Independent Submission                                        Y. Sheffer
Request for Comments: 8672                                        Intuit
Category: Experimental                                        D. Migault
ISSN: 2070-1721                                                 Ericsson
                                                            October 2019

                TLS Server Identity Pinning with Tickets

Abstract

   Misissued public-key certificates can prevent TLS clients from
   appropriately authenticating the TLS server.  Several alternatives
   have been proposed to detect this situation and prevent a client from
   establishing a TLS session with a TLS end point authenticated with an
   illegitimate public-key certificate.  These mechanisms are either not
   widely deployed or limited to public web browsing.

   This document proposes experimental extensions to TLS with opaque
   pinning tickets as a way to pin the server's identity.  During an
   initial TLS session, the server provides an original encrypted
   pinning ticket.  In subsequent TLS session establishment, upon
   receipt of the pinning ticket, the server proves its ability to
   decrypt the pinning ticket and thus the ownership of the pinning
   protection key.  The client can now safely conclude that the TLS
   session is established with the same TLS server as the original TLS
   session.  One of the important properties of this proposal is that no
   manual management actions are required.

Status of This Memo

   This document is not an Internet Standards Track specification; it is
   published for examination, experimental implementation, and
   evaluation.

   This document defines an Experimental Protocol for the Internet
   community.  This is a contribution to the RFC Series, independently
   of any other RFC stream.  The RFC Editor has chosen to publish this
   document at its discretion and makes no statement about its value for
   implementation or deployment.  Documents approved for publication by
   the RFC Editor are not candidates for any level of Internet Standard;
   see Section 2 of RFC 7841.

   Information about the current status of this document, any errata,
   and how to provide feedback on it may be obtained at
   https://www.rfc-editor.org/info/rfc8672.

Copyright Notice

   Copyright (c) 2019 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.

Table of Contents

   1.  Introduction
     1.1.  Conventions Used in This Document
     1.2.  Scope of Experimentation
   2.  Protocol Overview
     2.1.  Initial Connection
     2.2.  Subsequent Connections
     2.3.  Indexing the Pins
   3.  Message Definitions
   4.  Cryptographic Operations
     4.1.  Pinning Secret
     4.2.  Pinning Ticket
     4.3.  Pinning Protection Key
     4.4.  Pinning Proof
   5.  Operational Considerations
     5.1.  Protection Key Synchronization
     5.2.  Ticket Lifetime
     5.3.  Certificate Renewal
     5.4.  Certificate Revocation
     5.5.  Disabling Pinning
     5.6.  Server Compromise
     5.7.  Disaster Recovery
   6.  Security Considerations
     6.1.  Trust-on-First-Use (TOFU) and MITM Attacks
     6.2.  Pervasive Monitoring
     6.3.  Server-Side Error Detection
     6.4.  Client Policy and SSL Proxies
     6.5.  Client-Side Error Behavior
     6.6.  Stolen and Forged Tickets
     6.7.  Client Privacy
     6.8.  Ticket Protection Key Management
   7.  IANA Considerations
   8.  References
     8.1.  Normative References
     8.2.  Informative References
   Appendix A.  Previous Work
     A.1.  Comparison: HPKP
     A.2.  Comparison: TACK
   Acknowledgments
   Authors' Addresses

1.  Introduction

   Misissued public-key certificates can prevent TLS [RFC8446] clients
   from appropriately authenticating the TLS server.  This is a
   significant risk in the context of the global public key
   infrastructure (PKI), and similarly for large-scale deployments of
   certificates within enterprises.

   This document proposes experimental extensions to TLS with opaque
   pinning tickets as a way to pin the server's identity.  The approach
   is intended to be easy to implement and deploy, and reuses some of
   the ideas behind TLS session resumption [RFC5077].

   Ticket pinning is a second-factor server authentication method and is
   not proposed as a substitute for the authentication method provided
   in the TLS key exchange.  More specifically, the client only uses the
   pinning identity method after the TLS key exchange is successfully
   completed.  In other words, the pinning identity method is only
   performed over an authenticated TLS session.  Note that ticket
   pinning does not pin certificate information and therefore is truly
   an independent second-factor authentication.

   Ticket pinning is a trust-on-first-use (TOFU) mechanism, in that the
   first server authentication is only based on PKI certificate
   validation, but for any follow-on sessions, the client is further
   ensuring the server's identity based on the server's ability to
   decrypt the ticket, in addition to normal PKI certificate
   authentication.

   During initial TLS session establishment, the client requests a
   pinning ticket from the server.  Upon receiving the request the
   server generates a pinning secret that is expected to be
   unpredictable for peers other than the client or the server.  In our
   case, the pinning secret is generated from parameters exchanged
   during the TLS key exchange, so client and server can generate it
   locally and independently.  The server constructs the pinning ticket
   with the necessary information to retrieve the pinning secret.  The
   server then encrypts the ticket and returns the pinning ticket to the
   client with an associated pinning lifetime.

   The pinning lifetime value indicates for how long the server promises
   to retain the server-side ticket-encryption key, which allows it to
   complete the protocol exchange correctly and prove its identity.  The
   server commitment (and ticket lifetime) is typically on the order of
   weeks.

   Once the key exchange is completed, and the server is deemed
   authenticated, the client generates locally the pinning secret and
   caches the server's identifiers to index the pinning secret as well
   as the pinning ticket and its associated lifetime.

   When the client reestablishes a new TLS session with the server, it
   sends the pinning ticket to the server.  Upon receiving it, the
   server returns a proof of knowledge of the pinning secret.  Once the
   key exchange is completed, and the server has been authenticated, the
   client checks the pinning proof returned by the server using the
   client's stored pinning secret.  If the proof matches, the client can
   conclude that the server to which it is currently connecting is, in
   fact, the correct server.

