee.ietf.org/license-info): ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/1id-guidelines.txt: ---------------------------------------------------------------------------- == There are 6 instances of lines with non-ascii characters in the document. Checking nits according to https://www.ietf.org/id-info/checklist : ---------------------------------------------------------------------------- ** The document seems to lack an IANA Considerations section. (See Section 2.2 of https://www.ietf.org/id-info/checklist for how to handle the case when there are no actions for IANA.) -- The draft header indicates that this document obsoletes RFC7525, but the abstract doesn't seem to directly say this. It does mention RFC7525 though, so this could be OK. -- The draft header indicates that this document updates RFC6066, but the abstract doesn't seem to mention this, which it should. -- The draft header indicates that this document updates RFC5288, but the abstract doesn't seem to mention this, which it should. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year (Using the creation date from RFC5288, updated by this document, for RFC5378 checks: 2007-06-19) -- The document seems to lack a disclaimer for pre-RFC5378 work, but may have content which was first submitted before 10 November 2008. If you have contacted all the original authors and they are all willing to grant the BCP78 rights to the IETF Trust, then this is fine, and you can ignore this comment. If not, you may need to add the pre-RFC5378 disclaimer. (See the Legal Provisions document at https://trustee.ietf.org/license-info for more information.) -- The document date (28 August 2021) is 971 days in the past. Is this intentional? Checking references for intended status: Best Current Practice ---------------------------------------------------------------------------- (See RFCs 3967 and 4897 for information about using normative references to lower-maturity documents in RFCs) == Unused Reference: 'I-D.irtf-cfrg-aead-limits' is defined on line 1250, but no explicit reference was found in the text == Outdated reference: A later version (-19) exists of draft-ietf-httpbis-semantics-18 -- Possible downref: Normative reference to a draft: ref. 'I-D.ietf-httpbis-semantics' ** Obsolete normative reference: RFC 4492 (Obsoleted by RFC 8422) ** Downref: Normative reference to an Informational RFC: RFC 4949 ** Obsolete normative reference: RFC 5246 (Obsoleted by RFC 8446) ** Obsolete normative reference: RFC 6125 (Obsoleted by RFC 9525) ** Obsolete normative reference: RFC 6347 (Obsoleted by RFC 9147) ** Obsolete normative reference: RFC 7507 (Obsoleted by RFC 8996) ** Obsolete normative reference: RFC 8740 (Obsoleted by RFC 9113) == Outdated reference: A later version (-18) exists of draft-ietf-tls-esni-13 == Outdated reference: A later version (-08) exists of draft-irtf-cfrg-aead-limits-03 -- Obsolete informational reference (is this intentional?): RFC 2246 (Obsoleted by RFC 4346) -- Obsolete informational reference (is this intentional?): RFC 4346 (Obsoleted by RFC 5246) -- Obsolete informational reference (is this intentional?): RFC 4347 (Obsoleted by RFC 6347) -- Obsolete informational reference (is this intentional?): RFC 5077 (Obsoleted by RFC 8446) -- Obsolete informational reference (is this intentional?): RFC 6961 (Obsoleted by RFC 8446) -- Obsolete informational reference (is this intentional?): RFC 7525 (Obsoleted by RFC 9325) Summary: 8 errors (**), 0 flaws (~~), 6 warnings (==), 12 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 UTA Working Group Y. Sheffer 3 Internet-Draft Intuit 4 Obsoletes: 7525 (if approved) R. Holz 5 Updates: 5288, 6066 (if approved) University of Twente 6 Intended status: Best Current Practice P. Saint-Andre 7 Expires: 1 March 2022 Mozilla 8 T. Fossati 9 arm 10 28 August 2021 12 Recommendations for Secure Use of Transport Layer Security (TLS) and 13 Datagram Transport Layer Security (DTLS) 14 draft-ietf-uta-rfc7525bis-02 16 Abstract 18 Transport Layer Security (TLS) and Datagram Transport Layer Security 19 (DTLS) are widely used to protect data exchanged over application 20 protocols such as HTTP, SMTP, IMAP, POP, SIP, and XMPP. Over the 21 last few years, several serious attacks on TLS have emerged, 22 including attacks on its most commonly used cipher suites and their 23 modes of operation. This document provides recommendations for 24 improving the security of deployed services that use TLS and DTLS. 25 The recommendations are applicable to the majority of use cases. 27 This document was published as RFC 7525 when the industry was in the 28 midst of its transition to TLS 1.2. Years later this transition is 29 largely complete and TLS 1.3 is widely available. Given the new 30 environment, we believe new guidance is needed. 32 Status of This Memo 34 This Internet-Draft is submitted in full conformance with the 35 provisions of BCP 78 and BCP 79. 37 Internet-Drafts are working documents of the Internet Engineering 38 Task Force (IETF). Note that other groups may also distribute 39 working documents as Internet-Drafts. The list of current Internet- 40 Drafts is at https://datatracker.ietf.org/drafts/current/. 42 Internet-Drafts are draft documents valid for a maximum of six months 43 and may be updated, replaced, or obsoleted by other documents at any 44 time. It is inappropriate to use Internet-Drafts as reference 45 material or to cite them other than as "work in progress." 47 This Internet-Draft will expire on 1 March 2022. 49 Copyright Notice 51 Copyright (c) 2021 IETF Trust and the persons identified as the 52 document authors. All rights reserved. 54 This document is subject to BCP 78 and the IETF Trust's Legal 55 Provisions Relating to IETF Documents (https://trustee.ietf.org/ 56 license-info) in effect on the date of publication of this document. 57 Please review these documents carefully, as they describe your rights 58 and restrictions with respect to this document. Code Components 59 extracted from this document must include Simplified BSD License text 60 as described in Section 4.e of the Trust Legal Provisions and are 61 provided without warranty as described in the Simplified BSD License. 63 Table of Contents 65 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 3 66 2. Terminology . . . . . . . . . . . . . . . . . . . . . . . . . 4 67 3. General Recommendations . . . . . . . . . . . . . . . . . . . 5 68 3.1. Protocol Versions . . . . . . . . . . . . . . . . . . . . 5 69 3.1.1. SSL/TLS Protocol Versions . . . . . . . . . . . . . . 5 70 3.1.2. DTLS Protocol Versions . . . . . . . . . . . . . . . 6 71 3.1.3. Fallback to Lower Versions . . . . . . . . . . . . . 7 72 3.2. Strict TLS . . . . . . . . . . . . . . . . . . . . . . . 7 73 3.3. Compression . . . . . . . . . . . . . . . . . . . . . . . 8 74 3.4. TLS Session Resumption . . . . . . . . . . . . . . . . . 8 75 3.5. TLS Renegotiation . . . . . . . . . . . . . . . . . . . . 9 76 3.6. Post-Handshake Authentication . . . . . . . . . . . . . . 10 77 3.7. Server Name Indication . . . . . . . . . . . . . . . . . 10 78 3.8. Application-Layer Protocol Negotiation . . . . . . . . . 10 79 3.9. Zero Round Trip Time (0-RTT) Data in TLS 1.3 . . . . . . 11 80 4. Recommendations: Cipher Suites . . . . . . . . . . . . . . . 11 81 4.1. General Guidelines . . . . . . . . . . . . . . . . . . . 12 82 4.2. Recommended Cipher Suites . . . . . . . . . . . . . . . . 13 83 4.2.1. Implementation Details . . . . . . . . . . . . . . . 14 84 4.3. Cipher Suites for TLS 1.3 . . . . . . . . . . . . . . . . 14 85 4.4. Limits on Key Usage . . . . . . . . . . . . . . . . . . . 15 86 4.5. Public Key Length . . . . . . . . . . . . . . . . . . . . 15 87 4.6. Truncated HMAC . . . . . . . . . . . . . . . . . . . . . 16 88 5. Applicability Statement . . . . . . . . . . . . . . . . . . . 16 89 5.1. Security Services . . . . . . . . . . . . . . . . . . . . 17 90 5.2. Opportunistic Security . . . . . . . . . . . . . . . . . 18 91 6. Security Considerations . . . . . . . . . . . . . . . . . . . 18 92 6.1. Host Name Validation . . . . . . . . . . . . . . . . . . 18 93 6.2. AES-GCM . . . . . . . . . . . . . . . . . . . . . . . . . 19 94 6.2.1. Nonce Reuse in TLS 1.2 . . . . . . . . . . . . . . . 19 95 6.3. Forward Secrecy . . . . . . . . . . . . . . . . . . . . . 20 96 6.4. Diffie-Hellman Exponent Reuse . . . . . . . . . . . . . . 21 97 6.5. Certificate Revocation . . . . . . . . . . . . . . . . . 22 98 7. Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . 23 99 8. References . . . . . . . . . . . . . . . . . . . . . . . . . 23 100 8.1. Normative References . . . . . . . . . . . . . . . . . . 23 101 8.2. Informative References . . . . . . . . . . . . . . . . . 26 102 Appendix A. Differences from RFC 7525 . . . . . . . . . . . . . 32 103 Appendix B. Document History . . . . . . . . . . . . . . . . . . 33 104 B.1. draft-ietf-uta-rfc7525bis-02 . . . . . . . . . . . . . . 33 105 B.2. draft-ietf-uta-rfc7525bis-01 . . . . . . . . . . . . . . 33 106 B.3. draft-ietf-uta-rfc7525bis-00 . . . . . . . . . . . . . . 