tls E. Rescorla Internet-Draft RTFM, Inc. Intended status: Experimental K. Oku Expires: 10 September 2020 Fastly N. Sullivan Cloudflare C.A. Wood Apple, Inc. 9 March 2020 Encrypted Server Name Indication for TLS 1.3 draft-ietf-tls-esni-06 Abstract This document defines a simple mechanism for encrypting the Server Name Indication for TLS 1.3. Status of This Memo This Internet-Draft is submitted in full conformance with the provisions of BCP 78 and BCP 79. Internet-Drafts are working documents of the Internet Engineering Task Force (IETF). Note that other groups may also distribute working documents as Internet-Drafts. The list of current Internet- Drafts is at https://datatracker.ietf.org/drafts/current/. Internet-Drafts are draft documents valid for a maximum of six months and may be updated, replaced, or obsoleted by other documents at any time. It is inappropriate to use Internet-Drafts as reference material or to cite them other than as "work in progress." This Internet-Draft will expire on 10 September 2020. Copyright Notice Copyright (c) 2020 IETF Trust and the persons identified as the document authors. All rights reserved. This document is subject to BCP 78 and the IETF Trust's Legal Provisions Relating to IETF Documents (https://trustee.ietf.org/ license-info) in effect on the date of publication of this document. Please review these documents carefully, as they describe your rights and restrictions with respect to this document. Code Components extracted from this document must include Simplified BSD License text Rescorla, et al. Expires 10 September 2020 [Page 1] Internet-Draft TLS 1.3 SNI Encryption March 2020 as described in Section 4.e of the Trust Legal Provisions and are provided without warranty as described in the Simplified BSD License. Table of Contents 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 3 2. Conventions and Definitions . . . . . . . . . . . . . . . . . 4 3. Overview . . . . . . . . . . . . . . . . . . . . . . . . . . 4 3.1. Topologies . . . . . . . . . . . . . . . . . . . . . . . 4 3.2. ClientHello Encryption . . . . . . . . . . . . . . . . . 5 4. Encrypted ClientHello Configuration . . . . . . . . . . . . . 6 5. The "encrypted_client_hello" extension . . . . . . . . . . . 7 6. The "echo_nonce" extension . . . . . . . . . . . . . . . . . 8 6.1. Incorporating Outer Extensions . . . . . . . . . . . . . 9 7. Client Behavior . . . . . . . . . . . . . . . . . . . . . . . 10 7.1. Sending an encrypted ClientHello . . . . . . . . . . . . 10 7.2. Handling the server response . . . . . . . . . . . . . . 11 7.2.1. Accepted ECHO . . . . . . . . . . . . . . . . . . . . 12 7.2.2. Rejected ECHO . . . . . . . . . . . . . . . . . . . . 12 7.2.3. HelloRetryRequest . . . . . . . . . . . . . . . . . . 14 7.3. GREASE extensions . . . . . . . . . . . . . . . . . . . . 15 8. Client-Facing Server Behavior . . . . . . . . . . . . . . . . 15 9. Compatibility Issues . . . . . . . . . . . . . . . . . . . . 17 9.1. Misconfiguration and Deployment Concerns . . . . . . . . 17 9.2. Middleboxes . . . . . . . . . . . . . . . . . . . . . . . 18 10. TLS and HPKE CipherSuite Mapping . . . . . . . . . . . . . . 18 11. Security Considerations . . . . . . . . . . . . . . . . . . . 19 11.1. Why is cleartext DNS OK? . . . . . . . . . . . . . . . . 19 11.2. Optional Record Digests and Trial Decryption . . . . . . 19 11.3. Related Privacy Leaks . . . . . . . . . . . . . . . . . 20 11.4. Comparison Against Criteria . . . . . . . . . . . . . . 20 11.4.1. Mitigate against replay attacks . . . . . . . . . . 20 11.4.2. Avoid widely-deployed shared secrets . . . . . . . . 20 11.4.3. Prevent SNI-based DoS attacks . . . . . . . . . . . 21 11.4.4. Do not stick out . . . . . . . . . . . . . . . . . . 21 11.4.5. Forward secrecy . . . . . . . . . . . . . . . . . . 21 11.4.6. Proper security context . . . . . . . . . . . . . . 21 11.4.7. Split server spoofing . . . . . . . . . . . . . . . 21 11.4.8. Supporting multiple protocols . . . . . . . . . . . 22 11.5. Misrouting . . . . . . . . . . . . . . . . . . . . . . . 22 12. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 22 12.1. Update of the TLS ExtensionType Registry . . . . . . . . 22 12.2. Update of the TLS Alert Registry . . . . . . . . . . . . 22 13. References . . . . . . . . . . . . . . . . . . . . . . . . . 22 13.1. Normative References . . . . . . . . . . . . . . . . . . 22 13.2. Informative References . . . . . . . . . . . . . . . . . 24 Appendix A. Alternative SNI Protection Designs . . . . . . . . . 25 A.1. TLS-layer . . . . . . . . . . . . . . . . . . . . . . . . 25 Rescorla, et al. Expires 10 September 2020 [Page 2] Internet-Draft TLS 1.3 SNI Encryption March 2020 A.1.1. TLS in Early Data . . . . . . . . . . . . . . . . . . 25 A.1.2. Combined Tickets . . . . . . . . . . . . . . . . . . 25 A.2. Application-layer . . . . . . . . . . . . . . . . . . . . 26 A.2.1. HTTP/2 CERTIFICATE Frames . . . . . . . . . . . . . . 26 Appendix B. Total Client Hello Encryption . . . . . . . . . . . 26 Appendix C. Acknowledgements . . . . . . . . . . . . . . . . . . 27 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 27 1. Introduction DISCLAIMER: This is very early a work-in-progress design and has not yet seen significant (or really any) security analysis. It should not be used as a basis for building production systems. Although TLS 1.3 [RFC8446] encrypts most of the handshake, including the server certificate, there are several other channels that allow an on-path attacker to determine the domain name the client is trying to connect to, including: * Cleartext client DNS queries. * Visible server IP addresses, assuming the the server is not doing domain-based virtual hosting. * Cleartext Server Name Indication (SNI) [RFC6066] in ClientHello messages. DoH [I-D.ietf-doh-dns-over-https] and DPRIVE [RFC7858] [RFC8094] provide mechanisms for clients to conceal DNS lookups from network inspection, and many TLS servers host multiple domains on the same IP address. In such environments, SNI is an explicit signal used to determine the server's identity. Indirect mechanisms such as traffic analysis also exist. The TLS WG has extensively studied the problem of protecting SNI, but has been unable to develop a completely generic solution. [I-D.ietf-tls-sni-encryption] provides a description of the problem space and some of the proposed techniques. One of the more difficult problems is "Do not stick out" ([I-D.ietf-tls-sni-encryption]; Section 3.4): if only sensitive/private services use SNI encryption, then SNI encryption is a signal that a client is going to such a service. For this reason, much recent work has focused on concealing the fact that SNI is being protected. Unfortunately, the result often has undesirable performance consequences, incomplete coverage, or both. The design in this document takes a different approach: it assumes that private origins will co-locate with or hide behind a provider Rescorla, et al. Expires 10 September 2020 [Page 3] Internet-Draft TLS 1.3 SNI Encryption March 2020 (CDN, app server, etc.) which is able to activate encrypted SNI, by encrypting the entire ClientHello (ECHO), for all of the domains it hosts. As a result, the use of ECHO to protect the SNI does not indicate that the client is attempting to reach a private origin, but only that it is going to a particular service provider, which the observer could already tell from the IP address. 