   This document only applies to TLS 1.3.  We believe that the idea can
   also be retrofitted into earlier versions of the protocol, but this
   would require significant changes.  One example is that TLS 1.2
   [RFC5246] and earlier versions do not provide a generic facility of
   encrypted handshake extensions, such as is used here to transport the
   ticket.

   The main advantages of this protocol over earlier pinning solutions
   are the following:

   *  The protocol is at the TLS level, and as a result is not
      restricted to HTTP at the application level.

   *  The protocol is robust to changes in server IP address,
      certification authority (CA), and public key.  The server is
      characterized by the ownership of the pinning protection key,
      which is never provided to the client.  Server configuration
      parameters such as the CA and the public key may change without
      affecting the pinning ticket protocol.

   *  Once a single parameter is configured (the ticket's lifetime),
      operation is fully automated.  The server administrator need not
      bother with the management of backup certificates or explicit
      pins.

   *  For server clusters, we reuse the existing infrastructure
      [RFC5077] where it exists.

   *  Pinning errors, presumably resulting from man-in-the-middle (MITM)
      attacks, can be detected both by the client and the server.  This
      allows for server-side detection of MITM attacks using large-scale
      analytics, and with no need to rely on clients to explicitly
      report the error.

   A note on terminology: unlike other solutions in this space, we do
   not do "certificate pinning" (or "public key pinning"), since the
   protocol is oblivious to the server's certificate.  We prefer the
   term "server identity pinning" for this new solution.  In our
   solution, the server proves its identity by generating a proof that
   it can read and decrypt an encrypted ticket.  As a result, the
   identity proof relies on proof of ownership of the pinning protection
   key.  However, this key is never exchanged with the client or known
   by it, and so cannot itself be pinned.

1.1.  Conventions Used in This Document

   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.

1.2.  Scope of Experimentation

   This document describes an experimental extension to the TLS
   protocol.  This section defines constraints on this experiment and
   how it can yield useful information, potentially resulting in a
   standard.

   The protocol is designed so that if the server does not support it,
   the client and server fall back to a normal TLS exchange, with the
   exception of a single PinningTicket extension being initially sent by
   the client.  In addition, the protocol is designed only to strengthen
   the validation of the server's identity ("second factor").  As a
   result, implementation or even protocol errors should not result in
   weakened security compared to the normal TLS exchange.  Given these
   two points, experimentation can be run on the open Internet between
   consenting client and server implementations.

   The goal of the experiment is to prove that:

   *  Non-supporting clients and servers are unaffected.

   *  Connectivity between supporting clients and servers is retained
      under normal circumstances, whether the client connects to the
      server frequently (relative to the ticket's lifetime) or very
      rarely.

   *  Enterprise middleboxes do not interrupt such connectivity.

   *  Misissued certificates and rogue TLS-aware middleboxes do result
      in broken connectivity, and these cases are detected on the client
      and/or server side.  Clients and servers can be recovered even
      after such events and the normal connectivity restored.

   Following two years of successful deployment, the authors will
   publish a document that summarizes the experiment's findings and will
   resubmit the protocol for consideration as a Proposed Standard.

2.  Protocol Overview

   The protocol consists of two phases: the first time a particular
   client connects to a server, and subsequent connections.

   This protocol supports full TLS handshakes, as well as 0-RTT
   handshakes.  Below we present it in the context of a full handshake,
   but behavior in 0-RTT handshakes should be identical.

   The document presents some similarities with the ticket resumption
   mechanism described in [RFC5077].  However the scope of this document
   differs from session resumption mechanisms implemented with [RFC5077]
   or with other mechanisms.  Specifically, the pinning ticket does not
   carry any state associated with a TLS session and thus cannot be used
   for session resumption or client authentication.  Instead, the
   pinning ticket only contains the encrypted pinning secret.  The
   pinning ticket is used by the server to prove its ability to decrypt
   it, which implies ownership of the pinning protection key.

   [RFC5077] has been obsoleted by [RFC8446], and ticket resumption is
   now defined by Section 2.2 of [RFC8446].  This document references
   [RFC5077] as an informational document since it contains a more
   thorough discussion of stateless ticket resumption, and because
   ticket resumption benefits from significant operational experience
   with TLS 1.2 that is still widely deployed at the time of writing.
   This experience, as well as deployment experience, can easily be re-
   used for identity pinning.

   With TLS 1.3, session resumption is based on a Pre-Shared Key (PSK).
   This is orthogonal to this protocol.  With TLS 1.3, a TLS session can
   be established using PKI and a pinning ticket, and later resumed with
   PSK.

   However, the protocol described in this document addresses the
   problem of misissued certificates.  Thus, it is not expected to be
   used outside a certificate-based TLS key exchange, such as in PSK.
   As a result, PSK handshakes MUST NOT include the extension defined
   here.

2.1.  Initial Connection

   When a client first connects to a server, it requests a pinning
   ticket by sending an empty PinningTicket extension, and receives it
   as part of the server's first response, in the returned PinningTicket
   extension.

    Client                                               Server

    ClientHello
      + key_share
      + signature_algorithms
      + PinningTicket         -------->
                                                    ServerHello
                                                    + key_share
                                          {EncryptedExtensions
                                               + PinningTicket}
                                          {CertificateRequest*}
                                                 {Certificate*}
                                           {CertificateVerify*}
                              <--------              {Finished}
    {Certificate*}
    {CertificateVerify*}
    {Finished}                -------->
    [Application Data]        <------->      [Application Data]

           *  Indicates optional or situation-dependent
              messages that are not always sent.

           {} Indicates messages protected using keys
              derived from the ephemeral secret.