34 107 B.4. draft-sheffer-uta-rfc7525bis-00 . . . . . . . . . . . . . 34 108 B.5. draft-sheffer-uta-bcp195bis-00 . . . . . . . . . . . . . 34 109 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 34 111 1. Introduction 113 Transport Layer Security (TLS) [RFC5246] and Datagram Transport 114 Security Layer (DTLS) [RFC6347] are widely used to protect data 115 exchanged over application protocols such as HTTP, SMTP, IMAP, POP, 116 SIP, and XMPP. Over the years leading to 2015, several serious 117 attacks on TLS have emerged, including attacks on its most commonly 118 used cipher suites and their modes of operation. For instance, both 119 the AES-CBC [RFC3602] and RC4 [RFC7465] encryption algorithms, which 120 together have been the most widely deployed ciphers, have been 121 attacked in the context of TLS. A companion document [RFC7457] 122 provides detailed information about these attacks and will help the 123 reader understand the rationale behind the recommendations provided 124 here. 126 The TLS community reacted to these attacks in two ways: 128 * Detailed guidance was published on the use of TLS 1.2 and earlier 129 protocol versions. This guidance is included in the original 130 [RFC7525] and mostly retained in this revised version. 132 * A new protocol version was released, TLS 1.3 [RFC8446], which 133 largely mitigates or resolves these attacks. 135 Those who implement and deploy TLS and DTLS, in particular versions 136 1.2 or earlier of these protocols, need guidance on how TLS can be 137 used securely. This document provides guidance for deployed services 138 as well as for software implementations, assuming the implementer 139 expects his or her code to be deployed in environments defined in 140 Section 5. Concerning deployment, this document targets a wide 141 audience -- namely, all deployers who wish to add authentication (be 142 it one-way only or mutual), confidentiality, and data integrity 143 protection to their communications. 145 The recommendations herein take into consideration the security of 146 various mechanisms, their technical maturity and interoperability, 147 and their prevalence in implementations at the time of writing. 148 Unless it is explicitly called out that a recommendation applies to 149 TLS alone or to DTLS alone, each recommendation applies to both TLS 150 and DTLS. 152 This document attempts to minimize new guidance to TLS 1.2 153 implementations, and the overall approach is to encourage systems to 154 move to TLS 1.3. However this is not always practical. Newly 155 discovered attacks, as well as ecosystem changes, necessitated some 156 new requirements that apply to TLS 1.2 environments. Those are 157 summarized in Appendix A. 159 As noted, the TLS 1.3 specification resolves many of the 160 vulnerabilities listed in this document. A system that deploys TLS 161 1.3 should have fewer vulnerabilities than TLS 1.2 or below. This 162 document is being republished with this in mind, and with an explicit 163 goal to migrate most uses of TLS 1.2 into TLS 1.3. 165 These are minimum recommendations for the use of TLS in the vast 166 majority of implementation and deployment scenarios, with the 167 exception of unauthenticated TLS (see Section 5). Other 168 specifications that reference this document can have stricter 169 requirements related to one or more aspects of the protocol, based on 170 their particular circumstances (e.g., for use with a particular 171 application protocol); when that is the case, implementers are 172 advised to adhere to those stricter requirements. Furthermore, this 173 document provides a floor, not a ceiling, so stronger options are 174 always allowed (e.g., depending on differing evaluations of the 175 importance of cryptographic strength vs. computational load). 177 Community knowledge about the strength of various algorithms and 178 feasible attacks can change quickly, and experience shows that a Best 179 Current Practice (BCP) document about security is a point-in-time 180 statement. Readers are advised to seek out any errata or updates 181 that apply to this document. 183 2. Terminology 185 A number of security-related terms in this document are used in the 186 sense defined in [RFC4949]. 188 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 189 "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and 190 "OPTIONAL" in this document are to be interpreted as described in 191 BCP 14 [RFC2119] [RFC8174] when, and only when, they appear in all 192 capitals, as shown here. 194 3. General Recommendations 196 This section provides general recommendations on the secure use of 197 TLS. Recommendations related to cipher suites are discussed in the 198 following section. 200 3.1. Protocol Versions 202 3.1.1. SSL/TLS Protocol Versions 204 It is important both to stop using old, less secure versions of SSL/ 205 TLS and to start using modern, more secure versions; therefore, the 206 following are the recommendations concerning TLS/SSL protocol 207 versions: 209 * Implementations MUST NOT negotiate SSL version 2. 211 Rationale: Today, SSLv2 is considered insecure [RFC6176]. 213 * Implementations MUST NOT negotiate SSL version 3. 215 Rationale: SSLv3 [RFC6101] was an improvement over SSLv2 and 216 plugged some significant security holes but did not support strong 217 cipher suites. SSLv3 does not support TLS extensions, some of 218 which (e.g., renegotiation_info [RFC5746]) are security-critical. 219 In addition, with the emergence of the POODLE attack [POODLE], 220 SSLv3 is now widely recognized as fundamentally insecure. See 221 [DEP-SSLv3] for further details. 223 * Implementations MUST NOT negotiate TLS version 1.0 [RFC2246]. 225 Rationale: TLS 1.0 (published in 1999) does not support many 226 modern, strong cipher suites. In addition, TLS 1.0 lacks a per- 227 record Initialization Vector (IV) for CBC-based cipher suites and 228 does not warn against common padding errors. This and other 229 recommendations in this section are in line with [RFC8996]. 231 * Implementations MUST NOT negotiate TLS version 1.1 [RFC4346]. 233 Rationale: TLS 1.1 (published in 2006) is a security improvement 234 over TLS 1.0 but still does not support certain stronger cipher 235 suites. 237 NOTE: This recommendation has been changed from SHOULD NOT to MUST 238 NOT on the assumption that [I-D.ietf-tls-oldversions-deprecate] 239 will be published as an RFC before this document. 241 * Implementations MUST support TLS 1.2 [RFC5246] and MUST prefer to 242 negotiate TLS version 1.2 over earlier versions of TLS. 244 Rationale: Several stronger cipher suites are available only with 245 TLS 1.2 (published in 2008). In fact, the cipher suites 246 recommended by this document for TLS 1.2 (Section 4.2 below) are 247 only available in this version. 249 * Implementations SHOULD support TLS 1.3 [RFC8446] and if 250 implemented, MUST prefer to negotiate TLS 1.3 over earlier 251 versions of TLS. 253 Rationale: TLS 1.3 is a major overhaul to the protocol and 254 resolves many of the security issues with TLS 1.2. We note that 255 as long as TLS 1.2 is still allowed by a particular 256 implementation, even if it defaults to TLS 1.3, implementers MUST 257 still follow all the recommendations in this document. 259 * Implementations of "greenfield" protocols or deployments, where 260 there is no need to support legacy endpoints, SHOULD support TLS 261 1.3, with no negotiation of earlier versions. Similarly, we 262 RECOMMEND that new protocol designs that embed the TLS mechanisms 263 (such as QUIC has done [RFC9001]) include TLS 1.3. 265 Rationale: secure deployment of TLS 1.3 is significantly easier 266 and less error prone than the secure deployment of TLS 1.2. 268 This BCP applies to TLS 1.2, 1.3 and to earlier versions. It is not 269 safe for readers to assume that the recommendations in this BCP apply 270 to any future version of TLS. 272 3.1.2. DTLS Protocol Versions 274 DTLS, an adaptation of TLS for UDP datagrams, was introduced when TLS 275 1.1 was published. The following are the recommendations with 276 respect to DTLS: 278 * Implementations MUST NOT negotiate DTLS version 1.0 [RFC4347]. 280 Version 1.0 of DTLS correlates to version 1.1 of TLS (see above). 282 * Implementations MUST support and (unless a higher version is 283 available) MUST prefer to negotiate DTLS version 1.2 [RFC6347] 285 Version 1.2 of DTLS correlates to version 1.2 of TLS (see above). 286 (There is no version 1.1 of DTLS.) 288 * Implementations SHOULD support and, if available, MUST prefer to 289 negotiate DTLS version 1.3 as specified in [I-D.ietf-tls-dtls13]. 291 Version 1.3 of DTLS correlates to version 1.3 of TLS (see above). 293 3.1.3. Fallback to Lower Versions 295 Clients that "fall back" to lower versions of the protocol after the 296 server rejects higher versions of the protocol MUST NOT fall back to 297 SSLv3 or earlier. Implementations of TLS/DTLS 1.2 or earlier MUST 298 implement the Fallback SCSV mechanism [RFC7507] to prevent such 299 fallback being forced by an attacker. 