2. Conventions and Definitions The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in BCP 14 [RFC2119] [RFC8174] when, and only when, they appear in all capitals, as shown here. All TLS notation comes from [RFC8446]; Section 3. 3. Overview This document is designed to operate in one of two primary topologies shown below, which we call "Shared Mode" and "Split Mode" 3.1. Topologies +---------------------+ | | | 2001:DB8::1111 | | | Client <-----> | private.example.org | | | | public.example.com | | | +---------------------+ Server Figure 1: Shared Mode Topology In Shared Mode, the provider is the origin server for all the domains whose DNS records point to it and clients form a TLS connection directly to that provider, which has access to the plaintext of the connection. Rescorla, et al. Expires 10 September 2020 [Page 4] Internet-Draft TLS 1.3 SNI Encryption March 2020 +--------------------+ +---------------------+ | | | | | 2001:DB8::1111 | | 2001:DB8::EEEE | Client <------------------------------------>| | | public.example.com | | private.example.com | | | | | +--------------------+ +---------------------+ Client-Facing Server Backend Server Figure 2: Split Mode Topology In Split Mode, the provider is _not_ the origin server for private domains. Rather the DNS records for private domains point to the provider, and the provider's server relays the connection back to the backend server, which is the true origin server. The provider does not have access to the plaintext of the connection. 3.2. ClientHello Encryption ECHO works by encrypting the entire ClientHello, including the SNI and any additional extensions such as ALPN. This requires that each provider publish a public key and metadata which is used for ClientHello encryption for all the domains for which it serves directly or indirectly (via Split Mode). This document defines the format of the SNI encryption public key and metadata, referred to as an ECHO configuration, and delegates DNS publication details to [HTTPSSVC], though other delivery mechanisms are possible. In particular, if some of the clients of a private server are applications rather than Web browsers, those applications might have the public key and metadata preconfigured. When a client wants to form a TLS connection to any of the domains served by an ECHO-supporting provider, it constructs a ClientHello in the regular fashion containing the true SNI value (ClientHelloInner) and then encrypts it using the public key for the provider. It then constructs a new ClientHello (ClientHelloOuter) with an innocuous SNI (and potentially innocuous versions of other extensions such as ALPN [RFC7301]) and containing the encrypted ClientHelloInner as an extension. It sends ClientHelloOuter to the server. Upon receiving ClientHelloOuter, the server can then decrypt ClientHelloInner and either terminate the connection (in Shared Mode) or forward it to the backend server (in Split Mode). Note that both ClientHelloInner and ClientHelloOuter are both valid, complete ClientHello messages. ClientHelloOuter carries an encrypted representation of ClientHelloInner in a "encrypted_client_hello" extension, defined in Section 5. Rescorla, et al. Expires 10 September 2020 [Page 5] Internet-Draft TLS 1.3 SNI Encryption March 2020 4. Encrypted ClientHello Configuration ClientHello encryption configuration information is conveyed with the following ECHOConfigs structure. opaque HpkePublicKey<1..2^16-1>; uint16 HkpeKemId; // Defined in I-D.irtf-cfrg-hpke struct { opaque public_name<1..2^16-1>; HpkePublicKey public_key; HkpeKemId kem_id; CipherSuite cipher_suites<2..2^16-2>; uint16 maximum_name_length; Extension extensions<0..2^16-1>; } ECHOConfigContents; struct { uint16 version; uint16 length; select (ECHOConfig.version) { case 0xff03: ECHOConfigContents; } } ECHOConfig; ECHOConfig ECHOConfigs<1..2^16-1>; The ECHOConfigs structure contains one or more ECHOConfig structures in decreasing order of preference. This allows a server to support multiple versions of ECHO and multiple sets of ECHO parameters. The ECHOConfig structure contains the following fields: version The version of the structure. For this specification, that value SHALL be 0xff03. Clients MUST ignore any ECHOConfig structure with a version they do not understand. contents An opaque byte string whose contents depend on the version of the structure. For this specification, the contents are an ECHOConfigContents structure. The ECHOConfigContents structure contains the following fields: public_name The non-empty name of the entity trusted to update these encryption keys. This is used to repair misconfigurations, as described in Section 7.2. Rescorla, et al. Expires 10 September 2020 [Page 6] Internet-Draft TLS 1.3 SNI Encryption March 2020 public_key The HPKE [I-D.irtf-cfrg-hpke] public key which can be used by the client to encrypt the ClientHello. kem_id The HPKE [I-D.irtf-cfrg-hpke] KEM identifier corresponding to public_key. Clients MUST ignore any ECHOConfig structure with a key using a KEM they do not support. cipher_suites The list of TLS cipher suites which can be used by the client to encrypt the ClientHello. See Section 10 for information on how a cipher suite maps to corresponding HPKE algorithm identifiers. maximum_name_length the largest name the server expects to support. If the server supports arbitrary wildcard names, it SHOULD set this value to 256. Clients SHOULD reject ESNIConfig as invalid if maximum_name_length is greater than 256. extensions A list of extensions that the client can take into consideration when generating a Client Hello message. The purpose of the field is to provide room for additional features in the future. The format is defined in [RFC8446]; Section 4.2. The same interpretation rules apply: extensions MAY appear in any order, but there MUST NOT be more than one extension of the same type in the extensions block. An extension may be tagged as mandatory by using an extension type codepoint with the high order bit set to 1. A client which receives a mandatory extension they do not understand must reject the ECHOConfig content. Clients MUST parse the extension list and check for unsupported mandatory extensions. If an unsupported mandatory extension is present, clients MUST reject the ECHOConfig value. 