           [] Indicates messages protected using keys
              derived from the master secret.

   If a client supports the PinningTicket extension and does not have
   any pinning ticket associated with the server, the exchange is
   considered as an initial connection.  Other reasons the client may
   not have a pinning ticket include the client having flushed its
   pinning ticket store, or the committed lifetime of the pinning ticket
   having expired.

   Upon receipt of the PinningTicket extension, the server computes a
   pinning secret (Section 4.1) and sends the pinning ticket
   (Section 4.2) encrypted with the pinning protection key
   (Section 4.3).  The pinning ticket is associated with a lifetime
   value by which the server assumes the responsibility of retaining the
   pinning protection key and being able to decrypt incoming pinning
   tickets during the period indicated by the committed lifetime.

   Once the pinning ticket has been generated, the server returns the
   pinning ticket and the committed lifetime in a PinningTicket
   extension embedded in the EncryptedExtensions message.  We note that
   a PinningTicket extension MUST NOT be sent as part of a
   HelloRetryRequest.

   Upon receiving the pinning ticket, the client MUST NOT accept it
   until the key exchange is completed and the server authenticated.  If
   the key exchange is not completed successfully, the client MUST
   ignore the received pinning ticket.  Otherwise, the client computes
   the pinning secret and SHOULD cache the pinning secret and the
   pinning ticket for the duration indicated by the pinning ticket
   lifetime.  The client SHOULD clean up the cached values at the end of
   the indicated lifetime.

2.2.  Subsequent Connections

   When the client initiates a connection to a server it has previously
   seen (see Section 2.3 on identifying servers), it SHOULD send the
   pinning ticket for that server.  The pinning ticket, pinning secret,
   and pinning ticket lifetime computed during the establishment of the
   previous TLS session are designated in this document as the
   "original" ones, to distinguish them from a new ticket that may be
   generated during the current session.

   The server MUST extract the original pinning_secret value from the
   ticket and MUST respond with a PinningTicket extension, which
   includes:

   *  A proof that the server can understand the ticket that was sent by
      the client; this proof also binds the pinning ticket to the
      server's (current) public key, as well as the ongoing TLS session.
      The proof is mandatory and MUST be included if a pinning ticket
      was sent by the client.

   *  A fresh pinning ticket.  The main reason for refreshing the ticket
      on each connection is privacy: to avoid the ticket serving as a
      fixed client identifier.  While a fresh pinning ticket might be of
      zero length, it is RECOMMENDED to include a fresh ticket with a
      nonzero length with each response.

   If the server cannot validate the received ticket, that might
   indicate an earlier MITM attack on this client.  The server MUST then
   abort the connection with a handshake_failure alert and SHOULD log
   this failure.

   The client MUST verify the proof, and if it fails to do so, the
   client MUST issue a handshake_failure alert and abort the connection
   (see also Section 6.5).  It is important that the client does not
   attempt to "fall back" by omitting the PinningTicket extension.

   When the connection is successfully set up, i.e., after the Finished
   message is verified, the client SHOULD store the new ticket along
   with the corresponding pinning_secret, replacing the original ticket.

   Although this is an extension, if the client already has a ticket for
   a server, the client MUST interpret a missing PinningTicket extension
   in the server's response as an attack, because of the server's prior
   commitment to respect the ticket.  The client MUST abort the
   connection in this case.  See also Section 5.5 on ramping down
   support for this extension.

2.3.  Indexing the Pins

   Each pin is associated with a set of identifiers that include, among
   others, hostname, protocol (TLS or DTLS), and port number.  In other
   words, the pin for port TCP/443 may be different from that for DTLS,
   or from the pin for port TCP/8443.  These identifiers are expected to
   be relevant to characterize the identity of the server as well as the
   establishing TLS session.  When a hostname is used, it MUST be the
   value sent inside the Server Name Indication (SNI) extension.  This
   definition is similar to the concept of a Web Origin [RFC6454], but
   does not assume the existence of a URL.

   The purpose of ticket pinning is to pin the server identity.  As a
   result, any information orthogonal to the server's identity MUST NOT
   be considered in indexing.  More particularly, IP addresses are
   ephemeral and forbidden in SNI, and therefore pins MUST NOT be
   associated with IP addresses.  Similarly, CA names or public keys
   associated with server MUST NOT be used for indexing as they may
   change over time.

3.  Message Definitions

   This section defines the format of the PinningTicket extension.  We
   follow the message notation of [RFC8446].

    opaque pinning_ticket<0..2^16-1>;

    opaque pinning_proof<0..2^8-1>;

    struct {
      select (Role) {
        case client:
          pinning_ticket ticket<0..2^16-1>; //omitted on 1st connection

        case server:
          pinning_proof proof<0..2^8-1>; //no proof on 1st connection
          pinning_ticket ticket<0..2^16-1>; //omitted on ramp down
          uint32 lifetime;
      }
   } PinningTicketExtension;

   ticket    a pinning ticket sent by the client or returned by the
             server.  The ticket is opaque to the client.  The extension
             MUST contain exactly 0 or 1 tickets.

   proof     a demonstration by the server that it understands the
             received ticket and therefore that it is in possession of
             the secret that was used to generate it originally.  The
             extension MUST contain exactly 0 or 1 proofs.

   lifetime  the duration (in seconds) that the server commits to accept
             offered tickets in the future.

4.  Cryptographic Operations

   This section provides details on the cryptographic operations
   performed by the protocol peers.

4.1.  Pinning Secret

   The pinning secret is generated locally by the client and the server,
   which means they must use the same inputs to generate it.  This value
   must be generated before the ServerHello message is sent, as the
   server includes the corresponding pinning ticket in the same flight
   as the ServerHello message.  In addition, the pinning secret must be
   unpredictable to any party other than the client and the server.

