idnits 2.17.1 draft-ietf-uta-tls-bcp-02.txt: 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: ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/checklist : ---------------------------------------------------------------------------- No issues found here. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year -- The document date (August 24, 2014) is 3533 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) ** Obsolete normative reference: RFC 2818 (Obsoleted by RFC 9110) ** Obsolete normative reference: RFC 4492 (Obsoleted by RFC 8422) ** Obsolete normative reference: RFC 5246 (Obsoleted by RFC 8446) ** Obsolete normative reference: RFC 6125 (Obsoleted by RFC 9525) == Outdated reference: A later version (-01) exists of draft-ietf-tls-prohibiting-rc4-00 == Outdated reference: A later version (-05) exists of draft-ietf-uta-tls-attacks-01 -- Obsolete informational reference (is this intentional?): RFC 2246 (Obsoleted by RFC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 5077 (Obsoleted by RFC 8446) -- Obsolete informational reference (is this intentional?): RFC 6961 (Obsoleted by RFC 8446) Summary: 4 errors (**), 0 flaws (~~), 3 warnings (==), 5 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 UTA Y. Sheffer 3 Internet-Draft Porticor 4 Intended status: Best Current Practice R. Holz 5 Expires: February 25, 2015 TUM 6 P. Saint-Andre 7 &yet 8 August 24, 2014 10 Recommendations for Secure Use of TLS and DTLS 11 draft-ietf-uta-tls-bcp-02 13 Abstract 15 Transport Layer Security (TLS) and Datagram Transport Security Layer 16 (DTLS) are widely used to protect data exchanged over application 17 protocols such as HTTP, SMTP, IMAP, POP, SIP, and XMPP. Over the 18 last few years, several serious attacks on TLS have emerged, 19 including attacks on its most commonly used cipher suites and modes 20 of operation. This document provides recommendations for improving 21 the security of both software implementations and deployed services 22 that use TLS and DTLS. 24 Status of This Memo 26 This Internet-Draft is submitted in full conformance with the 27 provisions of BCP 78 and BCP 79. 29 Internet-Drafts are working documents of the Internet Engineering 30 Task Force (IETF). Note that other groups may also distribute 31 working documents as Internet-Drafts. The list of current Internet- 32 Drafts is at http://datatracker.ietf.org/drafts/current/. 34 Internet-Drafts are draft documents valid for a maximum of six months 35 and may be updated, replaced, or obsoleted by other documents at any 36 time. It is inappropriate to use Internet-Drafts as reference 37 material or to cite them other than as "work in progress." 39 This Internet-Draft will expire on February 25, 2015. 41 Copyright Notice 43 Copyright (c) 2014 IETF Trust and the persons identified as the 44 document authors. All rights reserved. 46 This document is subject to BCP 78 and the IETF Trust's Legal 47 Provisions Relating to IETF Documents 48 (http://trustee.ietf.org/license-info) in effect on the date of 49 publication of this document. Please review these documents 50 carefully, as they describe your rights and restrictions with respect 51 to this document. Code Components extracted from this document must 52 include Simplified BSD License text as described in Section 4.e of 53 the Trust Legal Provisions and are provided without warranty as 54 described in the Simplified BSD License. 56 Table of Contents 58 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 3 59 2. Conventions used in this document . . . . . . . . . . . . . . 3 60 3. General Recommendations . . . . . . . . . . . . . . . . . . . 4 61 3.1. Protocol Versions . . . . . . . . . . . . . . . . . . . . 4 62 3.2. Fallback to SSL . . . . . . . . . . . . . . . . . . . . . 4 63 3.3. Always Use TLS . . . . . . . . . . . . . . . . . . . . . 5 64 3.4. Compression . . . . . . . . . . . . . . . . . . . . . . . 5 65 3.5. Session Resumption . . . . . . . . . . . . . . . . . . . 5 66 3.6. Renegotiation . . . . . . . . . . . . . . . . . . . . . . 6 67 3.7. Server Name Indication . . . . . . . . . . . . . . . . . 