5. The "encrypted_client_hello" extension The encrypted ClientHelloInner is carried in an "encrypted_client_hello" extension, defined as follows: enum { encrypted_client_hello(TBD), (65535) } ExtensionType; For clients (in ClientHello), this extension contains the following ClientEncryptedCH structure: Rescorla, et al. Expires 10 September 2020 [Page 7] Internet-Draft TLS 1.3 SNI Encryption March 2020 struct { CipherSuite suite; opaque record_digest<0..2^16-1>; opaque enc<1..2^16-1>; opaque encrypted_ch<1..2^16-1>; } ClientEncryptedCH; suite The cipher suite used to encrypt ClientHelloInner. record_digest A cryptographic hash of the ECHOConfig structure from which the ECHO key was obtained, i.e., from the first byte of "version" to the end of the structure. This hash is computed using the hash function associated with "suite". enc The HPKE encapsulated key, used by servers to decrypt the corresponding encrypted_ch field. encrypted_ch The serialized and encrypted ClientHelloInner structure, AEAD-encrypted using cipher suite "suite" and the key generated as described below. If the server accepts ECHO, it does not send this extension. If it rejects ECHO, then it sends the following structure in EncryptedExtensions: struct { ECHOConfigs retry_configs; } ServerEncryptedCH; retry_configs An ESNIConfigs structure containing one or more ECHOConfig structures in decreasing order of preference that the client should use on subsequent connections to encrypt the ClientHelloInner structure. This protocol also defines the "echo_required" alert, which is sent by the client when it offered an "encrypted_server_name" extension which was not accepted by the server. 6. The "echo_nonce" extension When using ECHO, the client MUST also add an extension of type "echo_nonce" to the ClientHelloInner (but not to the outer ClientHello). This nonce ensures that the server's encrypted Certificate can only be read by the entity which sent this ClientHello. [[TODO: Describe HRR cut-and-paste 1 in Security Considerations.]] This extension is defined as follows: Rescorla, et al. Expires 10 September 2020 [Page 8] Internet-Draft TLS 1.3 SNI Encryption March 2020 enum { echo_nonce(0xffce), (65535) } ExtensionType; struct { uint8 nonce[16]; } ECHONonce; nonce A 16-byte nonce exported from the HPKE encryption context. See Section 7.1 for details about its computation. Finally, requirements in Section 7 and Section 8 require implementations to track, alongside each PSK established by a previous connection, whether the connection negotiated this extension with the "echo_accept" response type. If so, this is referred to as an "ECHO PSK". Otherwise, it is a "non-ECHO PSK". This may be implemented by adding a new field to client and server session states. 6.1. Incorporating Outer Extensions Some TLS 1.3 extensions can be quite large and having them both in the inner and outer ClientHello will lead to a very large overall size. One particularly pathological example is "key_share" with post-quantum algorithms. In order to reduce the impact of duplicated extensions, the client may use the "outer_extensions" extension. enum { outer_extension(TBD), (65535) } ExtensionType; struct { ExtensionType outer_extensions<2..254>; uint8 hash<32..255>; } OuterExtensions; OuterExtensions MUST only be used in ClientHelloInner. It consists of one or more ExtensionType values, each of which reference an extension in ClientHelloOuter, and a digest of the complete ClientHelloInner. When sending ClientHello, the client first computes ClientHelloInner, including any PSK binders, and then MAY substitute extensions which it knows will be duplicated in ClientHelloOuter. To do so, the client computes a hash H of the entire ClientHelloInner message with the same hash as for the KDF used to encrypt ClienHelloInner. Then, the client removes and and replaces extensions from ClientHelloInner with a single "outer_extensions" extension. The list of Rescorla, et al. Expires 10 September 2020 [Page 9] Internet-Draft TLS 1.3 SNI Encryption March 2020 outer_extensions include those which were removed from ClientHelloInner, in the order in which they were removed. The hash contains the full ClientHelloInner hash H computed above. This process is reversed by client-facing servers upon receipt. Specifically, the server replaces the "outer_extensions" with extensions contained in ClientHelloOuter. The server then computes a hash H' of the reconstructed ClientHelloInner. If H' does not equal OuterExtensions.hash, the server aborts the connection with an "illegal_parameter" alert. Clients SHOULD only use this mechanism for extensions which are large. All other extensions SHOULD appear in both ClientHelloInner and ClientHelloOuter even if they have identical values. 7. Client Behavior 7.1. Sending an encrypted ClientHello In order to send an encrypted ClientHello, the client first determines if it supports the server's chosen KEM, as identified by ECHOConfig.kem_id. If one is supported, the client MUST select an appropriate cipher suite from the list of suites offered by the server. If the client does not support the corresponding KEM or is unable to select an appropriate group or suite, it SHOULD ignore that ECHOConfig value and MAY attempt to use another value provided by the server. The client MUST NOT send ECHO using HPKE algorithms not advertised by the server. Given a compatible ECHOConfig with fields public_key and kem_id, carrying the HpkePublicKey and KEM identifier corresponding to the server, clients compute an HPKE encryption context as follows: pkR = HPKE.KEM.Unmarshal(ECHOConfig.public_key) enc, context = SetupBaseS(pkR, "tls13-echo") echo_nonce = context.Export("tls13-echo-nonce", 16) echo_hrr_key = context.Export("tls13-echo-hrr-key", 16) Note that the HPKE algorithm identifiers are those which match the client's chosen CipherSuite, according to Section 10. The client MAY replace any large, duplicated extensions in ClientHelloInner with the corresponding "outer_extensions" extension, as described in Section 6.1. The client then generates a ClientHelloInner value. In addition to the normal values, ClientHelloInner MUST also contain: * an "echo_nonce" extension Rescorla, et al. Expires 10 September 2020 [Page 10] Internet-Draft TLS 1.3 SNI Encryption March 2020 * TLS padding [RFC7685] Padding SHOULD be P = L - D bytes, where * L = ECHOConfig.maximum_name_length, rounded up to the nearest multiple of 16 * D = len(dns_name), where dns_name is the DNS name in the ClientHelloInner "server_name" extension When offering an encrypted ClientHello, the client MUST NOT offer to resume any non-ECHO PSKs. It additionally MUST NOT offer to resume any sessions for TLS 1.2 or