   The pinning secret is derived using the Derive-Secret function
   provided by TLS 1.3, described in Section 7.1 of [RFC8446].

   pinning secret = Derive-Secret(Handshake Secret, "pinning secret",
            ClientHello...ServerHello)

4.2.  Pinning Ticket

   The pinning ticket contains the pinning secret.  The pinning ticket
   is provided by the client to the server, which decrypts it in order
   to extract the pinning secret and responds with a pinning proof.  As
   a result, the characteristics of the pinning ticket are:

   *  Pinning tickets MUST be encrypted and integrity-protected using
      strong cryptographic algorithms.

   *  Pinning tickets MUST be protected with a long-term pinning
      protection key.

   *  Pinning tickets MUST include a pinning protection key ID or serial
      number as to enable the pinning protection key to be refreshed.

   *  The pinning ticket MAY include other information, in addition to
      the pinning secret.  When additional information is included, a
      careful review needs to be performed to evaluate its impact on
      privacy.

   The pinning ticket's format is not specified by this document, but a
   format similar to the one proposed by [RFC5077] is RECOMMENDED.

4.3.  Pinning Protection Key

   The pinning protection key is used only by the server and so remains
   server implementation specific.  [RFC5077] recommends the use of two
   keys, but when using Authenticated Encryption with Associated Data
   (AEAD) algorithms, only a single key is required.

   When a single server terminates TLS for multiple virtual servers
   using the SNI mechanism, it is strongly RECOMMENDED that the server
   use a separate protection key for each one of them, in order to allow
   migrating virtual servers between different servers while keeping
   pinning active.

   As noted in Section 5.1, if the server is actually a cluster of
   machines, the protection key MUST be synchronized between all the
   nodes that accept TLS connections to the same server name.  When
   [RFC5077] is deployed, an easy way to do it is to derive the
   protection key from the session-ticket protection key, which is
   already synchronized.  For example:

   pinning_protection_key = HKDF-Expand(resumption_protection_key,
                                 "pinning protection", L)

   Where resumption_protection_key is the ticket protection key defined
   in [RFC5077].  Both resumption_protection_key and
   pinning_protection_key are only used by the server.

   The above solution attempts to minimize code changes related to
   management of the resumption_protection_key.  The drawback is that
   this key would be used both to directly encrypt session tickets and
   to derive the pinning_protection_key, and such mixed usage of a
   single key is not in line with cryptographic best practices.  Where
   possible, it is RECOMMENDED that the resumption_protection_key be
   unrelated to the pinning_protection_key and that they are separately
   shared among the relevant servers.

4.4.  Pinning Proof

   The pinning proof is sent by the server to demonstrate that it has
   been able to decrypt the pinning ticket and to retrieve the pinning
   secret.  The proof must be unpredictable and must not be replayed.
   Similarly to the pinning ticket, the pinning proof is sent by the
   server in the ServerHello message.  In addition, it must not be
   possible for a MITM server with a fake certificate to obtain a
   pinning proof from the original server.

   In order to address these requirements, the pinning proof is bound to
   the TLS session as well as the public key of the server:

   pinning_proof_secret=Derive-Secret(Handshake Secret,
                "pinning proof 1", ClientHello...ServerHello)

   proof = HMAC(original_pinning_secret, "pinning proof 2" +
                pinning_proof_secret + Hash(server_public_key))

   where HMAC [RFC2104] uses the Hash algorithm that was negotiated in
   the handshake, and the same hash is also used over the server's
   public key.  The original_pinning_secret value refers to the secret
   value extracted from the ticket sent by the client, to distinguish it
   from a new pinning secret value that is possibly computed in the
   current exchange.  The server_public_key value is the DER
   representation of the public key, specifically the
   SubjectPublicKeyInfo structure as-is.

5.  Operational Considerations

   The main motivation behind the current protocol is to enable identity
   pinning without the need for manual operations.  Manual operations
   are susceptible to human error, and in the case of public key
   pinning, can easily result in "server bricking": the server becoming
   inaccessible to some or all of its users.  To achieve this goal,
   operations described in identity pinning are only performed within
   the current TLS session, and there is no dependence on any TLS
   configuration parameters such as CA identity or public keys.  As a
   result, configuration changes are unlikely to lead to desynchronized
   state between the client and the server.

5.1.  Protection Key Synchronization

   The only operational requirement when deploying this protocol is
   that, if the server is part of a cluster, protection keys (the keys
   used to encrypt tickets) MUST be synchronized between all cluster
   members.  The protocol is designed so that if resumption ticket
   protection keys [RFC5077] are already synchronized between cluster
   members, nothing more needs to be done.

   Moreover, synchronization does not need to be instantaneous, e.g.,
   protection keys can be distributed a few minutes or hours in advance
   of their rollover.  In such scenarios, each cluster member MUST be
   able to accept tickets protected with a new version of the protection
   key, even while it is still using an old version to generate keys.
   This ensures that, when a client receives a "new" ticket, it does not
   next hit a cluster member that still rejects this ticket.

   Misconfiguration can lead to the server's clock being off by a large
   amount of time.  Consider a case where a server's clock is
   misconfigured, for example, to be 1 year in the future, and the
   system is allowed to delete expired keys automatically.  The server
   will then delete many outstanding keys because they are now long
   expired and will end up rejecting valid tickets that are stored by
   clients.  Such a scenario could make the server inaccessible to a
   large number of clients.

   The decision to delete a key should at least consider the largest
   value of the ticket lifetime as well as the expected time
   desynchronization between the servers of the cluster and the time
   difference for distributing the new key among the different servers
   in the cluster.

