Network Working Group M. Thomson Internet-Draft Mozilla Intended status: Standards Track R. Hamilton Expires:September 22, 2016April 28, 2017 GoogleMarch 21,October 25, 2016Porting QUIC toUsing Transport Layer Security (TLS)draft-thomson-quic-tls-00 Abstract The QUIC experiment defines a custom security protocol. This was necessarytogain handshake latency improvements.Secure QUIC draft-thomson-quic-tls-01 Abstract This document describes howthat security protocol mightTransport Layer Security (TLS) can bereplaced with TLS.used to secure QUIC. 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 http://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 onSeptember 22, 2016.April 28, 2017. Copyright Notice Copyright (c) 2016 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 (http://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 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 . . . . . . . . . . . . . . . . . . . . . . . . 2 1.1. Notational Conventions . . . . . . . . . . . . . . . . . 3 2. Protocol Overview . . . . . . . . . . . . . . . . . . . . . . 3 2.1. Handshake Overview . . . . . . . . . . . . . . . . . . . 4 3.QUIC overTLSStructure . .in Stream 1 . . . . . . . . . . . . . . . . .5 4. Mapping of QUIC to QUIC over TLS. . . . . . 5 3.1. Handshake and Setup Sequence . . . . . . . .6 4.1. Protocol and Version Negotiation . . .. . . . . . 6 4. QUIC Record Protection . . .7 4.2. Source Address Validation. . . . . . . . . . . . . . . . 85. Record Protection4.1. Key Phases . . . . . . . . . . . . . . . . . . . . . . . 85.1.4.1.1. Retransmission of TLS HandshakeEncryption . . . . . . . . . .Messages . . . . . . 95.2.4.1.2. Key Update . . . . . . . . . . . . . . . . . . . . . 10 4.2. QUIC Key Expansion . .9 5.3. Sequence Number Reconstruction. . . . . . . . . . . . .10 5.4. Alternative Design: Exporters. . . . 11 4.3. QUIC AEAD application . . . . . . . . . .10 6. Pre-handshake QUIC Messages. . . . . . . . 12 4.4. Sequence Number Reconstruction . . . . . . . . .11 6.1. QUIC Extension. . . . 12 5. Pre-handshake QUIC Messages . . . . . . . . . . . . . . . . .11 6.2.13 5.1. Unprotected Frames Prior to Handshake Completion . . . .15 6.2.1.14 5.1.1. STREAM Frames . . . . . . . . . . . . . . . . . . . .15 6.2.2.14 5.1.2. ACK Frames . . . . . . . . . . . . . . . . . . . . . 156.2.3.5.1.3. WINDOW_UPDATE Frames . . . . . . . . . . . . . . . . 156.2.4. FEC5.1.4. Denial of Service with Unprotected Packets . . . . . 15 5.2. Use of 0-RTT Keys . . . . . . . . . . . . . . . . . . . . 166.3.5.3. Protected Frames Prior to Handshake Completion . . . . .16 7. Connection ID17 6. QUIC-Specific Additions to the TLS Handshake . . . . . . . . 18 6.1. Protocol and Version Negotiation . . . . . . . . . . . . 18 6.2. QUIC Extension . . . .17 8.. . . . . . . . . . . . . . . . . 18 6.3. Source Address Validation . . . . . . . . . . . . . . . . 19 6.4. Priming 0-RTT . . . . . . . . . . . . . . . . . . . . . . 19 7. Security Considerations . . . . . . . . . . . . . . . . . . .18 9.20 7.1. Packet Reflection Attack Mitigation . . . . . . . . . . . 20 7.2. Peer Denial of Service . . . . . . . . . . . . . . . . . 20 8. IANA Considerations . . . . . . . . . . . . . . . . . . . . .18 10.21 9. References . . . . . . . . . . . . . . . . . . . . . . . . .18 10.1.21 9.1. Normative References . . . . . . . . . . . . . . . . . .18 10.2.21 9.2. Informative References . . . . . . . . . . . . . . . . .1821 Appendix A. Acknowledgments . . . . . . . . . . . . . . . . . .1922 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . .1922 1. Introduction QUIC[I-D.tsvwg-quic-protocol][I-D.hamilton-quic-transport-protocol] provides a multiplexed transport for HTTP [RFC7230] semantics that provides several key advantages over HTTP/1.1 [RFC7230] or HTTP/2 [RFC7540] over TCP [RFC0793].The custom security protocol designed forThis document describes how QUIC can be secured using Transport Layer Security (TLS) version 1.3 [I-D.ietf-tls-tls13]. TLS 1.3 provides critical latency improvements for connectionestablishment.establishment over previous versions. Absent packet loss, most new connections can be establishedwithand secured within a single round trip; on subsequent connections between the same client and server, the client can often send application data immediately, that is, zero round trip setup.TLS 1.3 uses a similar design and aims to provide the same set of improvements.This document describes how the standardized TLS 1.3might serve ascan act a securitylayer forcomponent of QUIC. The same design could work for TLS 1.2, though few of the benefits QUIC provides would be realized due to the handshake latency in versions of TLS prior to 1.3.Alternative Designs: There are other designs that are possible; and many of these alternative designs are likely to be equally good. The point of this document is to articulate a coherent single design. Notes like this throughout the document are used describe points where alternatives were considered. Note: This is a rough draft. Many details have not been ironed out. Ryan is not responsible for any errors or omissions.1.1. Notational Conventions The words "MUST", "MUST NOT", "SHOULD", and "MAY" are used in this document. It's not shouting; when they are capitalized, they have the special meaning defined in [RFC2119]. 