below. The encrypted ClientHello value is then computed as: encrypted_ch = context.Seal("", ClientHelloInner) Finally, the client MUST generate a ClientHelloOuter message containing the "encrypted_client_hello" extension with the values as indicated above. In particular, * suite contains the client's chosen ciphersuite; * record_digest contains the digest of the corresponding ECHOConfig structure; * enc contains the encapsulated key as output by SetupBaseS; and * encrypted_ch contains the HPKE encapsulated key (enc) and the ClientHelloInner ciphertext (encrypted_ch_inner). The client MUST place the value of ECHOConfig.public_name in the ClientHelloOuter "server_name" extension. The remaining contents of the ClientHelloOuter MAY be identical to those in ClientHelloInner but MAY also differ. The ClientHelloOuter MUST NOT contain a "cached_info" extension [RFC7924] with a CachedObject entry whose CachedInformationType is "cert", since this indication would divulge the true server name. 7.2. Handling the server response As described in Section 8, the server MAY either accept ECHO and use ClientHelloInner or reject it and use ClientHelloOuter. However, there is no indication in ServerHello of which one the server has done and the client must therefore use trial decryption in order to determine this. Rescorla, et al. Expires 10 September 2020 [Page 11] Internet-Draft TLS 1.3 SNI Encryption March 2020 7.2.1. Accepted ECHO If the server used ClientHelloInner, the client proceeds with the connection as usual, authenticating the connection for the origin server. 7.2.2. Rejected ECHO If the server used ClientHelloOuter, the client proceeds with the handshake, authenticating for ECHOConfig.public_name as described in Section 7.2.2.1. If authentication or the handshake fails, the client MUST return a failure to the calling application. It MUST NOT use the retry keys. Otherwise, when the handshake completes successfully with the public name authenticated, the client MUST abort the connection with an "echo_required" alert. It then processes the "retry_keys" field from the server's "encrypted_server_name" extension. If one of the values contains a version supported by the client, it can regard the ECHO keys as securely replaced by the server. It SHOULD retry the handshake with a new transport connection, using that value to encrypt the ClientHello. The value may only be applied to the retry connection. The client MUST continue to use the previously-advertised keys for subsequent connections. This avoids introducing pinning concerns or a tracking vector, should a malicious server present client-specific retry keys to identify clients. If none of the values provided in "retry_keys" contains a supported version, the client can regard ECHO as securely disabled by the server. As below, it SHOULD then retry the handshake with a new transport connection and ECHO disabled. If the field contains any other value, the client MUST abort the connection with an "illegal_parameter" alert. If the server negotiates an earlier version of TLS, or if it does not provide an "encrypted_server_name" extension in EncryptedExtensions, the client proceeds with the handshake, authenticating for ECHOConfigContents.public_name as described in Section 7.2.2.1. If an earlier version was negotiated, the client MUST NOT enable the False Start optimization [RFC7918] for this handshake. If authentication or the handshake fails, the client MUST return a failure to the calling application. It MUST NOT treat this as a secure signal to disable ECHO. Otherwise, when the handshake completes successfully with the public name authenticated, the client MUST abort the connection with an Rescorla, et al. Expires 10 September 2020 [Page 12] Internet-Draft TLS 1.3 SNI Encryption March 2020 "echo_required" alert. The client can then regard ECHO as securely disabled by the server. It SHOULD retry the handshake with a new transport connection and ECHO disabled. [[TODO: Key replacement is significantly less scary than saying that ECHO-naive servers bounce ECHO off. Is it worth defining a strict mode toggle in the ECHO keys, for a deployment to indicate it is ready for that? ]] Clients SHOULD implement a limit on retries caused by "echo_retry_request" or servers which do not acknowledge the "encrypted_server_name" extension. If the client does not retry in either scenario, it MUST report an error to the calling application. 7.2.2.1. Authenticating for the public name When the server cannot decrypt or does not process the "encrypted_server_name" extension, it continues with the handshake using the cleartext "server_name" extension instead (see Section 8). Clients that offer ECHO then authenticate the connection with the public name, as follows: * If the server resumed a session or negotiated a session that did not use a certificate for authentication, the client MUST abort the connection with an "illegal_parameter" alert. This case is invalid because Section 7.1 requires the client to only offer ECHO-established sessions, and Section 8 requires the server to decline ECHO-established sessions if it did not accept ECHO. * The client MUST verify that the certificate is valid for ECHOConfigContents.public_name. If invalid, it MUST abort the connection with the appropriate alert. * If the server requests a client certificate, the client MUST respond with an empty Certificate message, denoting no client certificate. Note that authenticating a connection for the public name does not authenticate it for the origin. The TLS implementation MUST NOT report such connections as successful to the application. It additionally MUST ignore all session tickets and session IDs presented by the server. These connections are only used to trigger retries, as described in Section 7.2. This may be implemented, for instance, by reporting a failed connection with a dedicated error code. Rescorla, et al. Expires 10 September 2020 [Page 13] Internet-Draft TLS 1.3 SNI Encryption March 2020 7.2.3. HelloRetryRequest If the server sends a HelloRetryRequest in response to the ClientHello, the client sends a second updated ClientHello per the rules in [RFC8446]. However, at this point, the client does not know whether the server processed ClientHelloOuter or ClientHelloInner, and MUST regenerate both values to be acceptable. Note: if the inner and outer ClientHellos use different groups for their key shares or differ in some other way, then the HelloRetryRequest may actually be invalid for one or the other ClientHello, in which case a fresh ClientHello MUST be generated, ignoring the instructions in HelloRetryRequest. Otherwise, the usual rules for HelloRetryRequest processing apply. Clients bind encryption of the second ClientHelloInner to encryption of the first ClientHelloInner via the derived