5.2.  Ticket Lifetime

   The lifetime of the ticket is a commitment by the server to retain
   the ticket's corresponding protection key for this duration, so that
   the server can prove to the client that it knows the secret embedded
   in the ticket.  For production systems, the lifetime SHOULD be
   between 7 and 31 days.

5.3.  Certificate Renewal

   The protocol ensures that the client will continue speaking to the
   correct server even when the server's certificate is renewed.  In
   this sense, pinning is not associated with certificates, which is the
   reason we designate the protocol described in this document as
   "server identity pinning".

   Note that this property is not impacted by the use of the server's
   public key in the pinning proof because the scope of the public key
   used is only the current TLS session.

5.4.  Certificate Revocation

   The protocol is orthogonal to certificate validation in the sense
   that, if the server's certificate has been revoked or is invalid for
   some other reason, the client MUST refuse to connect to it regardless
   of any ticket-related behavior.

5.5.  Disabling Pinning

   A server implementing this protocol MUST have a "ramp down" mode of
   operation where:

   *  The server continues to accept valid pinning tickets and responds
      correctly with a proof.

   *  The server does not send back a new pinning ticket.

   After a while, no clients will hold valid tickets, and the feature
   may be disabled.  Note that clients that do not receive a new pinning
   ticket do not necessarily need to remove the original ticket.
   Instead, the client may keep using the ticket until its lifetime
   expires.  However, as detailed in Section 6.7, re-use of a ticket by
   the client may result in privacy concerns as the ticket value may be
   used to correlate TLS sessions.

   Issuing a new pinning ticket with a shorter lifetime would only delay
   the ramp down process, as the shorter lifetime can only affect
   clients that actually initiated a new connection.  Other clients
   would still see the original lifetime for their pinning tickets.

5.6.  Server Compromise

   If a server compromise is detected, the pinning protection key MUST
   be rotated immediately, but the server MUST still accept valid
   tickets that use the old, compromised key.  Clients that still hold
   old pinning tickets will remain vulnerable to MITM attacks, but those
   that connect to the correct server will immediately receive new
   tickets protected with the newly generated pinning protection key.

   The same procedure applies if the pinning protection key is
   compromised directly, e.g., if a backup copy is inadvertently made
   public.

5.7.  Disaster Recovery

   All web servers in production need to be backed up, so that they can
   be recovered if a disaster (including a malicious activity) ever
   wipes them out.  Backup often includes the certificate and its
   private key, which must be backed up securely.  The pinning secret,
   including earlier versions that are still being accepted, must be
   backed up regularly.  However since it is only used as an
   authentication second factor, it does not require the same level of
   confidentiality as the server's private key.

   Readers should note that [RFC5077] session resumption keys are more
   security sensitive and should normally not be backed up, but rather
   treated as ephemeral keys.  Even when servers derive pinning secrets
   from resumption keys (Section 4.1), they MUST NOT back up resumption
   keys.

6.  Security Considerations

   This section reviews several security aspects related to the proposed
   extension.

6.1.  Trust-on-First-Use (TOFU) and MITM Attacks

   This protocol is a trust-on-first-use protocol.  If a client
   initially connects to the "right" server, it will be protected
   against MITM attackers for the lifetime of each received ticket.  If
   it connects regularly (depending, of course, on the server-selected
   lifetime), it will stay constantly protected against fake
   certificates.

   However if it initially connects to an attacker, subsequent
   connections to the "right" server will fail.  Server operators might
   want to advise clients on how to remove corrupted pins, once such
   large-scale attacks are detected and remediated.

   The protocol is designed so that it is not vulnerable to an active
   MITM attacker who has real-time access to the original server.  The
   pinning proof includes a hash of the server's public key to ensure
   the client that the proof was in fact generated by the server with
   which it is initiating the connection.

6.2.  Pervasive Monitoring

   Some organizations, and even some countries, perform pervasive
   monitoring on their constituents [RFC7258].  This often takes the
   form of always-active SSL proxies.  Because of the TOFU property,
   this protocol does not provide any security in such cases.

   Pervasive monitoring may also result in privacy concerns detailed in
   Section 6.7.

6.3.  Server-Side Error Detection

   Uniquely, this protocol allows the server to detect clients that
   present incorrect tickets and therefore can be assumed to be victims
   of a MITM attack.  Server operators can use such cases as indications
   of ongoing attacks, similarly to fake certificate attacks that took
   place in a few countries in the past.

6.4.  Client Policy and SSL Proxies

   Like it or not, some clients are normally deployed behind an SSL
   proxy.  Similar to [RFC7469], it is acceptable to allow pinning to be
   disabled for some hosts according to local policy.  For example, a
   User Agent (UA) MAY disable pinning for hosts whose validated
   certificate chain terminates at a user-defined trust anchor, rather
   than a trust anchor built into the UA (or underlying platform).
   Moreover, a client MAY accept an empty PinningTicket extension from
   such hosts as a valid response.

6.5.  Client-Side Error Behavior

   When a client receives a malformed or empty PinningTicket extension
   from a pinned server, it MUST abort the handshake and handshake.  If the client
   retries the request, it MUST NOT retry
   with no omit the PinningTicket in the request. retry
   message.  Doing otherwise would expose the client to trivial fallback
   attacks, similar to those described in [RFC7507].

   However, this rule can negatively impact clients that move from
   behind SSL proxies into the open Internet, and vice versa, if the
   advice in Section 6.4 is not followed.  Therefore, it is RECOMMENDED
   that browser and library vendors provide a documented way to remove
   stored pins.

6.6.  Stolen and Forged Tickets

   An attacker gains no benefit from stealing pinning tickets, even in
   conjunction with other pinning parameters such as the associated
   pinning secret, since pinning tickets are used to secure the client
   rather than the server.  Similarly, it is useless to forge a ticket
   for a particular server.