2. Protocol Overview QUIC[I-D.tsvwg-quic-protocol][I-D.hamilton-quic-transport-protocol] can be separated into several modules: 1. The basic frame envelope describes the common packet layout. This layer includes connection identification, version negotiation, and includesthe indicatorsmarkers that allow theframing, public reset,framing andFEC modulespublic reset to be identified. 2. The public reset is an unprotectedframepacket that allows an intermediary (an entity that is not part of the security context) to request the termination of a QUIC connection. 3.The forward error correction (FEC) module provides redundant entropy that allows forVersion negotiation frames are used tobe repaired in eventagree on a common version ofloss.QUIC to use. 4. Framing comprises most of the QUIC protocol. Framing provides a number of different types of frame, each with a specific purpose. Framing supports frames for both congestion management and stream multiplexing. Framing additionally provides a liveness testing capability (the PING frame). 5.CryptoEncryption provides confidentiality and integrity protection for frames. All frames are protectedafterbased on keying material derived from thehandshake completesTLS connection running on stream 1. Prior to this, data is protected with the 0-RTT keys. 6. Multiplexed streams are the primary payload of QUIC. These provide reliable, in-order delivery of data and are used to carry the encryption handshake and transport parameters (stream 1), HTTP header fields (stream 3), and HTTP requests and responses. Frames for managing multiplexing include those for creating and destroying streams as well as flow control and priority frames. 7. Congestion management includes packet acknowledgment and other signal required to ensure effective use of available link capacity. 8. A complete TLS connection is run on stream 1. This includes the entire TLS record layer. As the TLS connection reaches certain states, keying material is provided to the QUIC encryption layer for protecting the remainder of the QUIC traffic. 9. HTTP mapping provides an adaptation to HTTP that is based on HTTP/2. The relative relationship of these components are pictorally represented in Figure 1.+----+------++-----+------+ |HSTLS | HTTP |+----+------+------------++-----+------+------------+ | Streams | Congestion |+-----------+------------++------------+------------+ | Frames|+--------+---------+ ++------------+ | | FEC +--------+ + +--------+------------++---------------------+ Public | Version | |Crypto| Encryption | Reset |+--+---------------------+--------+Nego. | +---+---------------------+--------+---------+ | Envelope |+---------------------------------++--------------------------------------------+ | UDP |+---------------------------------+ *HS = Crypto Handshake+--------------------------------------------+ Figure 1: QUIC Structure This documentdescribes a replacement ofdefines the cryptographic parts of QUIC. This includes the handshake messages that are exchanged on stream 1, plus the record protection that is used to encrypt and authenticate all other frames. 2.1. Handshake Overview TLS 1.3 provides two basic handshake modes of interest to QUIC: o A full handshake in which the client is able to send application data after one round trip and the server immediately after receiving the first message from the client. o A 0-RTT handshake in which the client uses information about the server to send immediately. This data can be replayed by an attacker so it MUST NOT carry a self-contained trigger for any non-idempotent action. A simplified TLS 1.3 handshake with 0-RTT application data is shown in Figure 2, see [I-D.ietf-tls-tls13] for moreoptions.options and details. Client Server ClientHello (Finished) (0-RTT Application Data) (end_of_early_data) --------> ServerHello {EncryptedExtensions} {ServerConfiguration} {Certificate} {CertificateVerify} {Finished} <-------- [Application Data] {Finished} --------> [Application Data] <-------> [Application Data] Figure 2: TLS Handshake with 0-RTT Two additional variations on this basic handshake exchange are relevant to this document: o The server can respond to a ClientHello with a HelloRetryRequest, which adds an additional round trip prior to the basic exchange. This is needed if the server wishes to request a different key exchange key from the client. HelloRetryRequestmightis alsobeused to verify that the client is correctly able to receive packets on the address it claims to have (see Section4.2).6.3). o A pre-shared key mode can be used for subsequent handshakes to avoid public key operations. Thismight beis the basis for0-RTT,0-RTT data, even if the remainder of the connection is protected by a new Diffie-Hellman exchange. 