301 Rationale: Some client implementations revert to lower versions of 302 TLS or even to SSLv3 if the server rejected higher versions of the 303 protocol. This fallback can be forced by a man-in-the-middle (MITM) 304 attacker. TLS 1.0 and SSLv3 are significantly less secure than TLS 305 1.2 but at least TLS 1.0 is still allowed by many web servers. As of 306 this writing, the Fallback SCSV solution is widely deployed and 307 proven as a robust solution to this problem. 309 3.2. Strict TLS 311 The following recommendations are provided to help prevent SSL 312 Stripping (an attack that is summarized in Section 2.1 of [RFC7457]): 314 * In cases where an application protocol allows implementations or 315 deployments a choice between strict TLS configuration and dynamic 316 upgrade from unencrypted to TLS-protected traffic (such as 317 STARTTLS), clients and servers SHOULD prefer strict TLS 318 configuration. 320 * Application protocols typically provide a way for the server to 321 offer TLS during an initial protocol exchange, and sometimes also 322 provide a way for the server to advertise support for TLS (e.g., 323 through a flag indicating that TLS is required); unfortunately, 324 these indications are sent before the communication channel is 325 encrypted. A client SHOULD attempt to negotiate TLS even if these 326 indications are not communicated by the server. 328 * HTTP client and server implementations MUST support the HTTP 329 Strict Transport Security (HSTS) header [RFC6797], in order to 330 allow Web servers to advertise that they are willing to accept 331 TLS-only clients. 333 * Web servers SHOULD use HSTS to indicate that they are willing to 334 accept TLS-only clients, unless they are deployed in such a way 335 that using HSTS would in fact weaken overall security (e.g., it 336 can be problematic to use HSTS with self-signed certificates, as 337 described in Section 11.3 of [RFC6797]). 339 Rationale: Combining unprotected and TLS-protected communication 340 opens the way to SSL Stripping and similar attacks, since an initial 341 part of the communication is not integrity protected and therefore 342 can be manipulated by an attacker whose goal is to keep the 343 communication in the clear. 345 3.3. Compression 347 In order to help prevent compression-related attacks (summarized in 348 Section 2.6 of [RFC7457]), when using TLS 1.2 implementations and 349 deployments SHOULD disable TLS-level compression (Section 6.2.2 of 350 [RFC5246]), unless the application protocol in question has been 351 shown not to be open to such attacks. Note: this recommendation 352 applies to TLS 1.2 only, because compression has been removed from 353 TLS 1.3. 355 Rationale: TLS compression has been subject to security attacks, such 356 as the CRIME attack. 358 Implementers should note that compression at higher protocol levels 359 can allow an active attacker to extract cleartext information from 360 the connection. The BREACH attack is one such case. These issues 361 can only be mitigated outside of TLS and are thus outside the scope 362 of this document. See Section 2.6 of [RFC7457] for further details. 364 3.4. TLS Session Resumption 366 Session resumption drastically reduces the number of TLS handshakes 367 and thus is an essential performance feature for most deployments. 369 Stateless session resumption with session tickets is a popular 370 strategy. For TLS 1.2, it is specified in [RFC5077]. For TLS 1.3, 371 an equivalent PSK-based mechanism is described in Section 4.6.1 of 372 [RFC8446]. When it is used, the resumption information MUST be 373 authenticated and encrypted to prevent modification or eavesdropping 374 by an attacker. Further recommendations apply to session tickets: 376 * A strong cipher suite MUST be used when encrypting the ticket (as 377 least as strong as the main TLS cipher suite). 379 * Ticket keys MUST be changed regularly, e.g., once every week, so 380 as not to negate the benefits of forward secrecy (see Section 6.3 381 for details on forward secrecy). 383 * For similar reasons, session ticket validity SHOULD be limited to 384 a reasonable duration (e.g., half as long as ticket key validity). 386 Rationale: session resumption is another kind of TLS handshake, and 387 therefore must be as secure as the initial handshake. This document 388 (Section 4) recommends the use of cipher suites that provide forward 389 secrecy, i.e. that prevent an attacker who gains momentary access to 390 the TLS endpoint (either client or server) and its secrets from 391 reading either past or future communication. The tickets must be 392 managed so as not to negate this security property. 394 TLS 1.3 provides the powerful option of forward secrecy even within a 395 long-lived connection that is periodically resumed. Section 2.2 of 396 [RFC8446] recommends that clients SHOULD send a "key_share" when 397 initiating session resumption. In order to gain forward secrecy, 398 this document recommends that server implementations SHOULD respond 399 with a "key_share", to complete an ECDHE exchange on each session 400 resumption. 402 TLS session resumption introduces potential privacy issues where the 403 server is able to track the client, in some cases indefinitely. See 404 [Sy2018] for more details. 406 3.5. TLS Renegotiation 408 Where handshake renegotiation is implemented, both clients and 409 servers MUST implement the renegotiation_info extension, as defined 410 in [RFC5746]. Note: this recommendation applies to TLS 1.2 only, 411 because renegotiation has been removed from TLS 1.3. 413 The most secure option for countering the Triple Handshake attack is 414 to refuse any change of certificates during renegotiation. In 415 addition, TLS clients SHOULD apply the same validation policy for all 416 certificates received over a connection. The [triple-handshake] 417 document suggests several other possible countermeasures, such as 418 binding the master secret to the full handshake (see [SESSION-HASH]) 419 and binding the abbreviated session resumption handshake to the 420 original full handshake. Although the latter two techniques are 421 still under development and thus do not qualify as current practices, 422 those who implement and deploy TLS are advised to watch for further 423 development of appropriate countermeasures. 425 3.6. Post-Handshake Authentication 427 Renegotiation in TLS 1.2 was replaced in TLS 1.3 by separate post- 428 handshake authentication and key update mechanisms. In the context 429 of protocols that multiplex requests over a single connection (such 430 as HTTP/2), post-handshake authentication has the same problems as 431 TLS 1.2 renegotiation. Multiplexed protocols SHOULD follow the 432 advice provided for HTTP/2 in [RFC8740]. 434 3.7. Server Name Indication 436 TLS implementations MUST support the Server Name Indication (SNI) 437 extension defined in Section 3 of [RFC6066] for those higher-level 438 protocols that would benefit from it, including HTTPS. However, the 439 actual use of SNI in particular circumstances is a matter of local 440 policy. Implementers are strongly encouraged to support TLS 441 Encrypted Client Hello (formerly called Encrypted SNI) once 442 [I-D.ietf-tls-esni] has been standardized. 444 Rationale: SNI supports deployment of multiple TLS-protected virtual 445 servers on a single address, and therefore enables fine-grained 446 security for these virtual servers, by allowing each one to have its 447 own certificate. However, SNI also leaks the target domain for a 448 given connection; this information leak will be plugged by use of TLS 449 Encrypted Client Hello. 451 In order to prevent the attacks described in [ALPACA], a server that 452 does not recognize the presented server name SHOULD NOT continue the 453 handshake and instead fail with a fatal-level 454 "unrecognized_name(112)" alert. Note that this recommendation 455 updates Section 3 of [RFC6066]: "If the server understood the 456 ClientHello extension but does not recognize the server name, the 457 server SHOULD take one of two actions: either abort the handshake by 458 sending a fatal-level "unrecognized_name(112)" alert or continue the 459 handshake." It is also RECOMMENDED that clients abort the handshake 460 if the server acknowledges the SNI hostname with a different hostname 461 than the one sent by the client. 463 3.8. Application-Layer Protocol Negotiation 465 TLS implementations (both client- and server-side) MUST support the 466 Application-Layer Protocol Negotiation (ALPN) extension [RFC7301]. 468 In order to prevent "cross-protocol" attacks resulting from failure 469 to ensure that a message intended for use in one protocol cannot be 470 mistaken for a message for use in another protocol, servers should 471 strictly enforce the behavior prescribed in Section 3.2 of [RFC7301]: 472 "In the event that the server supports no protocols that the client 473 advertises, then the server SHALL respond with a fatal 474 "no_application_protocol" alert." It is also RECOMMENDED that 475 clients abort the handshake if the server acknowledges the ALPN 476 extension, but does not select a protocol from the client list. 477 Failure to do so can result in attacks such those described in 478 [ALPACA]. 