6.7.  Client Privacy

   This protocol is designed so that an external attacker cannot link
   different requests to a single client, provided the client requests
   and receives a fresh ticket upon each connection.  This may be of
   concern particularly during ramp down, if the server does not provide
   a new ticket, and the client reuses the same ticket.  To reduce or
   avoid such privacy concerns, it is RECOMMENDED for the server to
   issue a fresh ticket with a reduced lifetime.  This would at least
   reduce the time period under in which the TLS session sessions of the client are
   correlated. can be
   linked.  The server MAY also issue tickets with a zero-second
   lifetime until it is confident all tickets are expired.

   On the other hand, the server to which the client is connecting can
   easily track the client.  This may be an issue when the client
   expects to connect to the server (e.g., a mail server) with multiple
   identities.  Implementations SHOULD allow the user to opt out of
   pinning, either in general or for particular servers.

   This document does not define the exact content of tickets.
   Including client-specific information in tickets would raise privacy
   concerns and is NOT RECOMMENDED.

6.8.  Ticket Protection Key Management

   While the ticket format is not mandated by this document, protecting
   the ticket using authenticated encryption is RECOMMENDED.  Some of
   the algorithms commonly used for authenticated encryption, e.g.,
   Galois/Counter Mode (GCM), are highly vulnerable to nonce reuse, and
   this problem is magnified in a cluster setting.  Therefore,
   implementations that choose AES-GCM or any AEAD equivalent MUST adopt
   one of these three alternatives:

   *  Partition the nonce namespace between cluster members and use
      monotonic counters on each member, e.g., by setting the nonce to
      the concatenation of the cluster member ID and an incremental
      counter.

   *  Generate random nonces but avoid the so-called birthday bound,
      i.e., never generate more than the maximum allowed number of
      encrypted tickets (2**64 for AES-128-GCM) for the same ticket
      pinning protection key.

   *  An alternative design that has been attributed to Karthik
      Bhargavan is as follows.  Start with a 128-bit master key K_master
      and then for each encryption, generate a 256-bit random nonce and
      compute: K = HKDF(K_master, Nonce || "key"), then N =
      HKDF(K_master, Nonce || "nonce").  Use these values to encrypt the
      ticket, AES-GCM(K, N, data).  This nonce should then be stored and
      transmitted with the ticket.

7.  IANA Considerations

   The IANA has allocated a TicketPinning extension value in the "TLS
   ExtensionType Values" registry.

   [RFC8447] defines the procedure, requirements, and the necessary
   information for the IANA to update the "TLS ExtensionType Values"
   registry [TLS-EXT].  The registration procedure is "Specification
   Required" [RFC8126].

   The TicketPinning extension is registered as follows.  (The extension
   is not limited to Private Use, and as such has its first byte in the
   range 0-254.)

   Value:  32

   Name:  ticket_pinning

   Recommended:  No

   TLS 1.3:  CH, EE (to indicate that the extension is present in
      ClientHello and EncryptedExtensions messages)

8.  References

8.1.  Normative References

   [RFC2119]  Bradner, S., "Key words for use in RFCs to Indicate
              Requirement Levels", BCP 14, RFC 2119,
              DOI 10.17487/RFC2119, March 1997,
              <https://www.rfc-editor.org/info/rfc2119>.

   [RFC8126]  Cotton, M., Leiba, B., and T. Narten, "Guidelines for
              Writing an IANA Considerations Section in RFCs", BCP 26,
              RFC 8126, DOI 10.17487/RFC8126, June 2017,
              <https://www.rfc-editor.org/info/rfc8126>.

   [RFC8174]  Leiba, B., "Ambiguity of Uppercase vs Lowercase in RFC
              2119 Key Words", BCP 14, RFC 8174, DOI 10.17487/RFC8174,
              May 2017, <https://www.rfc-editor.org/info/rfc8174>.

   [RFC8446]  Rescorla, E., "The Transport Layer Security (TLS) Protocol
              Version 1.3", RFC 8446, DOI 10.17487/RFC8446, August 2018,
              <https://www.rfc-editor.org/info/rfc8446>.

   [RFC8447]  Salowey, J. and S. Turner, "IANA Registry Updates for TLS
              and DTLS", RFC 8447, DOI 10.17487/RFC8447, August 2018,
              <https://www.rfc-editor.org/info/rfc8447>.

8.2.  Informative References

   [Netcraft] Mutton, P., "HTTP Public Key Pinning: You're doing it
              wrong!", March 2016,
              <https://news.netcraft.com/archives/2016/03/30/http-
              public-key-pinning-youre-doing-it-wrong.html>.

   [Oreo]     Berkman, O., Pinkas, B., and M. Yung, "Firm Grip
              Handshakes: A Tool for Bidirectional Vouching", Cryptology
              and Network Security pp. 142-157, 2012.

   [RFC2104]  Krawczyk, H., Bellare, M., and R. Canetti, "HMAC: Keyed-
              Hashing for Message Authentication", RFC 2104,
              DOI 10.17487/RFC2104, February 1997,
              <https://www.rfc-editor.org/info/rfc2104>.

   [RFC5077]  Salowey, J., Zhou, H., Eronen, P., and H. Tschofenig,
              "Transport Layer Security (TLS) Session Resumption without
              Server-Side State", RFC 5077, DOI 10.17487/RFC5077,
              January 2008, <https://www.rfc-editor.org/info/rfc5077>.

   [RFC5246]  Dierks, T. and E. Rescorla, "The Transport Layer Security
              (TLS) Protocol Version 1.2", RFC 5246,
              DOI 10.17487/RFC5246, August 2008,
              <https://www.rfc-editor.org/info/rfc5246>.