3.QUIC overTLSStructurein Stream 1 QUIC completes its cryptographic handshake on stream 1, which means that the negotiation of keying material happenswithinafter the QUICprotocol. QUIC over TLS does the same, relying onprotocol has started. This simplifies theordered delivery guarantees provided byuse of TLS since QUIC is able to ensure that the TLS handshake packets are delivered reliably and in order.+-----+---------+ | TLS | HTTP | +-----+----------+------------+ | Streams | Congestion | +----------------+------------+ | Frames | | +------------+ | | FEC +--------+ | +----------+------------+ Public | | | TLS Record Protection | Reset | +-----+-----------------------+--------+ | Envelope | +--------------------------------------+ | UDP | +--------------------------------------+ Figure 3: QUIC over TLS In this design theQUICenvelopeStream 1 carriesQUIC frames until thea complete TLShandshake completes. After the handshake successfully completesconnection. This includes thekey exchange, QUIC frames are then protected byTLS recordprotection. QUIC stream 1 is used to exchange TLS handshake packets.layer in its entirety. QUIC provides for reliable andin-orderin- order delivery of the TLS handshakemessages.messages on this stream. Prior to the completion of the TLS handshake, QUIC frames can be exchanged. However, these frames are not authenticated or confidentiality protected. Section65 covers some of the implications of thisdesign. Alternative Design: TLS could be used to protect the entiredesign and limitations on QUICenvelope.operation during this phase. Once complete, QUICversion negotiation could be subsumed by TLS and ALPN [RFC7301]. The only unprotected packetsframes arethen public resets and ACK frames, both of which could be given first octet values that would easily distinguish them from other TLS packets. This requires that theprotected using QUICsequence numbers be moved to the outside of the record.record protection, see Section 4.Mapping of QUIC to QUIC over TLS Several changes to the structure3.1. Handshake and Setup Sequence The integration of QUICare necessary to makewith alayered design practical. These changes produce theTLS handshake is shown in more detail in Figure4. In this handshake,3. QUICSTREAM"STREAM" frames on stream 1 carry the TLS handshake. QUIC is responsible for ensuring that the handshake packets arere- sentre-sent in case of loss and that they can be ordered correctly.QUIC operates without any record protection until the handshake completes, just as TLS over TCP does not include record protection for the handshake messages. Once complete, QUIC frames and forward error control (FEC) messages are encapsulated in using TLS record protection.Client Server @A QUIC STREAMFrame <stream 1>Frame(s) <1>: ClientHello + QUIC Setup Parameters(Finished)-------->(Replayable0-RTT Key -> @B @B QUIC STREAM Frame(s) <1>: (Finished) Replayable QUIC Frames <anystream>) (end_of_early_data <1>)stream> --------> QUIC STREAM Frame<1><1>: @B/A ServerHello{EncryptedExtensions} {ServerConfiguration} {Certificate} {CertificateVerify} {Finished}{Handshake Messages} <-------- 1-RTT Key -> @C QUIC Frames @C <--------[QUIC Frames/FEC]@B QUIC STREAMFrame <1>Frame(s) <1>: (end_of_early_data <1>) {Finished} -------->[QUIC Frames/FEC]@C QUIC Frames <------->[QUIC Frames/FEC]QUIC Frames @C Figure4:3: QUIC over TLS HandshakeThe remainder of this document describesIn Figure 3, symbols mean: o "<" and ">" enclose stream numbers. o "@" indicates thechanges tokey phase that is currently used for protecting QUIC packets. o "(" and ")" enclose messages that are protected with TLS 0-RTT handshake or application keys. o "{" and "}" enclose messages thatalloware protected by theprotocols to operate together. 4.1. Protocol and Version Negotiation The QUIC version negotiation mechanismTLS Handshake keys. If 0-RTT isused to negotiatenot possible, then theversion of QUIC that is used prior toclient does not send frames protected by thecompletion of0-RTT key (@B). The only key transition on thehandshake. However, this packetclient is from cleartext (@A) to 1-RTT protection (@C). If 0-RTT data is notauthenticated, enabling an active attackeraccepted by the server, then the server sends its handshake messages without protection (@A). The client still transitions from @A toforce a version downgrade. To ensure that@B, but it can stop sending 0-RTT data and progress immediately to 1-RTT data when it receives a cleartext ServerHello. 