6 68 4. Recommendations: Cipher Suites . . . . . . . . . . . . . . . 6 69 4.1. Cipher Suite Selection . . . . . . . . . . . . . . . . . 6 70 4.2. Public Key Length . . . . . . . . . . . . . . . . . . . . 8 71 4.3. Cipher Suite Negotiation Details . . . . . . . . . . . . 8 72 4.4. Modular vs. Elliptic Curve DH Cipher Suites . . . . . . . 9 73 5. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 10 74 6. Security Considerations . . . . . . . . . . . . . . . . . . . 10 75 6.1. Host Name Validation . . . . . . . . . . . . . . . . . . 10 76 6.2. AES-GCM . . . . . . . . . . . . . . . . . . . . . . . . . 10 77 6.3. Forward Secrecy . . . . . . . . . . . . . . . . . . . . . 10 78 6.4. Diffie Hellman Exponent Reuse . . . . . . . . . . . . . . 11 79 6.5. Certificate Revocation . . . . . . . . . . . . . . . . . 12 80 7. Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . 12 81 8. References . . . . . . . . . . . . . . . . . . . . . . . . . 12 82 8.1. Normative References . . . . . . . . . . . . . . . . . . 13 83 8.2. Informative References . . . . . . . . . . . . . . . . . 13 84 Appendix A. Change Log . . . . . . . . . . . . . . . . . . . . . 15 85 A.1. draft-ietf-tls-bcp-02 . . . . . . . . . . . . . . . . . . 15 86 A.2. draft-ietf-tls-bcp-01 . . . . . . . . . . . . . . . . . . 15 87 A.3. draft-ietf-tls-bcp-00 . . . . . . . . . . . . . . . . . . 16 88 A.4. draft-sheffer-tls-bcp-02 . . . . . . . . . . . . . . . . 16 89 A.5. draft-sheffer-tls-bcp-01 . . . . . . . . . . . . . . . . 16 90 A.6. draft-sheffer-tls-bcp-00 . . . . . . . . . . . . . . . . 16 91 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 16 93 1. Introduction 95 Transport Layer Security (TLS) and Datagram Transport Security Layer 96 (DTLS) are widely used to protect data exchanged over application 97 protocols such as HTTP, SMTP, IMAP, POP, SIP, and XMPP. Over the 98 last few years, several serious attacks on TLS have emerged, 99 including attacks on its most commonly used cipher suites and modes 100 of operation. For instance, both AES-CBC and RC4, which together 101 comprise most current usage, have been attacked in the context of 102 TLS. A companion document [I-D.ietf-uta-tls-attacks] provides 103 detailed information about these attacks. 105 Because of these attacks, those who implement and deploy TLS and DTLS 106 need updated guidance on how TLS can be used securely. Note that 107 this document provides guidance for deployed services, as well as 108 software implementations. In fact, this document calls for the 109 deployment of algorithms that are widely implemented but not yet 110 widely deployed. 112 The recommendations herein take into consideration the security of 113 various mechanisms, their technical maturity and interoperability, 114 and their prevalence in implementations at the time of writing. 115 These recommendations apply to both TLS and DTLS. TLS 1.3, when it 116 is standardized and deployed in the field, should resolve the current 117 vulnerabilities while providing significantly better functionality, 118 and will very likely obsolete this document. 120 These are minimum recommendations for the general use of TLS. 121 Individual specifications may have stricter requirements related to 122 one or more aspects of the protocol, based on their particular 123 circumstances. When that is the case, implementers MUST adhere to 124 those stricter requirements. 126 Community knowledge about the strength of various algorithms and 127 feasible attacks can change quickly, and experience shows that a 128 security BCP is a point-in-time statement. Readers are advised to 129 seek out any errata or updates that apply to this document. 131 2. Conventions used in this document 133 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 134 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this 135 document are to be interpreted as described in [RFC2119]. 137 3. General Recommendations 139 This section provides general recommendations on the secure use of 140 TLS. Recommendations related to cipher suites are discussed in the 141 following section. 