480 Protocol developers are strongly encouraged to register an ALPN 481 identifier for their protocols. This applies to new protocols, as 482 well as well-established protocols such as SMTP. 484 3.9. Zero Round Trip Time (0-RTT) Data in TLS 1.3 486 The 0-RTT early data feature is new in TLS 1.3. It provides improved 487 latency when TLS connections are resumed, at the potential cost of 488 security. As a result, it requires special attention from 489 implementers on both the server and the client side. Typically this 490 extends to both the TLS library as well as protocol layers above it. 492 For use in HTTP-over-TLS, readers are referred to [RFC8470] for 493 guidance. 495 For QUIC-on-TLS, refer to Sec. 9.2 of [RFC9001]. 497 For other protocols, generic guidance is given in Sec. 8 and 498 Appendix E.5 of [RFC8446]. Given the complexity, we RECOMMEND to 499 avoid this feature altogether unless an explicit specification exists 500 for the application protocol in question to clarify when 0-RTT is 501 appropriate and secure. This can take the form of an IETF RFC, a 502 non-IETF standard, or even documentation associated with a non- 503 standard protocol. 505 4. Recommendations: Cipher Suites 507 TLS and its implementations provide considerable flexibility in the 508 selection of cipher suites. Unfortunately, some available cipher 509 suites are insecure, some do not provide the targeted security 510 services, and some no longer provide enough security. Incorrectly 511 configuring a server leads to no or reduced security. This section 512 includes recommendations on the selection and negotiation of cipher 513 suites. 515 4.1. General Guidelines 517 Cryptographic algorithms weaken over time as cryptanalysis improves: 518 algorithms that were once considered strong become weak. Such 519 algorithms need to be phased out over time and replaced with more 520 secure cipher suites. This helps to ensure that the desired security 521 properties still hold. SSL/TLS has been in existence for almost 20 522 years and many of the cipher suites that have been recommended in 523 various versions of SSL/TLS are now considered weak or at least not 524 as strong as desired. Therefore, this section modernizes the 525 recommendations concerning cipher suite selection. 527 * Implementations MUST NOT negotiate the cipher suites with NULL 528 encryption. 530 Rationale: The NULL cipher suites do not encrypt traffic and so 531 provide no confidentiality services. Any entity in the network 532 with access to the connection can view the plaintext of contents 533 being exchanged by the client and server. 534 Nevertheless, this document does not discourage software from 535 implementing NULL cipher suites, since they can be useful for 536 testing and debugging. 538 * Implementations MUST NOT negotiate RC4 cipher suites. 540 Rationale: The RC4 stream cipher has a variety of cryptographic 541 weaknesses, as documented in [RFC7465]. Note that DTLS 542 specifically forbids the use of RC4 already. 544 * Implementations MUST NOT negotiate cipher suites offering less 545 than 112 bits of security, including so-called "export-level" 546 encryption (which provide 40 or 56 bits of security). 548 Rationale: Based on [RFC3766], at least 112 bits of security is 549 needed. 40-bit and 56-bit security are considered insecure today. 550 TLS 1.1 and 1.2 never negotiate 40-bit or 56-bit export ciphers. 552 * Implementations SHOULD NOT negotiate cipher suites that use 553 algorithms offering less than 128 bits of security. 555 Rationale: Cipher suites that offer between 112-bits and 128-bits 556 of security are not considered weak at this time; however, it is 557 expected that their useful lifespan is short enough to justify 558 supporting stronger cipher suites at this time. 128-bit ciphers 559 are expected to remain secure for at least several years, and 560 256-bit ciphers until the next fundamental technology 561 breakthrough. Note that, because of so-called "meet-in-the- 562 middle" attacks [Multiple-Encryption], some legacy cipher suites 563 (e.g., 168-bit 3DES) have an effective key length that is smaller 564 than their nominal key length (112 bits in the case of 3DES). 565 Such cipher suites should be evaluated according to their 566 effective key length. 568 * Implementations SHOULD NOT negotiate cipher suites based on RSA 569 key transport, a.k.a. "static RSA". 571 Rationale: These cipher suites, which have assigned values 572 starting with the string "TLS_RSA_WITH_*", have several drawbacks, 573 especially the fact that they do not support forward secrecy. 575 * Implementations MUST support and prefer to negotiate cipher suites 576 offering forward secrecy, such as those in the Ephemeral Diffie- 577 Hellman and Elliptic Curve Ephemeral Diffie-Hellman ("DHE" and 578 "ECDHE") families. 580 Rationale: Forward secrecy (sometimes called "perfect forward 581 secrecy") prevents the recovery of information that was encrypted 582 with older session keys, thus limiting the amount of time during 583 which attacks can be successful. See Section 6.3 for a detailed 584 discussion. 586 4.2. Recommended Cipher Suites 588 Given the foregoing considerations, implementation and deployment of 589 the following cipher suites is RECOMMENDED: 591 * TLS_DHE_RSA_WITH_AES_128_GCM_SHA256 593 * TLS_ECDHE_RSA_WITH_AES_128_GCM_SHA256 595 * TLS_DHE_RSA_WITH_AES_256_GCM_SHA384 597 * TLS_ECDHE_RSA_WITH_AES_256_GCM_SHA384 599 These cipher suites are supported only in TLS 1.2 and not in earlier 600 protocol versions, because they are authenticated encryption (AEAD) 601 algorithms [RFC5116]. 603 Typically, in order to prefer these suites, the order of suites needs 604 to be explicitly configured in server software. (See [BETTERCRYPTO] 605 for helpful deployment guidelines, but note that its recommendations 606 differ from the current document in some details.) It would be ideal 607 if server software implementations were to prefer these suites by 608 default. 610 Some devices have hardware support for AES-CCM but not AES-GCM, so 611 they are unable to follow the foregoing recommendations regarding 612 cipher suites. There are even devices that do not support public key 613 cryptography at all, but they are out of scope entirely. 615 4.2.1. Implementation Details 617 Clients SHOULD include TLS_ECDHE_RSA_WITH_AES_128_GCM_SHA256 as the 618 first proposal to any server, unless they have prior knowledge that 619 the server cannot respond to a TLS 1.2 client_hello message. 621 Servers MUST prefer this cipher suite over weaker cipher suites 622 whenever it is proposed, even if it is not the first proposal. 624 Clients are of course free to offer stronger cipher suites, e.g., 625 using AES-256; when they do, the server SHOULD prefer the stronger 626 cipher suite unless there are compelling reasons (e.g., seriously 627 degraded performance) to choose otherwise. 629 This document does not change the mandatory-to-implement TLS cipher 630 suite(s) prescribed by TLS. To maximize interoperability, RFC 5246 631 mandates implementation of the TLS_RSA_WITH_AES_128_CBC_SHA cipher 632 suite, which is significantly weaker than the cipher suites 633 recommended here. (The GCM mode does not suffer from the same 634 weakness, caused by the order of MAC-then-Encrypt in TLS 635 [Krawczyk2001], since it uses an AEAD mode of operation.) 636 Implementers should consider the interoperability gain against the 637 loss in security when deploying the TLS_RSA_WITH_AES_128_CBC_SHA 638 cipher suite. Other application protocols specify other cipher 639 suites as mandatory to implement (MTI). 641 Note that some profiles of TLS 1.2 use different cipher suites. For 642 example, [RFC6460] defines a profile that uses the 643 TLS_ECDHE_ECDSA_WITH_AES_128_GCM_SHA256 and 644 TLS_ECDHE_ECDSA_WITH_AES_256_GCM_SHA384 cipher suites. 646 [RFC4492] allows clients and servers to negotiate ECDH parameters 647 (curves). Both clients and servers SHOULD include the "Supported 648 Elliptic Curves" extension [RFC4492]. For interoperability, clients 649 and servers SHOULD support the NIST P-256 (secp256r1) curve 650 [RFC4492]. In addition, clients SHOULD send an ec_point_formats 651 extension with a single element, "uncompressed". 653 4.3. Cipher Suites for TLS 1.3 655 This document does not specify any cipher suites for TLS 1.3. 656 Readers are referred to Sec. 9.1 of [RFC8446] for cipher suite 657 recommendations. 659 4.4. Limits on Key Usage 661 All ciphers have an upper limit on the amount of traffic that can be 662 securely encrypted with any given key. In the case of AEAD cipher 663 suites, the limit is typically determined by the cipher's integrity 664 guarantees. When the amount of traffic for a particular connection 665 has reached the limit, an implementation SHOULD perform a new 666 handshake (or in TLS 1.3, a Key Update) to rotate the session key. 668 For all AES-GCM cipher suites recommended for TLS 1.2 in this 669 document, the limit for one connection is 2^24.5 full-size records 670 (about 24 million). This is the same number as for TLS 1.3 with the 671 equivalent cipher suites. 