   [RFC6454]  Barth, A., "The Web Origin Concept", RFC 6454,
              DOI 10.17487/RFC6454, December 2011,
              <https://www.rfc-editor.org/info/rfc6454>.

   [RFC6962]  Laurie, B., Langley, A., and E. Kasper, "Certificate
              Transparency", RFC 6962, DOI 10.17487/RFC6962, June 2013,
              <https://www.rfc-editor.org/info/rfc6962>.

   [RFC7258]  Farrell, S. and H. Tschofenig, "Pervasive Monitoring Is an
              Attack", BCP 188, RFC 7258, DOI 10.17487/RFC7258, May
              2014, <https://www.rfc-editor.org/info/rfc7258>.

   [RFC7469]  Evans, C., Palmer, C., and R. Sleevi, "Public Key Pinning
              Extension for HTTP", RFC 7469, DOI 10.17487/RFC7469, April
              2015, <https://www.rfc-editor.org/info/rfc7469>.

   [RFC7507]  Moeller, B. and A. Langley, "TLS Fallback Signaling Cipher
              Suite Value (SCSV) for Preventing Protocol Downgrade
              Attacks", RFC 7507, DOI 10.17487/RFC7507, April 2015,
              <https://www.rfc-editor.org/info/rfc7507>.

   [RFC8555]  Barnes, R., Hoffman-Andrews, J., McCarney, D., and J.
              Kasten, "Automatic Certificate Management Environment
              (ACME)", RFC 8555, DOI 10.17487/RFC8555, March 2019,
              <https://www.rfc-editor.org/info/rfc8555>.

   [TLS-EXT]  IANA, "TLS Extension Type Value",
              <https://www.iana.org/assignments/tls-extensiontype-
              values/>.

   [TLS-TACK] Marlinspike, M., "Trust Assertions for Certificate Keys",
              Work in Progress, Internet-Draft, draft-perrin-tls-tack-
              02, 7 January 2013,
              <https://tools.ietf.org/html/draft-perrin-tls-tack-02>.

Appendix A.  Previous Work

   The global PKI system relies on the trust of a CA issuing
   certificates.  As a result, a corrupted trusted CA may issue a
   certificate for any organization without the organization's approval
   (a misissued or "fake" certificate), and use the certificate to
   impersonate the organization.  There are many attempts to resolve
   these weaknesses, including the Certificate Transparency (CT)
   protocol [RFC6962], HTTP Public Key Pinning (HPKP) [RFC7469], and
   Trust Assertions for Certificate Keys (TACK) [TLS-TACK].

   CT requires cooperation of a large portion of the hundreds of extant
   certificate authorities (CAs) before it can be used "for real", in
   enforcing mode.  It is noted that the relevant industry forum (CA/
   Browser Forum) is indeed pushing for such extensive adoption.
   However the public nature of CT often makes it inappropriate for
   enterprise use because many organizations are not willing to expose
   their internal infrastructure publicly.

   TACK has some similarities to the current proposal, but work on it
   seems to have stalled.  Appendix A.2 compares our proposal to TACK.

   HPKP is an IETF standard, but so far has proven hard to deploy.  HPKP
   pins (fixes) a public key, one of the public keys listed in the
   certificate chain.  As a result, HPKP needs to be coordinated with
   the certificate management process.  Certificate management impacts
   HPKP and thus increases the probability of HPKP failures.  This risk
   is made even higher given the fact that, even though work has been
   done in the Automated Certificate Management Environment (ACME)
   working group to automate certificate management, in many or even
   most cases, certificates are still managed manually.  As a result,
   HPKP cannot be completely automated, resulting in error-prone manual
   configuration.  Such errors could prevent the web server from being
   accessed by some clients.  In addition, HPKP uses an HTTP header,
   which makes this solution HTTPS specific and not generic to TLS.  On
   the other hand, the current document provides a solution that is
   independent of the server's certificate management, and that can be
   entirely and easily automated.  Appendix A.1 compares HPKP to the
   current document in more detail.

   The ticket pinning proposal augments these mechanisms with a much
   easier to implement and deploy solution for server identity pinning,
   by reusing some of the ideas behind TLS session resumption.

   This section compares ticket pinning to two earlier proposals, HPKP
   and TACK.

A.1.  Comparison: HPKP

   The current IETF standard for pinning the identity of web servers is
   HPKP [RFC7469].

   The main differences between HPKP and the current document are the
   following:

   *  HPKP limits its scope to HTTPS, while the current document
      considers all application above TLS.

   *  HPKP pins the public key of the server (or another public key
      along the certificate chain), and as such, is highly dependent on
      the management of certificates.  Such dependency increases the
      potential error surface, especially as certificate management is
      not yet largely automated.  The current proposal, on the other
      hand, is independent of certificate management.

   *  HPKP pins public keys that are public and used for the standard
      TLS authentication.  Identity pinning relies on the ownership of
      the pinning key, which is not disclosed to the public and not
      involved in the standard TLS authentication.  As a result,
      identity pinning is a completely independent, second-factor
      authentication mechanism.

   *  HPKP relies on a backup key to recover the misissuance of a key.
      We believe such backup mechanisms add excessive complexity and
      cost.  Reliability of the current mechanism is primarily based on
      its being highly automated.

   *  HPKP relies on the client to report errors to the report-uri.  The
      current document does not need any out-of-band mechanism, and the
      server is informed automatically.  This provides an easier and
      more reliable health monitoring.

   On the other hand, HPKP shares the following aspects with identity
   pinning:

   *  Both mechanisms provide hard failure.  With HPKP, only the client
      is aware of the failure, while with the current proposal both
      client and server are informed of the failure.  This provides room
      for further mechanisms to automatically recover from such
      failures.