143 3.1. Protocol Versions 145 It is important both to stop using old, less secure versions of SSL/ 146 TLS and to start using modern, more secure versions. Therefore: 148 o Implementations MUST NOT negotiate SSL version 2. 150 Rationale: SSLv2 is considered today as insecure [RFC6176]. 152 o Implementations MUST NOT negotiate SSL version 3. 154 Rationale: SSLv3 [RFC6101] was an improvement over SSLv2 and 155 plugged some significant security holes, but did not support 156 strong cipher suites. In addition, SSLv3 does not support TLS 157 extensions, some of which are considered security-critical today. 159 o Implementations SHOULD NOT negotiate TLS version 1.0 [RFC2246]. 161 Rationale: TLS 1.0 (published in 1999) does not support many 162 modern, strong cipher suites. 164 o Implementations MAY negotiate TLS version 1.1 [RFC4346]. 166 Rationale: TLS 1.1 (published in 2006) is a security improvement 167 over TLS 1.0, but still does not support certain stronger cipher 168 suites. 170 o Implementations MUST support, and prefer to negotiate, TLS version 171 1.2 [RFC5246]. 173 Rationale: Several stronger cipher suites are available only with 174 TLS 1.2 (published in 2008). 176 This BCP applies to TLS 1.2. It is not safe for readers to assume 177 that the recommendations in this BCP apply to any future version of 178 TLS. 180 3.2. Fallback to SSL 182 Some client implementations revert to lower versions of TLS or even 183 to SSLv3 if the server rejected higher versions of the protocol. 185 This fall back can be forced by a man in the middle (MITM) attacker. 186 By default, such clients MUST NOT fall back to SSLv3. 188 Rationale: TLS 1.0 and SSLv3 are significantly less secure than TLS 189 1.2, the version recommended by this document. While TLS 1.0-only 190 servers are still quite common, IP scans show that SSLv3-only servers 191 amount to only about 3% of the current Web server population. 193 3.3. Always Use TLS 195 Combining unprotected and TLS-protected communication opens the way 196 to SSL Stripping and similar attacks. Therefore: 198 o In cases where an application protocol allows implementations or 199 deployments a choice between strict TLS configuration and dynamic 200 upgrade from unencrypted to TLS-protected traffic (such as 201 STARTTLS), clients and servers SHOULD prefer strict TLS 202 configuration. 204 o When applicable, Web servers SHOULD advertise that they are 205 willing to accept TLS-only clients, using the HTTP Strict 206 Transport Security (HSTS) header [RFC6797]. 208 3.4. Compression 210 Implementations and deployments SHOULD disable TLS-level compression 211 ([RFC5246], Sec. 6.2.2), because it has been subject to security 212 attacks. 214 Implementers should note that compression at higher protocol levels 215 can allow an active attacker to extract cleartext information from 216 the connection. The BREACH attack is one such case. These issues 217 can only be mitigated outside of TLS and are thus out of scope of the 218 current document. See Sec. 2.5 of [I-D.ietf-uta-tls-attacks] for 219 further details. 221 3.5. Session Resumption 223 If TLS session resumption is used, care ought to be taken to do so 224 safely. In particular, the resumption information (either session 225 IDs [RFC5246] or session tickets [RFC5077]) MUST be authenticated and 226 encrypted to prevent modification or eavesdropping by an attacker. 227 Further recommendations apply to session tickets: 229 o A strong cipher suite MUST be used when encrypting the ticket (as 230 least as strong as the main TLS cipher suite). 232 o Ticket keys MUST be changed regularly, e.g. once every week, so as 233 not to negate the benefits of forward secrecy (see Section 6.3 for 234 details on forward secrecy). 236 o Session ticket validity SHOULD be limited to a reasonable duration 237 (e.g. 1 day), for similar reasons. 239 3.6. Renegotiation 241 Where handshake renegotiation is implemented, both clients and 242 servers MUST implement the renegotiation_info extension, as defined 243 in [RFC5746]. 