673 // TODO: refer to {{I-D.irtf-cfrg-aead-limits}} once it has added the 674 // derivation for TLS 1.2, which is different from TLS 1.3. 675 // Different derivation, same numbers. 677 For all TLS 1.3 cipher suites, readers are referred to Section 5.5 of 678 [RFC8446]. 680 4.5. Public Key Length 682 When using the cipher suites recommended in this document, two public 683 keys are normally used in the TLS handshake: one for the Diffie- 684 Hellman key agreement and one for server authentication. Where a 685 client certificate is used, a third public key is added. 687 With a key exchange based on modular exponential (MODP) Diffie- 688 Hellman groups ("DHE" cipher suites), DH key lengths of at least 2048 689 bits are REQUIRED. 691 Rationale: For various reasons, in practice, DH keys are typically 692 generated in lengths that are powers of two (e.g., 2^10 = 1024 bits, 693 2^11 = 2048 bits, 2^12 = 4096 bits). Because a DH key of 1228 bits 694 would be roughly equivalent to only an 80-bit symmetric key 695 [RFC3766], it is better to use keys longer than that for the "DHE" 696 family of cipher suites. A DH key of 1926 bits would be roughly 697 equivalent to a 100-bit symmetric key [RFC3766]. A DH key of 2048 698 bits (equivalent to a 112-bit symmetric key) is the minimum allowed 699 by the latest revision of [NIST.SP.800-56A], as of this writing (see 700 in particular Appendix D). 702 As noted in [RFC3766], correcting for the emergence of a TWIRL 703 machine would imply that 1024-bit DH keys yield about 65 bits of 704 equivalent strength and that a 2048-bit DH key would yield about 92 705 bits of equivalent strength. The Logjam attack [Logjam] further 706 demonstrates that 1024-bit Diffie Hellman parameters should be 707 avoided. 709 With regard to ECDH keys, implementers are referred to the IANA 710 "Supported Groups Registry" (former "EC Named Curve Registry"), 711 within the "Transport Layer Security (TLS) Parameters" registry 712 [IANA_TLS], and in particular to the "recommended" groups. Curves of 713 less than 224 bits MUST NOT be used. This recommendation is in-line 714 with the latest revision of [NIST.SP.800-56A]. 716 When using RSA, servers SHOULD authenticate using certificates with 717 at least a 2048-bit modulus for the public key. In addition, the use 718 of the SHA-256 hash algorithm is RECOMMENDED and SHA-1 or MD5 MUST 719 NOT be used (see [CAB-Baseline] for more details). Clients MUST 720 indicate to servers that they request SHA-256, by using the 721 "Signature Algorithms" extension defined in TLS 1.2. 723 4.6. Truncated HMAC 725 Implementations MUST NOT use the Truncated HMAC extension, defined in 726 Section 7 of [RFC6066]. 728 Rationale: the extension does not apply to the AEAD cipher suites 729 recommended above. However it does apply to most other TLS cipher 730 suites. Its use has been shown to be insecure in [PatersonRS11]. 732 5. Applicability Statement 734 The recommendations of this document primarily apply to the 735 implementation and deployment of application protocols that are most 736 commonly used with TLS and DTLS on the Internet today. Examples 737 include, but are not limited to: 739 * Web software and services that wish to protect HTTP traffic with 740 TLS. 742 * Email software and services that wish to protect IMAP, POP3, or 743 SMTP traffic with TLS. 745 * Instant-messaging software and services that wish to protect 746 Extensible Messaging and Presence Protocol (XMPP) or Internet 747 Relay Chat (IRC) traffic with TLS. 749 * Realtime media software and services that wish to protect Secure 750 Realtime Transport Protocol (SRTP) traffic with DTLS. 752 This document does not modify the implementation and deployment 753 recommendations (e.g., mandatory-to-implement cipher suites) 754 prescribed by existing application protocols that employ TLS or DTLS. 755 If the community that uses such an application protocol wishes to 756 modernize its usage of TLS or DTLS to be consistent with the best 757 practices recommended here, it needs to explicitly update the 758 existing application protocol definition (one example is [TLS-XMPP], 759 which updates [RFC6120]). 761 Designers of new application protocols developed through the Internet 762 Standards Process [RFC2026] are expected at minimum to conform to the 763 best practices recommended here, unless they provide documentation of 764 compelling reasons that would prevent such conformance (e.g., 765 widespread deployment on constrained devices that lack support for 766 the necessary algorithms). 768 5.1. Security Services 770 This document provides recommendations for an audience that wishes to 771 secure their communication with TLS to achieve the following: 773 * Confidentiality: all application-layer communication is encrypted 774 with the goal that no party should be able to decrypt it except 775 the intended receiver. 777 * Data integrity: any changes made to the communication in transit 778 are detectable by the receiver. 780 * Authentication: an endpoint of the TLS communication is 781 authenticated as the intended entity to communicate with. 783 With regard to authentication, TLS enables authentication of one or 784 both endpoints in the communication. In the context of opportunistic 785 security [RFC7435], TLS is sometimes used without authentication. As 786 discussed in Section 5.2, considerations for opportunistic security 787 are not in scope for this document. 789 If deployers deviate from the recommendations given in this document, 790 they need to be aware that they might lose access to one of the 791 foregoing security services. 793 This document applies only to environments where confidentiality is 794 required. It recommends algorithms and configuration options that 795 enforce secrecy of the data in transit. 797 This document also assumes that data integrity protection is always 798 one of the goals of a deployment. In cases where integrity is not 799 required, it does not make sense to employ TLS in the first place. 800 There are attacks against confidentiality-only protection that 801 utilize the lack of integrity to also break confidentiality (see, for 802 instance, [DegabrieleP07] in the context of IPsec). 804 This document addresses itself to application protocols that are most 805 commonly used on the Internet with TLS and DTLS. Typically, all 806 communication between TLS clients and TLS servers requires all three 807 of the above security services. This is particularly true where TLS 808 clients are user agents like Web browsers or email software. 810 This document does not address the rarer deployment scenarios where 811 one of the above three properties is not desired, such as the use 812 case described in Section 5.2 below. As another scenario where 813 confidentiality is not needed, consider a monitored network where the 814 authorities in charge of the respective traffic domain require full 815 access to unencrypted (plaintext) traffic, and where users 816 collaborate and send their traffic in the clear. 818 5.2. Opportunistic Security 820 There are several important scenarios in which the use of TLS is 821 optional, i.e., the client decides dynamically ("opportunistically") 822 whether to use TLS with a particular server or to connect in the 823 clear. This practice, often called "opportunistic security", is 824 described at length in [RFC7435] and is often motivated by a desire 825 for backward compatibility with legacy deployments. 827 In these scenarios, some of the recommendations in this document 828 might be too strict, since adhering to them could cause fallback to 829 cleartext, a worse outcome than using TLS with an outdated protocol 830 version or cipher suite. 832 6. Security Considerations 834 This entire document discusses the security practices directly 835 affecting applications using the TLS protocol. This section contains 836 broader security considerations related to technologies used in 837 conjunction with or by TLS. 839 6.1. Host Name Validation 841 Application authors should take note that some TLS implementations do 842 not validate host names. If the TLS implementation they are using 843 does not validate host names, authors might need to write their own 844 validation code or consider using a different TLS implementation. 846 It is noted that the requirements regarding host name validation 847 (and, in general, binding between the TLS layer and the protocol that 848 runs above it) vary between different protocols. For HTTPS, these 849 requirements are defined by Sections 4.3.3, 4.3.4 and 4.3.5 of 850 [I-D.ietf-httpbis-semantics]. 852 Readers are referred to [RFC6125] for further details regarding 853 generic host name validation in the TLS context. In addition, that 854 RFC contains a long list of example protocols, some of which 855 implement a policy very different from HTTPS. 