echo_hrr_key by modifying HPKE setup as follows: pkR = HPKE.KEM.Unmarshal(ECHOConfig.public_key) enc, context = SetupPSKS(pkR, "tls13-echo-hrr", echo_hrr_key, "") echo_nonce = context.Export("tls13-echo-hrr-nonce", 16) Clients then encrypt the second ClientHelloInner using this new HPKE context. In doing so, the encrypted value is also authenticated by echo_hrr_key. Client-facing servers perform the corresponding process when decrypting second ClientHelloInner messages. In particular, upon receipt of a second ClientHello message with a ClientEncryptedCH value, servers setup their HPKE context and decrypt ClientEncryptedCH as follows: context = SetupPSKR(ClientEncryptedCH.enc, skR, "tls13-echo-hrr", echo_hrr_key, "") ClientHelloInner = context.Open("", ClientEncryptedCH.encrypted_ch) echo_nonce = context.Export("tls13-echo-hrr-nonce", 16) [[OPEN ISSUE: Should we be using the PSK input or the info input? On the one hand, the requirements on info seem weaker, but maybe actually this needs to be secret? Analysis needed.]] [[OPEN ISSUE: This, along with trial decryption is pretty gross. It would just be a lot easier if we were willing to have the server indicate whether ECHO had been accepted or not. Given that the server is supposed to only reject ECHO when it doesn't know the key, and this is easy to probe for, can we just instead have an extension to indicate what has happened.]] Rescorla, et al. Expires 10 September 2020 [Page 14] Internet-Draft TLS 1.3 SNI Encryption March 2020 7.3. GREASE extensions If the client attempts to connect to a server and does not have an ECHOConfig structure available for the server, it SHOULD send a GREASE [I-D.ietf-tls-grease] "encrypted_client_hello" extension as follows: * Select a supported cipher suite, named group, and padded_length value. The padded_length value SHOULD be 260 (sum of the maximum DNS name length and TLS encoding overhead) or a multiple of 16 less than 260. Set the "suite" field to the selected cipher suite. These selections SHOULD vary to exercise all supported configurations, but MAY be held constant for successive connections to the same server in the same session. * Set the "enc" field to a randomly-generated valid encapsulated public key output by the HPKE KEM. * Set the "record_digest" field to a randomly-generated string of hash_length bytes, where hash_length is the length of the hash function associated with the chosen cipher suite. * Set the "encrypted_client_hello" field to a randomly-generated string of [TODO] bytes. If the server sends an "encrypted_client_hello" extension, the client MUST check the extension syntactically and abort the connection with a "decode_error" alert if it is invalid. Offering a GREASE extension is not considered offering an encrypted ClientHello for purposes of requirements in Section 7. In particular, the client MAY offer to resume sessions established without ECHO. 8. Client-Facing Server Behavior Upon receiving an "encrypted_client_hello" extension, the client- facing server MUST check that it is able to negotiate TLS 1.3 or greater. If not, it MUST abort the connection with a "handshake_failure" alert. The ClientEncryptedCH value is said to match a known ECHOConfig if there exists an ECHOConfig that can be used to successfully decrypt ClientEncryptedCH.encrypted_ch. This matching procedure should be done using one of the following two checks: 1. Compare ClientEncryptedCH.record_digest against cryptographic hashes of known ECHOConfig and choose the one that matches. Rescorla, et al. Expires 10 September 2020 [Page 15] Internet-Draft TLS 1.3 SNI Encryption March 2020 2. Use trial decryption of ClientEncryptedCH.encrypted_ch with known ECHOConfig and choose the one that succeeds. Some uses of ECHO, such as local discovery mode, may omit the ClientEncryptedCH.record_digest since it can be used as a tracking vector. In such cases, trial decryption should be used for matching ClientEncryptedCH to known ECHOConfig. Unless specified by the application using (D)TLS or externally configured on both sides, implementations MUST use the first method. If the ClientEncryptedCH value does not match any known ECHOConfig structure, it MUST ignore the extension and proceed with the connection, with the following added behavior: * It MUST include the "encrypted_client_hello" extension with the "retry_keys" field set to one or more ECHOConfig structures with up-to-date keys. Servers MAY supply multiple ECHOConfig values of different versions. This allows a server to support multiple versions at once. * The server MUST ignore all PSK identities in the ClientHello which correspond to ECHO PSKs. ECHO PSKs offered by the client are associated with the ECHO name. The server was unable to decrypt then ECHO name, so it should not resume them when using the cleartext SNI name. This restriction allows a client to reject resumptions in Section 7.2.2.1. Note that an unrecognized ClientEncryptedCH.record_digest value may be a GREASE ECHO extension (see Section 7.3), so it is necessary for servers to proceed with the connection and rely on the client to abort if ECHO was required. In particular, the unrecognized value alone does not indicate a misconfigured ECHO advertisement (Section 9.1). Instead, servers can measure occurrences of the "echo_required" alert to detect this case. If the ClientEncryptedCH value does match a known ECHOConfig, the server then decrypts ClientEncryptedCH.encrypted_ch, using the private key skR corresponding to ESNIConfig, as follows: context = SetupBaseR(ClientEncryptedCH.enc, skR, "tls13-echo") ClientHelloInner = context.Open("", ClientEncryptedCH.encrypted_ch) echo_nonce = context.Export("tls13-echo-nonce", 16) echo_hrr_key = context.Export("tls13-echo-hrr-key", 16) If decryption fails, the server MUST abort the connection with a "decrypt_error" alert. Moreover, if there is no "echo_nonce" extension, or if its value does not match the derived echo_nonce, the server MUST abort the connection with a "decrypt_error" alert. Next, Rescorla, et al. Expires 10 September 2020 [Page 16] Internet-Draft TLS 1.3 SNI Encryption March 2020 the server MUST scan ClientHelloInner for any "outer_extension" extensions and substitute their values with the values in ClientHelloOuter. It MUST first verify that the hash found in the extension matches the hash of the extension to be interpolated in and if it does not, abort the connections with a "decrypt_error" alert. Upon determining the true SNI, the client-facing server then either serves the connection directly (if in Shared Mode), in which case it executes the steps in the following section, or forwards the TLS connection to the backend server (if in Split Mode). In the latter case, it does not make any changes to the TLS messages, but just blindly forwards them. If the server sends a NewSessionTicket message, the corresponding ECHO PSK MUST be ignored by all other servers in the deployment when not negotiating ECHO, including servers which do not implement this specification. 