245 To counter the Triple Handshake attack, we adopt the recommendation 246 from [triple-handshake]: TLS clients SHOULD ensure that all 247 certificates received over a connection are valid for the current 248 server endpoint, and abort the handshake if they are not. In some 249 usages, it may be simplest to refuse any change of certificates 250 during renegotiation. 252 3.7. Server Name Indication 254 TLS implementations MUST support the Server Name Indication (SNI) 255 extension for those higher level protocols which would benefit from 256 it, including HTTPS. However, unlike implementation, the use of SNI 257 in particular circumstances is a matter of local policy. 259 4. Recommendations: Cipher Suites 261 TLS and its implementations provide considerable flexibility in the 262 selection of cipher suites. Unfortunately many available cipher 263 suites are insecure, and so misconfiguration can easily result in 264 reduced security. This section includes recommendations on the 265 selection and negotiation of cipher suites. 267 4.1. Cipher Suite Selection 269 It is important both to stop using old, insecure cipher suites and to 270 start using modern, more secure cipher suites. Therefore: 272 o Implementations MUST NOT negotiate the NULL cipher suites. 274 Rationale: The NULL cipher suites offer no encryption whatsoever 275 and thus are completely insecure. 277 o Implementations MUST NOT negotiate RC4 cipher suites 278 Rationale: The RC4 stream cipher has a variety of cryptographic 279 weaknesses, as documented in [I-D.ietf-tls-prohibiting-rc4]. 281 o Implementations MUST NOT negotiate cipher suites offering only so- 282 called "export-level" encryption (including algorithms with 40 283 bits or 56 bits of security). 285 Rationale: These cipher suites are deliberately "dumbed down" and 286 are very easy to break. 288 o Applications MUST NOT negotiate cipher suites of less than 112 289 bits of security. 291 o Implementations SHOULD NOT negotiate cipher suites that use 292 algorithms offering less than 128 bits of security. Note that 293 some legacy cipher suites (e.g. 168-bit 3DES) have an effective 294 key length which is smaller than their nominal key length (112 295 bits in the case of 3DES). Such cipher suites should be evaluated 296 according to their effective key length. 298 Rationale: Although these cipher suites are not actively subject 299 to breakage, their useful lifespan is short enough that stronger 300 cipher suites are desirable. 128-bit ciphers are expected to 301 remain secure for at least several years, and 256-bit ciphers 302 "until the next fundamental technology breakthrough". 304 o Implementations MUST support, and SHOULD prefer to negotiate, 305 cipher suites offering forward secrecy, such as those in the 306 Ephemeral Diffie-Hellman and Elliptic Curve Ephemeral Diffie 307 Hellman ("DHE" and "ECDHE") families. 309 Rationale: Forward secrecy (sometimes called "perfect forward 310 secrecy") prevents the recovery of information that was encrypted 311 with older session keys, thus limiting the amount of time during 312 which attacks can be successful. 314 Given the foregoing considerations, implementation of the following 315 cipher suites is RECOMMENDED (see [RFC5289] for details): 317 o TLS_DHE_RSA_WITH_AES_128_GCM_SHA256 319 o TLS_ECDHE_RSA_WITH_AES_128_GCM_SHA256 321 o TLS_DHE_RSA_WITH_AES_256_GCM_SHA384 323 o TLS_ECDHE_RSA_WITH_AES_256_GCM_SHA384 324 We suggest that TLS_ECDHE_RSA_WITH_AES_128_GCM_SHA256 be preferred in 325 general. 327 It is noted that those cipher suites are supported only in TLS 1.2 328 since they are authenticated encryption (AEAD) algorithms [RFC5116]. 330 [RFC4492] allows clients and servers to negotiate ECDH parameters 331 (curves). For interoperability, clients and servers SHOULD support 332 the NIST P-256 (secp256r1) curve [RFC4492]. In addition, clients 333 SHOULD send an ec_point_formats extension with a single element, 334 "uncompressed". 336 4.2. Public Key Length 338 With a key exchange based on modular Diffie-Hellman ("DHE" cipher 339 suites), key lengths of at least 2048 bits are RECOMMENDED. 