4. QUICversion downgradeRecord Protection QUIC provides a record protection layer that isnot forcedresponsible for authenticated encryption of packets. The record protection layer uses keys provided byan attacker, version information is copied intothe TLShandshake, which providesconnection and authenticated encryption to provide confidentiality and integrity protection for the content of packets. Different keys are used for QUICnegotiation.and TLS record protection. Having separate QUIC and TLS record protection means that TLS records can be protected by two different keys. Thisdoesn't prevent version downgrade duringredundancy is maintained for thehandshake, though it does prevent a connection from completing with a downgraded version, see Section 6.1. ISSUE:sake of simplicity. 4.1. Key Phases The transition to use of a new QUICversion negotiation has poor performance inkey occurs immediately after sending theeventTLS handshake messages that produced the key transition. Every time that aclientnew set of keys isforced to downgrade from their preferred version. 4.2. Source Address Validation QUIC implementations describe a source address token. Thisused for protecting outbound messages, the KEY_PHASE bit in the public flags isan opaque blob that a server provides to clients when they first use a given source address.toggled. Theclient returns this token in subsequentKEY_PHASE bit on unencrypted messagesas a return routeability check. That is, the client returns this token to prove that itisable to receive packets at0. The KEY_PHASE bit on thesource address that it claims. Since this tokenpublic flags isopaque and consumed only bytheserver, it can be includedmost significant bit (0x80). The KEY_PHASE bit allows a recipient to detect a change in keying material without needing to receive theTLS 1.3 configuration identifier for 0-RTT handshakes. Serversmessage thatuse 0-RTTtriggers the change. This avoids head-of-line blocking around transitions between keys without relying on trial decryption. The following transitions areadviseddefined: o The client transitions toprovide new configuration identifiersusing 0-RTT keys afterevery handshakesending the ClientHello. This causes the KEY_PHASE bit on packets sent by the client toavoid passive linkability of connectionsbe set to 1. o The server transitions to using 0-RTT keys before sending the ServerHello, but only if the early data from thesame client. A server thatclient isunder load might includeaccepted. This transition causes thesame information inKEY_PHASE bit on packets sent by thecookie extension/field of a HelloRetryRequest. (Note:server to be set to 1. If thecurrent version of TLS 1.3 does not includeserver rejects 0-RTT data, theability to includeserver's handshake messages are sent without QUIC-level record protection with acookie in HelloRetryRequest.) 5. Record Protection EachKEY_PHASE of 0. TLS handshake messages will still be protected by TLS recordis encapsulated inprotection based on theQUIC envelope.TLS handshake traffic keys. o The server transitions to using 1-RTT keys after sending its Finished message. Thisprovides length information, which means thatcauses thelength field canKEY_PHASE bit to bedroppedset to 0 if early data was accepted, and 1 if the server rejected early data. o The client transitions to 1-RTT keys after sending its Finished message. Subsequent messages from the client will then have a KEY_PHASE of 0 if 0-RTT data was sent, and 1 otherwise. o Both peers start sending messages protected by a new key immediately after sending a TLSrecord.KeyUpdate message. Thesequence numbervalue of the KEY_PHASE bit is changed each time. At each point, both keying material (see Section 4.2) and the AEAD function used by TLSrecord protectionischanged to dealinterchanged with thepotentialvalues that are currently in use for protecting outbound packets. Once a change of keys has been made, packetsto be dropped or lost. The QUICwith higher sequencenumbernumbers MUST use the new keying material until a newer set of keys (and AEAD) are used. The exception to this isused in placethat retransmissions ofthe monotonically increasingTLSrecord sequence number. This meanshandshake packets MUST use the keys that they were originally protected with. Once a packet protected by a new key has been received, a recipient SHOULD retain theTLS record protection employed is closerprevious keys for a short period. Retaining old keys allows the recipient toDTLSdecode reordered packets around a change inboth its form and the guarantees that are provided. QUICkeys. Keys SHOULD be discarded when an endpoints hasa single, contiguous sequence number space. In comparison, TLS restarts itsreceived all packets with sequencenumber each time that record protection keys are changed. Thenumbers lower than the lowest sequence numberrestart in TLS ensuresused for the new key, or when it determines thata compromisereordering ofthe current trafficthose packets is unlikely. 