857 If the host name is discovered indirectly and in an insecure manner 858 (e.g., by an insecure DNS query for an MX or SRV record), it SHOULD 859 NOT be used as a reference identifier [RFC6125] even when it matches 860 the presented certificate. This proviso does not apply if the host 861 name is discovered securely (for further discussion, see [DANE-SRV] 862 and [DANE-SMTP]). 864 Host name validation typically applies only to the leaf "end entity" 865 certificate. Naturally, in order to ensure proper authentication in 866 the context of the PKI, application clients need to verify the entire 867 certification path in accordance with [RFC5280] (see also [RFC6125]). 869 6.2. AES-GCM 871 Section 4.2 above recommends the use of the AES-GCM authenticated 872 encryption algorithm. Please refer to Section 11 of [RFC5246] for 873 general security considerations when using TLS 1.2, and to Section 6 874 of [RFC5288] for security considerations that apply specifically to 875 AES-GCM when used with TLS. 877 6.2.1. Nonce Reuse in TLS 1.2 879 The existence of deployed TLS stacks that mistakenly reuse the AES- 880 GCM nonce is documented in [Boeck2016], showing there is an actual 881 risk of AES-GCM getting implemented in an insecure way and thus 882 making TLS sessions that use an AES-GCM cipher suite vulnerable to 883 attacks such as [Joux2006]. (See [CVE] records: CVE-2016-0270, CVE- 884 2016-10213, CVE-2016-10212, CVE-2017-5933.) 886 While this problem has been fixed in TLS 1.3, which enforces a 887 deterministic method to generate nonces from record sequence numbers 888 and shared secrets for all of its AEAD cipher suites (including AES- 889 GCM), TLS 1.2 implementations could still choose their own 890 (potentially insecure) nonce generation methods. 892 It is therefore RECOMMENDED that TLS 1.2 implementations use the 893 64-bit sequence number to populate the "nonce_explicit" part of the 894 GCM nonce, as described in the first two paragraphs of Section 5.3 of 895 [RFC8446]. Note that this recommendation updates Section 3 of 896 [RFC5288]: "The nonce_explicit MAY be the 64-bit sequence number." 898 We note that at the time of writing there are no cipher suites 899 defined for nonce reuse resistant algorithms such as AES-GCM-SIV 900 [RFC8452]. 902 6.3. Forward Secrecy 904 Forward secrecy (also called "perfect forward secrecy" or "PFS" and 905 defined in [RFC4949]) is a defense against an attacker who records 906 encrypted conversations where the session keys are only encrypted 907 with the communicating parties' long-term keys. 909 Should the attacker be able to obtain these long-term keys at some 910 point later in time, the session keys and thus the entire 911 conversation could be decrypted. 913 In the context of TLS and DTLS, such compromise of long-term keys is 914 not entirely implausible. It can happen, for example, due to: 916 * A client or server being attacked by some other attack vector, and 917 the private key retrieved. 919 * A long-term key retrieved from a device that has been sold or 920 otherwise decommissioned without prior wiping. 922 * A long-term key used on a device as a default key [Heninger2012]. 924 * A key generated by a trusted third party like a CA, and later 925 retrieved from it either by extortion or compromise 926 [Soghoian2011]. 928 * A cryptographic break-through, or the use of asymmetric keys with 929 insufficient length [Kleinjung2010]. 931 * Social engineering attacks against system administrators. 933 * Collection of private keys from inadequately protected backups. 935 Forward secrecy ensures in such cases that it is not feasible for an 936 attacker to determine the session keys even if the attacker has 937 obtained the long-term keys some time after the conversation. It 938 also protects against an attacker who is in possession of the long- 939 term keys but remains passive during the conversation. 941 Forward secrecy is generally achieved by using the Diffie-Hellman 942 scheme to derive session keys. The Diffie-Hellman scheme has both 943 parties maintain private secrets and send parameters over the network 944 as modular powers over certain cyclic groups. The properties of the 945 so-called Discrete Logarithm Problem (DLP) allow the parties to 946 derive the session keys without an eavesdropper being able to do so. 947 There is currently no known attack against DLP if sufficiently large 948 parameters are chosen. A variant of the Diffie-Hellman scheme uses 949 Elliptic Curves instead of the originally proposed modular 950 arithmetic. 952 Unfortunately, many TLS/DTLS cipher suites were defined that do not 953 feature forward secrecy, e.g., TLS_RSA_WITH_AES_256_CBC_SHA256. This 954 document therefore advocates strict use of forward-secrecy-only 955 ciphers. 957 6.4. Diffie-Hellman Exponent Reuse 959 For performance reasons, many TLS implementations reuse Diffie- 960 Hellman and Elliptic Curve Diffie-Hellman exponents across multiple 961 connections. Such reuse can result in major security issues: 963 * If exponents are reused for too long (e.g., even more than a few 964 hours), an attacker who gains access to the host can decrypt 965 previous connections. In other words, exponent reuse negates the 966 effects of forward secrecy. 968 * TLS implementations that reuse exponents should test the DH public 969 key they receive for group membership, in order to avoid some 970 known attacks. These tests are not standardized in TLS at the 971 time of writing. See [RFC6989] for recipient tests required of 972 IKEv2 implementations that reuse DH exponents. 974 * Under certain conditions, the use of static DH keys, or of 975 ephemeral DH keys that are reused across multiple connections, can 976 lead to timing attacks (such as those described in [RACCOON]) on 977 the shared secrets used in Diffie-Hellman key exchange. 979 To address these concerns, TLS implementations SHOULD NOT use static 980 DH keys and SHOULD NOT reuse ephemeral DH keys across multiple 981 connections. 983 // TODO: revisit when draft-bartle-tls-deprecate-ffdhe becomes a TLS 984 // WG item, since it specifies MUST NOT rather than SHOULD NOT. 986 6.5. Certificate Revocation 988 The following considerations and recommendations represent the 989 current state of the art regarding certificate revocation, even 990 though no complete and efficient solution exists for the problem of 991 checking the revocation status of common public key certificates 992 [RFC5280]: 994 * Although Certificate Revocation Lists (CRLs) are the most widely 995 supported mechanism for distributing revocation information, they 996 have known scaling challenges that limit their usefulness (despite 997 workarounds such as partitioned CRLs and delta CRLs). 999 * Proprietary mechanisms that embed revocation lists in the Web 1000 browser's configuration database cannot scale beyond a small 1001 number of the most heavily used Web servers. 1003 * The On-Line Certification Status Protocol (OCSP) [RFC6960] 1004 presents both scaling and privacy issues. In addition, clients 1005 typically "soft-fail", meaning that they do not abort the TLS 1006 connection if the OCSP server does not respond. (However, this 1007 might be a workaround to avoid denial-of-service attacks if an 1008 OCSP responder is taken offline.) 1010 * The TLS Certificate Status Request extension (Section 8 of 1011 [RFC6066]), commonly called "OCSP stapling", resolves the 1012 operational issues with OCSP. However, it is still ineffective in 1013 the presence of a MITM attacker because the attacker can simply 1014 ignore the client's request for a stapled OCSP response. 1016 * OCSP stapling as defined in [RFC6066] does not extend to 1017 intermediate certificates used in a certificate chain. Although 1018 the Multiple Certificate Status extension [RFC6961] addresses this 1019 shortcoming, it is a recent addition without much deployment. 1021 * Both CRLs and OCSP depend on relatively reliable connectivity to 1022 the Internet, which might not be available to certain kinds of 1023 nodes (such as newly provisioned devices that need to establish a 1024 secure connection in order to boot up for the first time). 1026 With regard to common public key certificates, servers SHOULD support 1027 the following as a best practice given the current state of the art 1028 and as a foundation for a possible future solution: 1030 1. OCSP [RFC6960] 1031 2. Both the status_request extension defined in [RFC6066] and the 1032 status_request_v2 extension defined in [RFC6961] (This might 1033 enable interoperability with the widest range of clients.) 1035 3. The OCSP stapling extension defined in [RFC6961] 1037 The considerations in this section do not apply to scenarios where 1038 the DANE-TLSA resource record [RFC6698] is used to signal to a client 1039 which certificate a server considers valid and good to use for TLS 1040 connections. 1042 7. Acknowledgments 1044 The following acknowledgments are inherited from [RFC7525]. 