341 Rationale: because Diffie-Hellman keys of 1024 bits are estimated to 342 be roughly equivalent to 80-bit symmetric keys, it is better to use 343 longer keys for the "DHE" family of cipher suites. Unfortunately, 344 some existing software cannot handle (or cannot easily handle) key 345 lengths greater than 1024 bits. The most common workaround for these 346 systems is to prefer the "ECDHE" family of cipher suites instead of 347 the "DHE" family. For modular groups, key lengths of at least 2048 348 bits are estimated to be roughly equivalent to 112-bit symmetric keys 349 and might be sufficient for at least the next 10 years. 351 Servers SHOULD authenticate using 2048-bit certificates. In 352 addition, the use of SHA-256 fingerprints is RECOMMENDED (see 353 [CAB-Baseline] for more details). Clients SHOULD indicate to servers 354 that they request SHA-256, by using the "Signature Algorithms" 355 extension defined in TLS 1.2. 357 4.3. Cipher Suite Negotiation Details 359 Clients SHOULD include TLS_ECDHE_RSA_WITH_AES_128_GCM_SHA256 as the 360 first proposal to any server, unless they have prior knowledge that 361 the server cannot respond to a TLS 1.2 client_hello message. 363 Servers SHOULD prefer this cipher suite whenever it is proposed, even 364 if it is not the first proposal. 366 Both clients and servers SHOULD include the "Supported Elliptic 367 Curves" extension [RFC4492]. 369 Clients are of course free to offer stronger cipher suites, e.g. 370 using AES-256; when they do, the server SHOULD prefer the stronger 371 cipher suite unless there are compelling reasons (e.g., seriously 372 degraded performance) to choose otherwise. 374 Note that other profiles of TLS 1.2 exist that use different cipher 375 suites. For example, [RFC6460] defines a profile that uses the 376 TLS_ECDHE_ECDSA_WITH_AES_128_GCM_SHA256 and 377 TLS_ECDHE_ECDSA_WITH_AES_256_GCM_SHA384 cipher suites. 379 This document is not an application profile standard, in the sense of 380 Sec. 9 of [RFC5246]. As a result, clients and servers are still 381 REQUIRED to support the mandatory TLS cipher suite, 382 TLS_RSA_WITH_AES_128_CBC_SHA. 384 4.4. Modular vs. Elliptic Curve DH Cipher Suites 386 Not all TLS implementations support both modular and EC Diffie- 387 Hellman groups, as required by Section 4.1. Some implementations are 388 severely limited in the length of DH values. When such 389 implementations need to be accommodated, we recommend using (in 390 priority order): 392 1. Elliptic Curve DHE with negotiated parameters [RFC5289] 394 2. TLS_DHE_RSA_WITH_AES_128_GCM_SHA256 [RFC5288], with 2048-bit 395 Diffie-Hellman parameters 397 3. The same cipher suite, with 1024-bit parameters. 399 Rationale: Elliptic Curve Cryptography is not universally deployed 400 for several reasons, including its complexity compared to modular 401 arithmetic and longstanding IPR concerns. On the other hand, there 402 are two related issues hindering effective use of modular Diffie- 403 Hellman cipher suites in TLS: 405 o There are no protocol mechanisms to negotiate the DH groups or 406 parameter lengths supported by client and server. 408 o There are widely deployed client implementations that reject 409 received DH parameters if they are longer than 1024 bits. 411 We note that with DHE and ECDHE cipher suites, the TLS master key 412 only depends on the Diffie Hellman parameters and not on the strength 413 of the RSA certificate; moreover, 1024 bit modular DH parameters are 414 generally considered insufficient at this time. 416 With modular ephemeral DH, deployers SHOULD carefully evaluate 417 interoperability vs. security considerations when configuring their 418 TLS endpoints. 420 5. IANA Considerations 422 This document requests no actions of IANA. [Note to RFC Editor: 423 please remove this whole section before publication.] 425 6. Security Considerations 427 This entire document discusses security practices, and this section 428 adds a few security considerations and includes further discussion of 429 particular recommendations. 