0-RTT keys SHOULD be retained until the handshake is complete. The KEY_PHASE bit does notallow an attacker to truncate thedirectly indicate which keys are in use. Depending on whether 0-RTT datathat iswas sentafter a key update by sending additionaland accepted, packetsunderprotected with keys derived from theold key (causing new packetssame secret might be marked with different KEY_PHASE values. 4.1.1. Retransmission of TLS Handshake Messages TLS handshake messages need to bediscarded).retransmitted with the same level of cryptographic protection that was originally used to protect them. Newer keys cannot be used to protect QUICdoes not rely on there beingpackets that carry TLS messages. A client would be unable to decrypt retransmissions of acontinuous sequenceserver's handshake messages that are protected using the 1-RTT keys, since the calculation of the application datapackets; QUIC uses authenticated repair mechansims that operate abovekeys depends on thelayercontents ofencryption. QUIC can therefore operate without restarting sequence numbers. 5.1. TLS Handshake Encryption TLS 1.3 adds encryption forthe handshake messages. Thisintroduces an additional transition between different record protection keys duringrestriction means thehandshake. A consequencecreation ofthis is that it becomes more importantan exception toexplicitly identifythetransition from one set ofrequirement to always use new keys for sending once they are available. A server MUST mark the retransmitted handshake messages with the same KEY_PHASE as the original messages to allow a recipient to distinguish thenext (see Section 5.2). 5.2.messages. 4.1.2. Key UpdateEach time thatOnce the TLSrecord protection keys are changed,handshake is complete, the KEY_PHASE bit allows for the processing of messages without having to receive the TLS KeyUpdate messageinitiatingthat triggers thechange could be lost.key update. Thisresults in subsequent packets being indecipherableallows endpoints to start using updated keys immediately without thepeerconcern thatreceives them. Key changes happena lost KeyUpdate will cause their messages to be indecipherable to their peer.. An endpoint MUST NOT initiate more than one key update at a time. A new key update cannot be sent until theconclusion of the handshake and and immediately afterendpoint has received a matching KeyUpdatemessage. TLS relies onmessage from its peer; or, if the endpoint did not initiate the original key update, it has received anordered, reliable transport and therefore provides no other mechanismacknowledgment of its own KeyUpdate. This ensures that there are at most two keys toensuredistinguish between at any one time, for which the KEY_PHASE bit is sufficient. Initiating Peer Responding Peer @M KeyUpdate New Keys -> @N @N QUIC Frames --------> KeyUpdate @N <-------- -- Initiating Peer can initiate another KeyUpdate -- @N Acknowledgment --------> -- Responding Peer can initiate another KeyUpdate -- Figure 4: Key Update As shown in Figure 3 and Figure 4, there is never a situation where there are more than two different sets of keying material that might be received by apeer receives the message initiatingpeer. A server cannot initiate a keychange prior to receivingupdate until it has received thesubsequent messages that areclient's Finished message. Otherwise, packets protectedusingby thenew key.updated keys could be confused for retransmissions of handshake messages. Asimilar mechanism here would introduce head-of-line blocking. The simplest solution here is to stealclient cannot initiate asingle bit fromkey update until it has received an acknowledgment that its Finished message has been received. Note: This models theunprotected part ofkey changes in theQUIC header that signalshandshake as a keyupdates, similar to how DTLS signalsupdate initiated by theepoch on each packet. The epoch bit is encoded into 0x80 ofserver, with theQUIC public flags. Each timeFinished message in theepoch bit changes, an attempt is made to updateplace of KeyUpdate. 4.2. QUIC Key Expansion The following table shows QUIC keys, when they are generated and the TLS secret from which they are derived: +-------+----------------------+----------------------------+ | Key | TLS Secret | Phase | +-------+----------------------+----------------------------+ | 0-RTT | early_traffic_secret | "QUIC 0-RTT key expansion" | | | | | | 1-RTT | traffic_secret_N | "QUIC 1-RTT key expansion" | +-------+----------------------+----------------------------+ 0-RTT keys are those keys that are used in resumed connections prior toread. Peersthe completion of the TLS handshake. Data sent using 0-RTT keys might be replayed and so has some restrictions on its use, see Section 5.2. 0-RTT keys areprohibited fromused after sending or receiving a ClientHello. 