1046 Thanks to RJ Atkinson, Uri Blumenthal, Viktor Dukhovni, Stephen 1047 Farrell, Daniel Kahn Gillmor, Paul Hoffman, Simon Josefsson, Watson 1048 Ladd, Orit Levin, Ilari Liusvaara, Johannes Merkle, Bodo Moeller, 1049 Yoav Nir, Massimiliano Pala, Kenny Paterson, Patrick Pelletier, Tom 1050 Ritter, Joe St. Sauver, Joe Salowey, Rich Salz, Brian Smith, Sean 1051 Turner, and Aaron Zauner for their feedback and suggested 1052 improvements. Thanks also to Brian Smith, who has provided a great 1053 resource in his "Proposal to Change the Default TLS Ciphersuites 1054 Offered by Browsers" [Smith2013]. Finally, thanks to all others who 1055 commented on the TLS, UTA, and other discussion lists but who are not 1056 mentioned here by name. 1058 Robert Sparks and Dave Waltermire provided helpful reviews on behalf 1059 of the General Area Review Team and the Security Directorate, 1060 respectively. 1062 During IESG review, Richard Barnes, Alissa Cooper, Spencer Dawkins, 1063 Stephen Farrell, Barry Leiba, Kathleen Moriarty, and Pete Resnick 1064 provided comments that led to further improvements. 1066 Ralph Holz gratefully acknowledges the support by Technische 1067 Universitaet Muenchen. 1069 The authors gratefully acknowledge the assistance of Leif Johansson 1070 and Orit Levin as the working group chairs and Pete Resnick as the 1071 sponsoring Area Director. 1073 8. References 1075 8.1. Normative References 1077 [I-D.ietf-httpbis-semantics] 1078 Fielding, R. T., Nottingham, M., and J. Reschke, "HTTP 1079 Semantics", Work in Progress, Internet-Draft, draft-ietf- 1080 httpbis-semantics-18, 18 August 2021, 1081 . 1084 [I-D.ietf-tls-dtls13] 1085 Rescorla, E., Tschofenig, H., and N. Modadugu, "The 1086 Datagram Transport Layer Security (DTLS) Protocol Version 1087 1.3", Work in Progress, Internet-Draft, draft-ietf-tls- 1088 dtls13-43, 30 April 2021, 1089 . 1092 [I-D.ietf-tls-oldversions-deprecate] 1093 Moriarty, K. and S. Farrell, "Deprecating TLS 1.0 and TLS 1094 1.1", Work in Progress, Internet-Draft, draft-ietf-tls- 1095 oldversions-deprecate-12, 21 January 2021, 1096 . 1099 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 1100 Requirement Levels", BCP 14, RFC 2119, 1101 DOI 10.17487/RFC2119, March 1997, 1102 . 1104 [RFC3766] Orman, H. and P. Hoffman, "Determining Strengths For 1105 Public Keys Used For Exchanging Symmetric Keys", BCP 86, 1106 RFC 3766, DOI 10.17487/RFC3766, April 2004, 1107 . 1109 [RFC4492] Blake-Wilson, S., Bolyard, N., Gupta, V., Hawk, C., and B. 1110 Moeller, "Elliptic Curve Cryptography (ECC) Cipher Suites 1111 for Transport Layer Security (TLS)", RFC 4492, 1112 DOI 10.17487/RFC4492, May 2006, 1113 . 1115 [RFC4949] Shirey, R., "Internet Security Glossary, Version 2", 1116 FYI 36, RFC 4949, DOI 10.17487/RFC4949, August 2007, 1117 . 1119 [RFC5246] Dierks, T. and E. Rescorla, "The Transport Layer Security 1120 (TLS) Protocol Version 1.2", RFC 5246, 1121 DOI 10.17487/RFC5246, August 2008, 1122 . 1124 [RFC5288] Salowey, J., Choudhury, A., and D. McGrew, "AES Galois 1125 Counter Mode (GCM) Cipher Suites for TLS", RFC 5288, 1126 DOI 10.17487/RFC5288, August 2008, 1127 . 1129 [RFC5746] Rescorla, E., Ray, M., Dispensa, S., and N. Oskov, 1130 "Transport Layer Security (TLS) Renegotiation Indication 1131 Extension", RFC 5746, DOI 10.17487/RFC5746, February 2010, 1132 . 1134 [RFC6066] Eastlake 3rd, D., "Transport Layer Security (TLS) 1135 Extensions: Extension Definitions", RFC 6066, 1136 DOI 10.17487/RFC6066, January 2011, 1137 . 1139 [RFC6125] Saint-Andre, P. and J. Hodges, "Representation and 1140 Verification of Domain-Based Application Service Identity 1141 within Internet Public Key Infrastructure Using X.509 1142 (PKIX) Certificates in the Context of Transport Layer 1143 Security (TLS)", RFC 6125, DOI 10.17487/RFC6125, March 1144 2011, . 1146 [RFC6176] Turner, S. and T. Polk, "Prohibiting Secure Sockets Layer 1147 (SSL) Version 2.0", RFC 6176, DOI 10.17487/RFC6176, March 1148 2011, . 1150 [RFC6347] Rescorla, E. and N. Modadugu, "Datagram Transport Layer 1151 Security Version 1.2", RFC 6347, DOI 10.17487/RFC6347, 1152 January 2012, . 1154 [RFC7301] Friedl, S., Popov, A., Langley, A., and E. Stephan, 1155 "Transport Layer Security (TLS) Application-Layer Protocol 1156 Negotiation Extension", RFC 7301, DOI 10.17487/RFC7301, 1157 July 2014, . 1159 [RFC7465] Popov, A., "Prohibiting RC4 Cipher Suites", RFC 7465, 1160 DOI 10.17487/RFC7465, February 2015, 1161 . 1163 [RFC7507] Moeller, B. and A. Langley, "TLS Fallback Signaling Cipher 1164 Suite Value (SCSV) for Preventing Protocol Downgrade 1165 Attacks", RFC 7507, DOI 10.17487/RFC7507, April 2015, 1166 . 1168 [RFC8174] Leiba, B., "Ambiguity of Uppercase vs Lowercase in RFC 1169 2119 Key Words", BCP 14, RFC 8174, DOI 10.17487/RFC8174, 1170 May 2017, . 1172 [RFC8446] Rescorla, E., "The Transport Layer Security (TLS) Protocol 1173 Version 1.3", RFC 8446, DOI 10.17487/RFC8446, August 2018, 1174 . 1176 [RFC8740] Benjamin, D., "Using TLS 1.3 with HTTP/2", RFC 8740, 1177 DOI 10.17487/RFC8740, February 2020, 1178 . 1180 [RFC8996] Moriarty, K. and S. Farrell, "Deprecating TLS 1.0 and TLS 1181 1.1", BCP 195, RFC 8996, DOI 10.17487/RFC8996, March 2021, 1182 . 1184 8.2. Informative References 1186 [ALPACA] Brinkmann, M., Dresen, C., Merget, R., Poddebniak, D., 1187 Müller, J., Somorovsky, J., Schwenk, J., and S. Schinzel, 1188 "ALPACA: Application Layer Protocol Confusion - Analyzing 1189 and Mitigating Cracks in TLS Authentication", 30th USENIX 1190 Security Symposium (USENIX Security 21) , 2021, 1191 . 1194 [BETTERCRYPTO] 1195 bettercrypto.org, "Applied Crypto Hardening", April 2015, 1196 . 1198 [Boeck2016] 1199 Böck, H., Zauner, A., Devlin, S., Somorovsky, J., and P. 1200 Jovanovic, "Nonce-Disrespecting Adversaries: Practical 1201 Forgery Attacks on GCM in TLS", May 2016, 1202 . 1204 [CAB-Baseline] 1205 CA/Browser Forum, "Baseline Requirements for the Issuance 1206 and Management of Publicly-Trusted Certificates Version 1207 1.1.6", 2013, . 1209 [CVE] MITRE, "Common Vulnerabilities and Exposures", 1210 . 1212 [DANE-SMTP] 1213 Dukhovni, V. and W. Hardaker, "SMTP Security via 1214 Opportunistic DNS-Based Authentication of Named Entities 1215 (DANE) Transport Layer Security (TLS)", RFC 7672, 1216 DOI 10.17487/RFC7672, October 2015, 1217 . 1219 [DANE-SRV] Finch, T., Miller, M., and P. Saint-Andre, "Using DNS- 1220 Based Authentication of Named Entities (DANE) TLSA Records 1221 with SRV Records", RFC 7673, DOI 10.17487/RFC7673, October 1222 2015, . 1224 [DegabrieleP07] 1225 Degabriele, J. and K. Paterson, "Attacking the IPsec 1226 Standards in Encryption-only Configurations", 2007 IEEE 1227 Symposium on Security and Privacy (SP '07), 1228 DOI 10.1109/sp.2007.8, May 2007, 1229 . 1231 [DEP-SSLv3] 1232 Barnes, R., Thomson, M., Pironti, A., and A. Langley, 1233 "Deprecating Secure Sockets Layer Version 3.0", RFC 7568, 1234 DOI 10.17487/RFC7568, June 2015, 1235 . 1237 [Heninger2012] 1238 Heninger, N., Durumeric, Z., Wustrow, E., and J.A. 1239 Halderman, "Mining Your Ps and Qs: Detection of Widespread 1240 Weak Keys in Network Devices", Usenix Security 1241 Symposium 2012, 2012. 1243 [I-D.ietf-tls-esni] 1244 Rescorla, E., Oku, K., Sullivan, N., and C. A. Wood, "TLS 1245 Encrypted Client Hello", Work in Progress, Internet-Draft, 1246 draft-ietf-tls-esni-13, 12 August 2021, 1247 . 1250 [I-D.irtf-cfrg-aead-limits] 1251 Günther, F., Thomson, M., and C. A. Wood, "Usage Limits on 1252 AEAD Algorithms", Work in Progress, Internet-Draft, draft- 1253 irtf-cfrg-aead-limits-03, 12 July 2021, 1254 . 1257 [IANA_TLS] IANA, "Transport Layer Security (TLS) Parameters", 1258 . 1260 [Joux2006] Joux, A., "Authentication Failures in NIST version of 1261 GCM", 2006, . 1265 [Kleinjung2010] 1266 Kleinjung, T., Aoki, K., Franke, J., Lenstra, A., Thomé, 1267 E., Bos, J., Gaudry, P., Kruppa, A., Montgomery, P., 1268 Osvik, D., te Riele, H., Timofeev, A., and P. Zimmermann, 1269 "Factorization of a 768-Bit RSA Modulus", Advances in 1270 Cryptology - CRYPTO 2010 pp. 333-350, 1271 DOI 10.1007/978-3-642-14623-7_18, 2010, 1272 . 1274 [Krawczyk2001] 1275 Krawczyk, H., "The Order of Encryption and Authentication 1276 for Protecting Communications (Or: How Secure is SSL?)", 1277 CRYPTO 01, 2001, 1278 . 1280 [Logjam] Adrian, D., Bhargavan, K., Durumeric, Z., Gaudry, P., 1281 Green, M., Halderman, J., Heninger, N., Springall, D., 1282 Thomé, E., Valenta, L., VanderSloot, B., Wustrow, E., 1283 Zanella-Béguelin, S., and P. Zimmermann, "Imperfect 1284 Forward Secrecy: How Diffie-Hellman Fails in Practice", 1285 Proceedings of the 22nd ACM SIGSAC Conference on Computer 1286 and Communications Security, DOI 10.1145/2810103.2813707, 1287 October 2015, . 1289 [Multiple-Encryption] 1290 Merkle, R. and M. Hellman, "On the security of multiple 1291 encryption", Communications of the ACM Vol. 24, pp. 1292 465-467, DOI 10.1145/358699.358718, July 1981, 1293 . 1295 [NIST.SP.800-56A] 1296 Barker, E., Chen, L., Roginsky, A., Vassilev, A., and R. 1297 Davis, "Recommendation for pair-wise key-establishment 1298 schemes using discrete logarithm cryptography", National 1299 Institute of Standards and Technology report, 1300 DOI 10.6028/nist.sp.800-56ar3, April 2018, 1301 . 1303 [PatersonRS11] 1304 Paterson, K., Ristenpart, T., and T. Shrimpton, "Tag Size 1305 Does Matter: Attacks and Proofs for the TLS Record 1306 Protocol", Lecture Notes in Computer Science pp. 372-389, 1307 DOI 10.1007/978-3-642-25385-0_20, 2011, 1308 . 