431 6.1. Host Name Validation 433 Application authors should take note that TLS implementations 434 frequently do not validate host names, and must therefore determine 435 if the TLS implementation they are using does, and if not write their 436 own validation code or consider changing the TLS implementation. 438 It is noted that the requirements regarding host name validation (and 439 in general, binding between the TLS layer and the protocol that runs 440 above it) vary between different protocols. For HTTPS, these 441 requirements are defined by Sec. 3 of [RFC2818]. 443 Readers are referred to [RFC6125] for further details regarding 444 generic host name validation in the TLS context. In addition, the 445 RFC contains a long list of example protocols, some of which 446 implement a policy very different from HTTPS. 448 6.2. AES-GCM 450 Please refer to [RFC5246], Sec. 11 for general security 451 considerations when using TLS 1.2, and to [RFC5288], Sec. 6 for 452 security considerations that apply specifically to AES-GCM when used 453 with TLS. 455 6.3. Forward Secrecy 457 Forward secrecy (also often called Perfect Forward Secrecy or "PFS") 458 is a defense against an attacker who records encrypted conversations 459 where the session keys are only encrypted with the communicating 460 parties' long-term keys. Should the attacker be able to obtain these 461 long-term keys at some point later in time, he will be able to 462 decrypt the session keys and thus the entire conversation. In the 463 context of TLS and DTLS, such compromise of long-term keys is not 464 entirely implausible. It can happen, for example, due to: 466 o A client or server being attacked by some other attack vector, and 467 the private key retrieved. 469 o A long-term key retrieved from a device that has been sold or 470 otherwise decommissioned without prior wiping. 472 o A long-term key used on a device as a default key [Heninger2012]. 474 o A key generated by a Trusted Third Party like a CA, and later 475 retrieved from it either by extortion or compromise 476 [Soghoian2011]. 478 o A cryptographic break-through, or the use of asymmetric keys with 479 insufficient length [Kleinjung2010]. 481 PFS ensures in such cases that the session keys cannot be determined 482 even by an attacker who obtains the long-term keys some time after 483 the conversation. It also protects against an attacker who is in 484 possession of the long-term keys, but remains passive during the 485 conversation. 487 PFS is generally achieved by using the Diffie-Hellman scheme to 488 derive session keys. The Diffie-Hellman scheme has both parties 489 maintain private secrets and send parameters over the network as 490 modular powers over certain cyclic groups. The properties of the so- 491 called Discrete Logarithm Problem (DLP) allow to derive the session 492 keys without an eavesdropper being able to do so. There is currently 493 no known attack against DLP if sufficiently large parameters are 494 chosen. A variant of the Diffie-Hellman scheme uses Elliptic Curves 495 instead of the originally proposed modular arithmetics. 497 Unfortunately, many TLS/DTLS cipher suites were defined that do not 498 feature PFS, e.g. TLS_RSA_WITH_AES_256_CBC_SHA256. We thus advocate 499 strict use of PFS-only ciphers. 501 6.4. Diffie Hellman Exponent Reuse 503 For performance reasons, many TLS implementations reuse Diffie- 504 Hellman and Elliptic Curve Diffie-Hellman exponents across multiple 505 connections. Such reuse can result in major security issues: 507 o If exponents are reused for a long time (e.g., more than a few 508 hours), an attacker who gains access to the host can decrypt 509 previous connections. In other words, exponent reuse negates the 510 effects of forward secrecy. 512 o TLS implementations that reuse exponents should test the DH public 513 key they receive, in order to avoid some known attacks. These 514 tests are not standardized in TLS at the time of writing. See 515 [RFC6989] for recipient tests required of IKEv2 implementations 516 that reuse DH exponents. 