   *  Both mechanisms are subject to a server compromise in which users
      are provided with an invalid ticket (e.g., a random one) or HTTP
      header with a very long lifetime.  For identity pinning, this
      lifetime SHOULD NOT be longer than 31 days.  In both cases,
      clients will not be able to reconnect the server during this
      lifetime.  With the current proposal, an attacker needs to
      compromise the TLS layer, while with HPKP, the attacker needs to
      compromise the HTTP server.  Arguably, the TLS-level compromise is
      typically more difficult for the attacker.

   Unfortunately HPKP has not seen wide deployment yet.  As of March
   2016, the number of servers using HPKP was less than 3000 [Netcraft].
   This may simply be due to inertia, but we believe the main reason is
   the interactions between HPKP and manual certificate management that
   is needed to implement HPKP for enterprise servers.  The penalty for
   making mistakes (e.g., being too early or too late to deploy new
   pins) is that the server becomes unusable for some of the clients.

   To demonstrate this point, we present a list of the steps involved in
   deploying HPKP on a security-sensitive web server.

   1.   Generate two public/private key pairs on a computer that is not
        the live server.  The second one is the "backup1" key pair.

        openssl genrsa -out "example.com.key" 2048;

        openssl genrsa -out "example.com.backup1.key" 2048;

   2.   Generate hashes for both of the public keys.  These will be used
        in the HPKP header:

        openssl rsa -in "example.com.key" -outform der -pubout | \
        openssl dgst -sha256 -binary | openssl enc -base64

        openssl rsa -in "example.com.backup1.key" -outform der \
        -pubout | openssl dgst -sha256 -binary | openssl enc -base64

   3.   Generate a single CSR (Certificate Signing Request) for the
        first key pair, where you include the domain name in the CN
        (Common Name) field:

        openssl req -new -subj "/C=GB/ST=Area/L=Town/O=Org/ \
        CN=example.com" -key "example.com.key" -out "example.com.csr";

   4.   Send this CSR to the CA and go though the dance to prove you own
        the domain.  The CA will give you a single certificate that will
        typically expire within a year or two.

   5.   On the live server, upload and set up the first key pair and its
        certificate.  At this point, you can add the "Public-Key-Pins"
        header, using the two hashes you created in step 2.

        Note that only the first key pair has been uploaded to the
        server so far.

   6.   Store the second (backup1) key pair somewhere safe, probably
        somewhere encrypted like a password manager.  It won't expire,
        as it's just a key pair; it just needs to be ready for when you
        need to get your next certificate.

   7.   Time passes -- probably just under a year (if waiting for a
        certificate to expire), or maybe sooner if you find that your
        server has been compromised, and you need to replace the key
        pair and certificate.

   8.   Create a new CSR using the "backup1" key pair, and get a new
        certificate from your CA.

   9.   Generate a new backup key pair (backup2), get its hash, and
        store it in a safe place (again, not on the live server).

   10.  Replace your old certificate and old key pair, update the
        "Public-Key-Pins" header to remove the old hash, and add the new
        "backup2" key pair.

   Note that in the above steps, both the certificate issuance as well
   as the storage of the backup key pair involve manual steps.  Even
   with an automated CA that runs the ACME protocol, protocol [RFC8555], key
   backup would be a challenge to automate.

A.2.  Comparison: TACK

   Compared with HPKP, TACK [TLS-TACK] is more similar to the current
   document.  It can even be argued that this document is a symmetric-
   cryptography variant of TACK.  That said, there are still a few
   significant differences:

   *  Probably the most important difference is that with TACK,
      validation of the server certificate is no longer required, and in
      fact TACK specifies it as a "MAY" requirement ([TLS-TACK],
      Section 5.3).  With ticket pinning, certificate validation by the
      client remains a MUST requirement, and the ticket acts only as a
      second factor.  If the pinning secret is compromised, the server's
      security is not immediately at risk.

   *  Both TACK and the current document are mostly orthogonal to the
      server certificate as far as their life cycle, and so both can be
      deployed with no manual steps.

   *  TACK uses Elliptic Curve Digital Signature Algorithm (ECDSA) to
      sign the server's public key.  This allows cooperating clients to
      share server assertions between themselves.  This is an optional
      TACK feature, and one that cannot be done with pinning tickets.

   *  TACK allows multiple servers to share its public keys.  Such
      sharing is disallowed by the current document.

   *  TACK does not allow the server to track a particular client, and
      so has better privacy properties than the current document.

   *  TACK has an interesting way to determine the pin's lifetime,
      setting it to the time period since the pin was first observed,
      with a hard upper bound of 30 days.  The current document makes
      the lifetime explicit, which may be more flexible to deploy.  For
      example, web sites that are only visited rarely by users may opt
      for a longer period than other sites that expect users to visit on
      a daily basis.

Acknowledgments

   The original idea behind this proposal was published in [Oreo] by
   Moti Yung, Benny Pinkas, and Omer Berkman.  The current protocol is
   but a distant relative of the original Oreo protocol, and any errors
   are the responsibility of the authors of this document alone.

   We would like to thank Adrian Farrel, Dave Garrett, Daniel Kahn
   Gillmor, Alexey Melnikov, Yoav Nir, Eric Rescorla, Benjamin Kaduk,
   and Rich Salz for their comments on this document.  Special thanks to
   Craig Francis for contributing the HPKP deployment script, and to
   Ralph Holz for several fruitful discussions.

Authors' Addresses

   Yaron Sheffer
   Intuit

   Email: yaronf.ietf@gmail.com

   Daniel Migault
   Ericsson

   Email: daniel.migault@ericsson.com