9. Compatibility Issues Unlike most TLS extensions, placing the SNI value in an ECHO extension is not interoperable with existing servers, which expect the value in the existing cleartext extension. Thus server operators SHOULD ensure servers understand a given set of ECHO keys before advertising them. Additionally, servers SHOULD retain support for any previously-advertised keys for the duration of their validity However, in more complex deployment scenarios, this may be difficult to fully guarantee. Thus this protocol was designed to be robust in case of inconsistencies between systems that advertise ECHO keys and servers, at the cost of extra round-trips due to a retry. Two specific scenarios are detailed below. 9.1. Misconfiguration and Deployment Concerns It is possible for ECHO advertisements and servers to become inconsistent. This may occur, for instance, from DNS misconfiguration, caching issues, or an incomplete rollout in a multi-server deployment. This may also occur if a server loses its ECHO keys, or if a deployment of ECHO must be rolled back on the server. The retry mechanism repairs inconsistencies, provided the server is authoritative for the public name. If server and advertised keys mismatch, the server will respond with echo_retry_requested. If the server does not understand the "encrypted_server_name" extension at all, it will ignore it as required by [RFC8446]; Section 4.1.2. Provided the server can present a certificate valid for the public Rescorla, et al. Expires 10 September 2020 [Page 17] Internet-Draft TLS 1.3 SNI Encryption March 2020 name, the client can safely retry with updated settings, as described in Section 7.2. Unless ECHO is disabled as a result of successfully establishing a connection to the public name, the client MUST NOT fall back to using unencrypted ClientHellos, as this allows a network attacker to disclose the contents of this ClientHello, including the SNI. It MAY attempt to use another server from the DNS results, if one is provided. Client-facing servers with non-uniform cryptographic configurations across backend origin servers segment the ECHO anonymity set based on these configurations. For example, if a client-facing server hosts k backend origin servers, and exactly one of those backend origin servers supports a different set of cryptographic algorithms than the other (k - 1) servers, it may be possible to identify this single server based on the contents of the ServerHello as this message is not encrypted. 9.2. Middleboxes A more serious problem is MITM proxies which do not support this extension. [RFC8446]; Section 9.3 requires that such proxies remove any extensions they do not understand. The handshake will then present a certificate based on the public name, without echoing the "encrypted_server_name" extension to the client. Depending on whether the client is configured to accept the proxy's certificate as authoritative for the public name, this may trigger the retry logic described in Section 7.2 or result in a connection failure. A proxy which is not authoritative for the public name cannot forge a signal to disable ECHO. A non-conformant MITM proxy which instead forwards the ECHO extension, substituting its own KeyShare value, will result in the client-facing server recognizing the key, but failing to decrypt the SNI. This causes a hard failure. Clients SHOULD NOT attempt to repair the connection in this case. 10. TLS and HPKE CipherSuite Mapping Per {{RFC8446}, TLS ciphersuites define an AEAD and hash algorithm. In contrast, HPKE defines separate AEAD algorithms and key derivation functions. The table below lists the mapping between ciphersuites and HPKE identifiers. TLS_AES_128_CCM_SHA256 and TLS_AES_128_CCM_8_SHA256 are not supported ECHO ciphersuites as they have no HPKE equivalent. Rescorla, et al. Expires 10 September 2020 [Page 18] Internet-Draft TLS 1.3 SNI Encryption March 2020 +------------------------------+------------------+-------------+ | TLS CipherSuite | HPKE AEAD | HPKE KDF | +==============================+==================+=============+ | TLS_AES_128_GCM_SHA256 | AES-GCM-128 | HKDF-SHA256 | +------------------------------+------------------+-------------+ | TLS_AES_256_GCM_SHA384 | AES-GCM-256 | HKDF-SHA256 | +------------------------------+------------------+-------------+ | TLS_CHACHA20_POLY1305_SHA256 | ChaCha20Poly1305 | HKDF-SHA256 | +------------------------------+------------------+-------------+ Table 1 11. Security Considerations 11.1. Why is cleartext DNS OK? In comparison to [I-D.kazuho-protected-sni], wherein DNS Resource Records are signed via a server private key, ECHO records have no authenticity or provenance information. This means that any attacker which can inject DNS responses or poison DNS caches, which is a common scenario in client access networks, can supply clients with fake ECHO records (so that the client encrypts data to them) or strip the ECHO record from the response. However, in the face of an attacker that controls DNS, no encryption scheme can work because the attacker can replace the IP address, thus blocking client connections, or substituting a unique IP address which is 1:1 with the DNS name that was looked up (modulo DNS wildcards). Thus, allowing the ECHO records in the clear does not make the situation significantly worse. Clearly, DNSSEC (if the client validates and hard fails) is a defense against this form of attack, but DoH/DPRIVE are also defenses against DNS attacks by attackers on the local network, which is a common case where ClientHello and SNI encryption are desired. Moreover, as noted in the introduction, SNI encryption is less useful without encryption of DNS queries in transit via DoH or DPRIVE mechanisms. 11.2. Optional Record Digests and Trial Decryption Supporting optional record digests and trial decryption opens oneself up to DoS attacks. Specifically, an adversary may send malicious ClientHello messages, i.e., those which will not decrypt with any known ECHO key, in order to force decryption. Servers that support this feature should, for example, implement some form of rate limiting mechanism to limit the damage caused by such attacks. Rescorla, et al. Expires 10 September 2020 [Page 19] Internet-Draft TLS 1.3 SNI Encryption March 2020 11.3. Related Privacy Leaks ECHO requires encrypted DNS to be an effective privacy protection mechanism. However, verifying the server's identity from the Certificate message, particularly when using the X509 CertificateType, may result in additional network traffic that may reveal the server identity. Examples of this traffic may include requests for revocation information, such as OCSP or CRL traffic, or requests for repository information, such as authorityInformationAccess. It may also include implementation- specific traffic for additional information sources as part of verification. Implementations SHOULD avoid leaking information that may identify the server. Even when sent over an encrypted transport, such requests may result in indirect exposure of the server's identity, such as indicating a specific CA or service being used. To mitigate this risk, servers SHOULD deliver such information in-band when possible, such as through the use of OCSP stapling, and clients SHOULD take steps to minimize or protect such requests during certificate validation. 