1-RTT keys are used after the TLS handshake completes. There are potentially multiple sets of 1-RTT keys; new 1-RTT keys are created by sending a TLS KeyUpdatemessages until they seemessage. 1-RTT keys are used after sending areciprocalFinished or KeyUpdateto preventmessage. The complete key expansion uses thechance that a transition is undetectedsame process for key expansion asa resultdefined in Section 7.3 oftwo changes[I-D.ietf-tls-tls13]. For example, the Client Write Key for the data sent immediately after sending the TLS Finished message is: label = "QUIC 1-RTT key expansion, client write key" client_write = HKDF-Expand-Label(traffic_secret_0, label, "", key_length) This results inthis bit.a label input to HKDF that includes a two-octet length field, the string "TLS 1.3, QUIC 1-RTT key expansion, client write key" and a zero octet. ThetransitionQUIC record protection initially starts without keying material. When the TLS state machine produces the corresponding secret, new keys are generated fromcleartextthe TLS connection and used toencrypted packetsprotect the QUIC record protection. The Authentication Encryption with Associated Data (AEAD) [RFC5116] function used isexempt from this limit ofthe same onekey change. Two key changes occur duringthat is negotiated for use with thehandshake. The server sends packets inTLS connection. For example, if TLS is using theclear, plusTLS_ECDHE_ECDSA_WITH_AES_128_GCM_SHA256, the AEAD_AES_128_GCM function is used. 4.3. QUIC AEAD application Regular QUIC packets are protectedusing handshakeby an AEAD [RFC5116]. Version negotiation andapplication data keys. With onlypublic reset packets are not protected. Once TLS has provided asingle bit available to discriminate between keys,key, the contents of regular QUIC packets immediately after any TLS messages have been sent are protectedwithby theapplication data keys will haveAEAD selected by TLS. The key, K, for thesame bit valueAEAD is either the Client Write Key or the Server Write Key, derived ascleartext packets. This condition will be easily identified and handled, likelydefined in Section 4.2. The nonce, N, for the AEAD is formed bydiscardingcombining either theapplicationClient Write IV or Server Write IV with the sequence numbers. The 48 bits of the reconstructed QUIC sequence number (see Section 4.4) in network byte order is left-padded with zeros to the N_MAX parameter of the AEAD (see Section 4 of [RFC5116]). The exclusive OR of the padded sequence number and the IV forms the AEAD nonce. The associated data,sinceA, for theencrypted packets will be highly unlikelyAEAD is an empty sequence. The input plaintext, P, for the AEAD is the contents of the QUIC frame following the packet number, as described in [I-D.hamilton-quic-transport-protocol] The output ciphertext, C, of the AEAD is transmitted in place of P. Prior to TLS providing keys, no record protection is performed and the plaintext, P, is transmitted unmodified. Note: QUIC defined a null-encryption that had an additional, hash- based checksum for cleartext packets. This might bevalid. 5.3.added here, but it is more complex. 4.4. Sequence Number Reconstruction Each peer maintains a 48-bitsendsequence number that is incremented witheachevery packet that issent (even retransmissions).sent, including retransmissions. The least significant 8-, 16-, 32-, or 48-bits of this number is encoded in the QUIC sequence number field in every packet. A16-bit send epoch number is maintained; the epoch is incremented each time new record protection keying material is used. The least significant bit of the epoch number is encoded into the epoch bit (0x80) of the QUIC public flags. Areceiver maintains the same values, but recovers values based on the packets it receives. This is based on the sequence number of packets that it has received. A simple scheme predicts the receive sequence number of an incoming packet by incrementing the sequence number of the most recent packet to be successfully decrypted by one and expecting the sequence number to be within a range centered on that value.The receive epoch value is incremented each time that the epoch bit (0x80) changes. The sequence number used for record protection isA more sophisticated algorithm can almost double the64-bit value obtainedsearch space byconcatenatingchecking backwards from theepoch andmost recent sequencenumber, both in network byte order. 