1310 [POODLE] US-CERT, "SSL 3.0 Protocol Vulnerability and POODLE 1311 Attack", October 2014, 1312 . 1314 [RACCOON] Merget, R., Brinkmann, M., Aviram, N., Somorovsky, J., 1315 Mittmann, J., and J. Schwenk, "Raccoon Attack: Finding and 1316 Exploiting Most-Significant-Bit-Oracles in TLS-DH(E)", 1317 30th USENIX Security Symposium (USENIX Security 21) , 1318 2021, . 1321 [RFC2026] Bradner, S., "The Internet Standards Process -- Revision 1322 3", BCP 9, RFC 2026, DOI 10.17487/RFC2026, October 1996, 1323 . 1325 [RFC2246] Dierks, T. and C. Allen, "The TLS Protocol Version 1.0", 1326 RFC 2246, DOI 10.17487/RFC2246, January 1999, 1327 . 1329 [RFC3602] Frankel, S., Glenn, R., and S. Kelly, "The AES-CBC Cipher 1330 Algorithm and Its Use with IPsec", RFC 3602, 1331 DOI 10.17487/RFC3602, September 2003, 1332 . 1334 [RFC4346] Dierks, T. and E. Rescorla, "The Transport Layer Security 1335 (TLS) Protocol Version 1.1", RFC 4346, 1336 DOI 10.17487/RFC4346, April 2006, 1337 . 1339 [RFC4347] Rescorla, E. and N. Modadugu, "Datagram Transport Layer 1340 Security", RFC 4347, DOI 10.17487/RFC4347, April 2006, 1341 . 1343 [RFC5077] Salowey, J., Zhou, H., Eronen, P., and H. Tschofenig, 1344 "Transport Layer Security (TLS) Session Resumption without 1345 Server-Side State", RFC 5077, DOI 10.17487/RFC5077, 1346 January 2008, . 1348 [RFC5116] McGrew, D., "An Interface and Algorithms for Authenticated 1349 Encryption", RFC 5116, DOI 10.17487/RFC5116, January 2008, 1350 . 1352 [RFC5280] Cooper, D., Santesson, S., Farrell, S., Boeyen, S., 1353 Housley, R., and W. Polk, "Internet X.509 Public Key 1354 Infrastructure Certificate and Certificate Revocation List 1355 (CRL) Profile", RFC 5280, DOI 10.17487/RFC5280, May 2008, 1356 . 1358 [RFC6101] Freier, A., Karlton, P., and P. Kocher, "The Secure 1359 Sockets Layer (SSL) Protocol Version 3.0", RFC 6101, 1360 DOI 10.17487/RFC6101, August 2011, 1361 . 1363 [RFC6120] Saint-Andre, P., "Extensible Messaging and Presence 1364 Protocol (XMPP): Core", RFC 6120, DOI 10.17487/RFC6120, 1365 March 2011, . 1367 [RFC6460] Salter, M. and R. Housley, "Suite B Profile for Transport 1368 Layer Security (TLS)", RFC 6460, DOI 10.17487/RFC6460, 1369 January 2012, . 1371 [RFC6698] Hoffman, P. and J. Schlyter, "The DNS-Based Authentication 1372 of Named Entities (DANE) Transport Layer Security (TLS) 1373 Protocol: TLSA", RFC 6698, DOI 10.17487/RFC6698, August 1374 2012, . 1376 [RFC6797] Hodges, J., Jackson, C., and A. Barth, "HTTP Strict 1377 Transport Security (HSTS)", RFC 6797, 1378 DOI 10.17487/RFC6797, November 2012, 1379 . 1381 [RFC6960] Santesson, S., Myers, M., Ankney, R., Malpani, A., 1382 Galperin, S., and C. Adams, "X.509 Internet Public Key 1383 Infrastructure Online Certificate Status Protocol - OCSP", 1384 RFC 6960, DOI 10.17487/RFC6960, June 2013, 1385 . 1387 [RFC6961] Pettersen, Y., "The Transport Layer Security (TLS) 1388 Multiple Certificate Status Request Extension", RFC 6961, 1389 DOI 10.17487/RFC6961, June 2013, 1390 . 1392 [RFC6989] Sheffer, Y. and S. Fluhrer, "Additional Diffie-Hellman 1393 Tests for the Internet Key Exchange Protocol Version 2 1394 (IKEv2)", RFC 6989, DOI 10.17487/RFC6989, July 2013, 1395 . 1397 [RFC7435] Dukhovni, V., "Opportunistic Security: Some Protection 1398 Most of the Time", RFC 7435, DOI 10.17487/RFC7435, 1399 December 2014, . 1401 [RFC7457] Sheffer, Y., Holz, R., and P. Saint-Andre, "Summarizing 1402 Known Attacks on Transport Layer Security (TLS) and 1403 Datagram TLS (DTLS)", RFC 7457, DOI 10.17487/RFC7457, 1404 February 2015, . 1406 [RFC7525] Sheffer, Y., Holz, R., and P. Saint-Andre, 1407 "Recommendations for Secure Use of Transport Layer 1408 Security (TLS) and Datagram Transport Layer Security 1409 (DTLS)", BCP 195, RFC 7525, DOI 10.17487/RFC7525, May 1410 2015, . 1412 [RFC8452] Gueron, S., Langley, A., and Y. Lindell, "AES-GCM-SIV: 1413 Nonce Misuse-Resistant Authenticated Encryption", 1414 RFC 8452, DOI 10.17487/RFC8452, April 2019, 1415 . 1417 [RFC8470] Thomson, M., Nottingham, M., and W. Tarreau, "Using Early 1418 Data in HTTP", RFC 8470, DOI 10.17487/RFC8470, September 1419 2018, . 1421 [RFC9001] Thomson, M., Ed. and S. Turner, Ed., "Using TLS to Secure 1422 QUIC", RFC 9001, DOI 10.17487/RFC9001, May 2021, 1423 . 1425 [SESSION-HASH] 1426 Bhargavan, K., Ed., Delignat-Lavaud, A., Pironti, A., 1427 Langley, A., and M. Ray, "Transport Layer Security (TLS) 1428 Session Hash and Extended Master Secret Extension", 1429 RFC 7627, DOI 10.17487/RFC7627, September 2015, 1430 . 1432 [Smith2013] 1433 Smith, B., "Proposal to Change the Default TLS 1434 Ciphersuites Offered by Browsers.", 2013, 1435 . 1437 [Soghoian2011] 1438 Soghoian, C. and S. Stamm, "Certified Lies: Detecting and 1439 Defeating Government Interception Attacks Against SSL", 1440 SSRN Electronic Journal, DOI 10.2139/ssrn.1591033, 2010, 1441 . 1443 [Sy2018] Sy, E., Burkert, C., Federrath, H., and M. Fischer, 1444 "Tracking Users across the Web via TLS Session 1445 Resumption", Proceedings of the 34th Annual Computer 1446 Security Applications Conference, 1447 DOI 10.1145/3274694.3274708, December 2018, 1448 . 1450 [TLS-XMPP] Saint-Andre, P. and T. Alkemade, "Use of Transport Layer 1451 Security (TLS) in the Extensible Messaging and Presence 1452 Protocol (XMPP)", RFC 7590, DOI 10.17487/RFC7590, June 1453 2015, . 1455 [triple-handshake] 1456 Bhargavan, K., Lavaud, A., Fournet, C., Pironti, A., and 1457 P. Strub, "Triple Handshakes and Cookie Cutters: Breaking 1458 and Fixing Authentication over TLS", 2014 IEEE Symposium 1459 on Security and Privacy, DOI 10.1109/sp.2014.14, May 2014, 1460 . 1462 Appendix A. Differences from RFC 7525 1464 This revision of the Best Current Practices contains numerous 1465 changes, and this section is focused on the normative changes. 1467 * High level differences: 1469 - Clarified items (e.g. renegotiation) that only apply to TLS 1470 1.2. 1472 - Changed status of TLS 1.0 and 1.1 from SHOULD NOT to MUST NOT. 1474 - Added TLS 1.3 at a SHOULD level. 1476 - Similar changes to DTLS, pending publication of DTLS 1.3. 1478 - Specific guidance for multiplexed protocols. 1480 - MUST-level implementation requirement for ALPN, and more 1481 specific SHOULD-level guidance for ALPN and SNI. 1483 - New attacks since [RFC7457]: ALPACA, Raccoon, Logjam, "Nonce- 1484 Disrespecting Adversaries" 1486 * Differences specific to TLS 1.2: 1488 - Fallback SCSV as a MUST for TLS 1.2. 1490 - SHOULD-level guidance on AES-GCM nonce generation in TLS 1.2. 1492 - SHOULD NOT use static DH keys or reuse ephemeral DH keys across 1493 multiple connections. 1495 - 2048-bit DH now a MUST, ECDH minimal curve size is 224, vs. 192 1496 previously. 1498 * Differences specific to TLS 1.3: 1500 - New TLS 1.3 capabilities: 0-RTT. 1502 - Removed capabilities: renegotiation, compression. 1504 - Added mention of TLS Encrypted Client Hello, but no 1505 recommendation to use until it is finalized. 1507 - SHOULD-level requirement for forward secrecy in TLS 1.3 session 1508 resumption. 1510 - Generic SHOULD-level guidance to avoid 0-RTT unless it is 1511 documented for the particular protocol. 1513 Appendix B. Document History 1515 // Note to RFC Editor: please remove before publication. 1517 B.1. draft-ietf-uta-rfc7525bis-02 1519 * Adjusted text about ALPN support in application protocols 1521 * Incorporated text from draft-ietf-tls-md5-sha1-deprecate 1523 B.2. draft-ietf-uta-rfc7525bis-01 1525 * Many more changes, including: 1527 - SHOULD-level requirement for forward secrecy in TLS 1.3 session 1528 resumption. 1530 - Removed TLS 1.2 capabilities: renegotiation, compression. 1532 - Specific guidance for multiplexed protocols. 1534 - MUST-level implementation requirement for ALPN, and more 1535 specific SHOULD-level guidance for ALPN and SNI. 1537 - Generic SHOULD-level guidance to avoid 0-RTT unless it is 1538 documented for the particular protocol. 1540 - SHOULD-level guidance on AES-GCM nonce generation in TLS 1.2. 1542 - SHOULD NOT use static DH keys or reuse ephemeral DH keys across 1543 multiple connections. 1545 - 2048-bit DH now a MUST, ECDH minimal curve size is 224, up from 1546 192. 1548 B.3. draft-ietf-uta-rfc7525bis-00 1550 * Renamed: WG document. 1552 * Started populating list of changes from RFC 7525. 1554 * General rewording of abstract and intro for revised version. 1556 * Protocol versions, fallback. 1558 * Reference to ECHO. 1560 B.4. draft-sheffer-uta-rfc7525bis-00 1562 * Renamed, since the BCP number does not change. 1564 * Added an empty "Differences from RFC 7525" section. 1566 B.5. draft-sheffer-uta-bcp195bis-00 1568 * Initial release, the RFC 7525 text as-is, with some minor 1569 editorial changes to the references. 1571 Authors' Addresses 1573 Yaron Sheffer 1574 Intuit 1576 Email: yaronf.ietf@gmail.com 1578 Ralph Holz 1579 University of Twente 1581 Email: ralph.ietf@gmail.com 1583 Peter Saint-Andre 1584 Mozilla 1586 Email: stpeter@mozilla.com 1588 Thomas Fossati 1589 arm 1591 Email: thomas.fossati@arm.com