11.4. Comparison Against Criteria [I-D.ietf-tls-sni-encryption] lists several requirements for SNI encryption. In this section, we re-iterate these requirements and assess the ECHO design against them. 11.4.1. Mitigate against replay attacks Since servers process either ClientHelloInner or ClientHelloOuter, and ClientHelloInner contains an HPKE-derived nonce, it is not possible for an attacker to "cut and paste" the ECHO value in a different Client Hello and learn information from ClientHelloInner. This is because the attacker lacks access to the HPKE-derived nonce used to derive the handshake secrets. 11.4.2. Avoid widely-deployed shared secrets This design depends upon DNS as a vehicle for semi-static public key distribution. Server operators may partition their private keys however they see fit provided each server behind an IP address has the corresponding private key to decrypt a key. Thus, when one ECHO key is provided, sharing is optimally bound by the number of hosts that share an IP address. Server operators may further limit sharing by publishing different DNS records containing ECHOConfig values with different keys using a short TTL. Rescorla, et al. Expires 10 September 2020 [Page 20] Internet-Draft TLS 1.3 SNI Encryption March 2020 11.4.3. Prevent SNI-based DoS attacks This design requires servers to decrypt ClientHello messages with ClientEncryptedCH extensions carrying valid digests. Thus, it is possible for an attacker to force decryption operations on the server. This attack is bound by the number of valid TCP connections an attacker can open. 11.4.4. Do not stick out As more clients enable ECHO support, e.g., as normal part of Web browser functionality, with keys supplied by shared hosting providers, the presence of ECHO extensions becomes less suspicious and part of common or predictable client behavior. In other words, if all Web browsers start using ECHO, the presence of this value does not signal suspicious behavior to passive eavesdroppers. Additionally, this specification allows for clients to send GREASE ECHO extensions (see Section 7.3), which helps ensure the ecosystem handles the values correctly. 11.4.5. Forward secrecy This design is not forward secret because the server's ECHO key is static. However, the window of exposure is bound by the key lifetime. It is RECOMMENDED that servers rotate keys frequently. 11.4.6. Proper security context This design permits servers operating in Split Mode to forward connections directly to backend origin servers, thereby avoiding unnecessary MiTM attacks. 11.4.7. Split server spoofing Assuming ECHO records retrieved from DNS are authenticated, e.g., via DNSSEC or fetched from a trusted Recursive Resolver, spoofing a server operating in Split Mode is not possible. See Section 11.1 for more details regarding cleartext DNS. Authenticating the ECHOConfigs structure naturally authenticates the included public name. This also authenticates any retry signals from the server because the client validates the server certificate against the public name before retrying. Rescorla, et al. Expires 10 September 2020 [Page 21] Internet-Draft TLS 1.3 SNI Encryption March 2020 11.4.8. Supporting multiple protocols This design has no impact on application layer protocol negotiation. It may affect connection routing, server certificate selection, and client certificate verification. Thus, it is compatible with multiple protocols. 11.5. Misrouting Note that the backend server has no way of knowing what the SNI was, but that does not lead to additional privacy exposure because the backend server also only has one identity. This does, however, change the situation slightly in that the backend server might previously have checked SNI and now cannot (and an attacker can route a connection with an encrypted SNI to any backend server and the TLS connection will still complete). However, the client is still responsible for verifying the server's identity in its certificate. [[TODO: Some more analysis needed in this case, as it is a little odd, and probably some precise rules about handling ECHO and no SNI uniformly?]] 12. IANA Considerations 12.1. Update of the TLS ExtensionType Registry IANA is requested to create an entry, encrypted_server_name(0xffce), in the existing registry for ExtensionType (defined in [RFC8446]), with "TLS 1.3" column values being set to "CH, EE", and "Recommended" column being set to "Yes". 12.2. Update of the TLS Alert Registry IANA is requested to create an entry, echo_required(121) in the existing registry for Alerts (defined in [RFC8446]), with the "DTLS- OK" column being set to "Y". 13. References 13.1. Normative References [HTTPSSVC] Schwartz, B., Bishop, M., and E. Nygren, "Service binding and parameter specification via the DNS (DNS SVCB and HTTPSSVC)", Work in Progress, Internet-Draft, draft- nygren-dnsop-svcb-httpssvc-00, 23 September 2019, . Rescorla, et al. Expires 10 September 2020 [Page 22] Internet-Draft TLS 1.3 SNI Encryption March 2020 [I-D.ietf-tls-exported-authenticator] Sullivan, N., "Exported Authenticators in TLS", Work in Progress, Internet-Draft, draft-ietf-tls-exported- authenticator-11, 18 December 2019, . [I-D.irtf-cfrg-hpke] Barnes, R. and K. Bhargavan, "Hybrid Public Key Encryption", Work in Progress, Internet-Draft, draft-irtf- cfrg-hpke-02, 4 November 2019, . [RFC1035] Mockapetris, P.V., "Domain names - implementation and specification", STD 13, RFC 1035, DOI 10.17487/RFC1035, November 1987, . [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate Requirement Levels", BCP 14, RFC 2119, DOI 10.17487/RFC2119, March 1997, . [RFC6066] Eastlake 3rd, D., "Transport Layer Security (TLS) Extensions: Extension Definitions", RFC 6066, DOI 10.17487/RFC6066, January 2011, . [RFC6234] Eastlake 3rd, D. and T. Hansen, "US Secure Hash Algorithms (SHA and SHA-based HMAC and HKDF)", RFC 6234, DOI 10.17487/RFC6234, May 2011, . [RFC7685] Langley, A., "A Transport Layer Security (TLS) ClientHello Padding Extension", RFC 7685, DOI 10.17487/RFC7685, October 2015, . [RFC7918] Langley, A., Modadugu, N., and B. Moeller, "Transport Layer Security (TLS) False Start", RFC 7918, DOI 10.17487/RFC7918, August 2016, . [RFC7924] Santesson, S. and H. Tschofenig, "Transport Layer Security (TLS) Cached Information Extension", RFC 7924, DOI 10.17487/RFC7924, July 2016, . [RFC8174] Leiba, B., "Ambiguity of Uppercase vs Lowercase in RFC Rescorla, et al. Expires 10 September 2020 [Page 23] Internet-Draft TLS 1.3 SNI Encryption March 2020 2119 Key Words", BCP 14, RFC 8174, DOI 10.17487/RFC8174, May 2017, . [RFC8446] Rescorla, E., "The Transport Layer Security (TLS) Protocol Version 1.3", RFC 8446, DOI 10.17487/RFC8446, August 2018, . 