518 6.5. Certificate Revocation 520 Unfortunately there is currently no effective, Internet-scale 521 mechanism to affect certificate revocation: 523 o Certificate Revocation Lists (CRLs) are non-scalable and therefore 524 rarely used. 526 o The On-Line Certification Status Protocol (OCSP) presents both 527 scaling and privacy issues when used for heavy traffic Web 528 servers. In addition, clients typically "soft-fail", meaning they 529 do not abort the TLS connection if the OCSP server does not 530 respond. 532 o OCSP stapling (Sec. 8 of [RFC6066]) resolves the operational 533 issues with OCSP, but is still ineffective in the presence of a 534 MITM attacker because they can simply ignore the client's request 535 for a stapled OCSP response. 537 o OCSP stapling as defined in [RFC6066] does not extend to 538 intermediate certificates used in a certificate chain. [RFC6961] 539 addresses this shortcoming, but is a recent addition without much 540 deployment. 542 o Proprietary mechanisms that embed revocation lists in the Web 543 browser's configuration database cannot scale beyond a small 544 number of the most heavily used Web servers. 546 The current consensus appears to be that OCSP stapling, combined with 547 a "must staple" mechanism similar to HSTS, would finally resolve this 548 problem; in particular when used together with the extension defined 549 in [RFC6961]. But such a mechanism has not been standardized yet. 551 7. Acknowledgments 553 We would like to thank Stephen Farrell, Simon Josefsson, Watson Ladd, 554 Johannes Merkle, Bodo Moeller, Yoav Nir, Kenny Paterson, Patrick 555 Pelletier, Tom Ritter, Rich Salz, Aaron Zauner for their review. 556 Thanks to Brian Smith whose "browser cipher suites" page is a great 557 resource. Finally, thanks to all others who commented on the TLS, 558 UTA and other lists and are not mentioned here by name. 560 8. References 561 8.1. Normative References 563 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 564 Requirement Levels", BCP 14, RFC 2119, March 1997. 566 [RFC2818] Rescorla, E., "HTTP Over TLS", RFC 2818, May 2000. 568 [RFC4492] Blake-Wilson, S., Bolyard, N., Gupta, V., Hawk, C., and B. 569 Moeller, "Elliptic Curve Cryptography (ECC) Cipher Suites 570 for Transport Layer Security (TLS)", RFC 4492, May 2006. 572 [RFC5246] Dierks, T. and E. Rescorla, "The Transport Layer Security 573 (TLS) Protocol Version 1.2", RFC 5246, August 2008. 575 [RFC5288] Salowey, J., Choudhury, A., and D. McGrew, "AES Galois 576 Counter Mode (GCM) Cipher Suites for TLS", RFC 5288, 577 August 2008. 579 [RFC5289] Rescorla, E., "TLS Elliptic Curve Cipher Suites with 580 SHA-256/384 and AES Galois Counter Mode (GCM)", RFC 5289, 581 August 2008. 583 [RFC5746] Rescorla, E., Ray, M., Dispensa, S., and N. Oskov, 584 "Transport Layer Security (TLS) Renegotiation Indication 585 Extension", RFC 5746, February 2010. 587 [RFC6125] Saint-Andre, P. and J. Hodges, "Representation and 588 Verification of Domain-Based Application Service Identity 589 within Internet Public Key Infrastructure Using X.509 590 (PKIX) Certificates in the Context of Transport Layer 591 Security (TLS)", RFC 6125, March 2011. 593 [RFC6176] Turner, S. and T. Polk, "Prohibiting Secure Sockets Layer 594 (SSL) Version 2.0", RFC 6176, March 2011. 596 8.2. Informative References 598 [CAB-Baseline] 599 CA/Browser Forum, , "Baseline Requirements for the 600 Issuance and Management of Publicly-Trusted Certificates 601 Version 1.1.6", 2013, . 604 [Heninger2012] 605 Heninger, N., Durumeric, Z., Wustrow, E., and J. 606 Halderman, "Mining Your Ps and Qs: Detection of Widespread 607 Weak Keys in Network Devices", Usenix Security Symposium 608 2012, 2012. 610 [I-D.ietf-tls-prohibiting-rc4] 611 Popov, A., "Prohibiting RC4 Cipher Suites", draft-ietf- 612 tls-prohibiting-rc4-00 (work in progress), July 2014. 614 [I-D.ietf-uta-tls-attacks] 615 Sheffer, Y., Holz, R., and P. Saint-Andre, "Summarizing 616 Current Attacks on TLS and DTLS", draft-ietf-uta-tls- 617 attacks-01 (work in progress), June 2014. 619 [Kleinjung2010] 620 Kleinjung, T., "Factorization of a 768-Bit RSA Modulus", 621 CRYPTO 10, 2010, . 