13.2. Informative References [I-D.ietf-doh-dns-over-https] Hoffman, P. and P. McManus, "DNS Queries over HTTPS (DoH)", Work in Progress, Internet-Draft, draft-ietf-doh- dns-over-https-14, 16 August 2018, . [I-D.ietf-tls-grease] Benjamin, D., "Applying GREASE to TLS Extensibility", Work in Progress, Internet-Draft, draft-ietf-tls-grease-04, 22 August 2019, . [I-D.ietf-tls-sni-encryption] Huitema, C. and E. Rescorla, "Issues and Requirements for SNI Encryption in TLS", Work in Progress, Internet-Draft, draft-ietf-tls-sni-encryption-09, 28 October 2019, . [I-D.kazuho-protected-sni] Oku, K., "TLS Extensions for Protecting SNI", Work in Progress, Internet-Draft, draft-kazuho-protected-sni-00, 18 July 2017, . [RFC7301] Friedl, S., Popov, A., Langley, A., and E. Stephan, "Transport Layer Security (TLS) Application-Layer Protocol Negotiation Extension", RFC 7301, DOI 10.17487/RFC7301, July 2014, . [RFC7858] Hu, Z., Zhu, L., Heidemann, J., Mankin, A., Wessels, D., and P. Hoffman, "Specification for DNS over Transport Layer Security (TLS)", RFC 7858, DOI 10.17487/RFC7858, May 2016, . [RFC8094] Reddy, T., Wing, D., and P. Patil, "DNS over Datagram Transport Layer Security (DTLS)", RFC 8094, DOI 10.17487/RFC8094, February 2017, . Rescorla, et al. Expires 10 September 2020 [Page 24] Internet-Draft TLS 1.3 SNI Encryption March 2020 [SNIExtensibilityFailed] "Accepting that other SNI name types will never work", March 2016, . Appendix A. Alternative SNI Protection Designs Alternative approaches to encrypted SNI may be implemented at the TLS or application layer. In this section we describe several alternatives and discuss drawbacks in comparison to the design in this document. A.1. TLS-layer A.1.1. TLS in Early Data In this variant, TLS Client Hellos are tunneled within early data payloads belonging to outer TLS connections established with the client-facing server. This requires clients to have established a previous session --- and obtained PSKs --- with the server. The client-facing server decrypts early data payloads to uncover Client Hellos destined for the backend server, and forwards them onwards as necessary. Afterwards, all records to and from backend servers are forwarded by the client-facing server - unmodified. This avoids double encryption of TLS records. Problems with this approach are: (1) servers may not always be able to distinguish inner Client Hellos from legitimate application data, (2) nested 0-RTT data may not function correctly, (3) 0-RTT data may not be supported - especially under DoS - leading to availability concerns, and (4) clients must bootstrap tunnels (sessions), costing an additional round trip and potentially revealing the SNI during the initial connection. In contrast, encrypted SNI protects the SNI in a distinct Client Hello extension and neither abuses early data nor requires a bootstrapping connection. A.1.2. Combined Tickets In this variant, client-facing and backend servers coordinate to produce "combined tickets" that are consumable by both. Clients offer combined tickets to client-facing servers. The latter parse them to determine the correct backend server to which the Client Hello should be forwarded. This approach is problematic due to non- trivial coordination between client-facing and backend servers for ticket construction and consumption. Moreover, it requires a bootstrapping step similar to that of the previous variant. In contrast, encrypted SNI requires no such coordination. Rescorla, et al. Expires 10 September 2020 [Page 25] Internet-Draft TLS 1.3 SNI Encryption March 2020 A.2. Application-layer A.2.1. HTTP/2 CERTIFICATE Frames In this variant, clients request secondary certificates with CERTIFICATE_REQUEST HTTP/2 frames after TLS connection completion. In response, servers supply certificates via TLS exported authenticators [I-D.ietf-tls-exported-authenticator] in CERTIFICATE frames. Clients use a generic SNI for the underlying client-facing server TLS connection. Problems with this approach include: (1) one additional round trip before peer authentication, (2) non-trivial application-layer dependencies and interaction, and (3) obtaining the generic SNI to bootstrap the connection. In contrast, encrypted SNI induces no additional round trip and operates below the application layer. Appendix B. Total Client Hello Encryption The design described here only provides encryption for the SNI, but not for other extensions, such as ALPN. Another potential design would be to encrypt all of the extensions using the same basic structure as we use here for ECHO. That design has the following advantages: * It protects all the extensions from ordinary eavesdroppers * If the encrypted block has its own KeyShare, it does not necessarily require the client to use a single KeyShare, because the client's share is bound to the SNI by the AEAD (analysis needed). It also has the following disadvantages: * The client-facing server can still see the other extensions. By contrast we could introduce another EncryptedExtensions block that was encrypted to the backend server and not the client-facing server. * It requires a mechanism for the client-facing server to provide the extension-encryption key to the backend server and thus cannot be used with an unmodified backend server. * A conforming middlebox will strip every extension, which might result in a ClientHello which is just unacceptable to the server (more analysis needed). Rescorla, et al. Expires 10 September 2020 [Page 26] Internet-Draft TLS 1.3 SNI Encryption March 2020 Appendix C. Acknowledgements This document draws extensively from ideas in [I-D.kazuho-protected-sni], but is a much more limited mechanism because it depends on the DNS for the protection of the ECHO key. Richard Barnes, Christian Huitema, Patrick McManus, Matthew Prince, Nick Sullivan, Martin Thomson, and David Benjamin also provided important ideas and contributions. Authors' Addresses Eric Rescorla RTFM, Inc. Email: ekr@rtfm.com Kazuho Oku Fastly Email: kazuhooku@gmail.com Nick Sullivan Cloudflare Email: nick@cloudflare.com Christopher A. Wood Apple, Inc. Email: cawood@apple.com Rescorla, et al. Expires 10 September 2020 [Page 27]