623 [RFC2246] Dierks, T. and C. Allen, "The TLS Protocol Version 1.0", 624 RFC 2246, January 1999. 626 [RFC4346] Dierks, T. and E. Rescorla, "The Transport Layer Security 627 (TLS) Protocol Version 1.1", RFC 4346, April 2006. 629 [RFC5077] Salowey, J., Zhou, H., Eronen, P., and H. Tschofenig, 630 "Transport Layer Security (TLS) Session Resumption without 631 Server-Side State", RFC 5077, January 2008. 633 [RFC5116] McGrew, D., "An Interface and Algorithms for Authenticated 634 Encryption", RFC 5116, January 2008. 636 [RFC6066] Eastlake, D., "Transport Layer Security (TLS) Extensions: 637 Extension Definitions", RFC 6066, January 2011. 639 [RFC6101] Freier, A., Karlton, P., and P. Kocher, "The Secure 640 Sockets Layer (SSL) Protocol Version 3.0", RFC 6101, 641 August 2011. 643 [RFC6460] Salter, M. and R. Housley, "Suite B Profile for Transport 644 Layer Security (TLS)", RFC 6460, January 2012. 646 [RFC6797] Hodges, J., Jackson, C., and A. Barth, "HTTP Strict 647 Transport Security (HSTS)", RFC 6797, November 2012. 649 [RFC6961] Pettersen, Y., "The Transport Layer Security (TLS) 650 Multiple Certificate Status Request Extension", RFC 6961, 651 June 2013. 653 [RFC6989] Sheffer, Y. and S. Fluhrer, "Additional Diffie-Hellman 654 Tests for the Internet Key Exchange Protocol Version 2 655 (IKEv2)", RFC 6989, July 2013. 657 [Soghoian2011] 658 Soghoian, C. and S. Stamm, "Certified lies: Detecting and 659 defeating government interception attacks against SSL.", 660 Proc. 15th Int. Conf. Financial Cryptography and Data 661 Security , 2011. 663 [triple-handshake] 664 Delignat-Lavaud, A., Bhargavan, K., and A. Pironti, 665 "Triple Handshakes Considered Harmful: Breaking and Fixing 666 Authentication over TLS", 2014, . 669 Appendix A. Change Log 671 Note to RFC Editor: please remove this section before publication. 673 A.1. draft-ietf-tls-bcp-02 675 o Rearranged some sections for clarity and re-styled the text so 676 that normative text is followed by rationale where possible. 678 o Removed the recommendation to use Brainpool curves. 680 o Triple Handshake mitigation. 682 o MUST NOT negotiate algorithms lower than 112 bits of security. 684 o MUST implement SNI, but use per local policy. 686 o Changed SHOULD NOT negotiate or fall back to SSLv3 to MUST NOT. 688 o Added hostname validation. 690 o Non-normative discussion of DH exponent reuse. 692 A.2. draft-ietf-tls-bcp-01 694 o Clarified that specific TLS-using protocols may have stricter 695 requirements. 697 o Changed TLS 1.0 from MAY to SHOULD NOT. 699 o Added discussion of "optional TLS" and HSTS. 701 o Recommended use of the Signature Algorithm and Renegotiation Info 702 extensions. 704 o Use of a strong cipher for a resumption ticket: changed SHOULD to 705 MUST. 707 o Added an informational discussion of certificate revocation, but 708 no recommendations. 710 A.3. draft-ietf-tls-bcp-00 712 o Initial WG version, with only updated references. 714 A.4. draft-sheffer-tls-bcp-02 716 o Reorganized the content to focus on recommendations. 718 o Moved description of attacks to a separate document (draft- 719 sheffer-uta-tls-attacks). 721 o Strengthened recommendations regarding session resumption. 723 A.5. draft-sheffer-tls-bcp-01 725 o Clarified our motivation in the introduction. 727 o Added a section justifying the need for PFS. 729 o Added recommendations for RSA and DH parameter lengths. Moved 730 from DHE to ECDHE, with a discussion on whether/when DHE is 731 appropriate. 733 o Recommendation to avoid fallback to SSLv3. 735 o Initial information about browser support - more still needed! 737 o More clarity on compression. 739 o Client can offer stronger cipher suites. 741 o Discussion of the regular TLS mandatory cipher suite. 743 A.6. draft-sheffer-tls-bcp-00 745 o Initial version. 747 Authors' Addresses 748 Yaron Sheffer 749 Porticor 750 29 HaHarash St. 751 Hod HaSharon 4501303 752 Israel 754 Email: yaronf.ietf@gmail.com 756 Ralph Holz 757 Technische Universitaet Muenchen 758 Boltzmannstr. 3 759 Garching 85748 760 Germany 762 Email: holz@net.in.tum.de 764 Peter Saint-Andre 765 &yet 767 Email: ietf@stpeter.im