5.4. Alternative Design: Exporters An exporter could be used to provide keying materialfor aQUIC- specific record protection. This could draw on the selected cipher suite andreceived (or abandoned) packet. If a packet was received, then theTLS record protection design sopacket contains a sequence number thatthe overall effort required to design and analyzeiskept minimal. One concern with using exportersgreater than the most recent sequence number. If no such packet was found, the number isthat TLS doesn't define an exporter for use priorassumed to be in theend ofsmaller window centered on thehandshake. That meansnext sequence number, as in thecreation ofsimpler scheme. Note: QUIC has aspecial exporter for usesingle, contiguous sequence number space. In comparison, TLS restarts its sequence number each time that record protection keys are changed. The sequence number restart inprotecting 0-RTT data. That's a pretty sharp object to leave lying around, and it's not clear what the properties we could provide. (That doesn't meanTLS ensures thatthere wouldn't be demand for suchathing,compromise of thepossibility has already been raised.) An exporter-based scheme might optcurrent traffic keys does not allow an attacker tousetruncate thehandshake traffic keysdata that is sent after a key update by sending additional packets under the old key (causing new packets toprotectbe discarded). QUIC does not assume a reliable transport and is therefore required to handle attacks where packetsduring the handshake, relying instead onare dropped in other ways. TLS maintains a separateprotectionsequence number that is used for record protection on theTLS handshake records.connection that is hosted on stream 1. Thiscomplicates implementations somewhat, so an exporter might still be used. In the end, using an exporter doesn't alter the design significantly. Given the risks, a modificationsequence number is reset according to therecord protocol is probably safer. 6.rules in the TLS protocol. 5. Pre-handshake QUIC Messages Implementations MUST NOT exchange data on any stream other than stream 1 prior to the completion of the TLShandshake completing.handshake. However, QUIC requires the use of several types of frame for managing loss detection and recovery. In addition, it might be useful to use the data acquired during the exchange of unauthenticated messages for congestion management. This section generally only applies to TLS handshake messages from both peers and acknowledgments of the packets carrying those messages. In many cases, the need for servers to provide acknowledgments is minimal, since the messages that clients send are small and implicitly acknowledged by the server's responses. The actions that a peer takes as a result of receiving an unauthenticated packet needstobeto be limited. In particular, state established by these packets cannot be retained once record protection commences. There are several approaches possible for dealing with unauthenticated packets prior to handshake completion: o discard and ignore them o use them, but reset any state that is established once the handshake completes o use them and authenticate them afterwards; failing the handshake if they can't be authenticated o save them and use them when they can be properly authenticated o treat them as a fatal error Different strategies are appropriate for different types of data. This document proposes that all strategies are possible depending on the type of message. o Transport parameters and options are made usable and authenticated as part of the TLS handshake (see Section6.1).6.2). o Most unprotected messages are treated as fatal errors when received except for the small number necessary to permit the handshake to complete (see Section6.2).5.1). o Protected packets can either bediscarded, but can bediscarded or saved and later used (see Section6.3). 6.1. QUIC Extension A client describes characteristics of the transport protocol it intends to conduct with the server in a new QUIC-specific extension in its ClientHello. The server uses this information to determine whether it wants to continue the connection, request source address validation, or reject the connection. Having this information unencrypted permits this check to occur prior to committing the resources needed to complete the initial key exchange. If the server decides to complete the connection, it generates a corresponding response and includes it in the EncryptedExtensions message. These parameters are not confidentiality-protected when sent by the client, but the server response is protected by the handshake traffic keys. The entire exchange is integrity protected once the handshake completes. This information is not used by TLS, but can be passed to the QUIC protocol as initialization parmeters. The "quic_parameters" extension contains a declarative set of parameters that establish QUIC operating parameters and constrain the behaviour of a peer. The connection identifier and version are first negotiated using QUIC, and are included in the TLS handshake in order to provide integrity protection. enum { receive_buffer(0), (65535) } QuicTransportParameterType; struct { QuicTransportParameterType type; uint32 value; } QuicTransportParameter; uint32 QuicVersion; enum { (65535) } QuicOption; struct { uint64 connection_id; QuicVersion quic_version; QuicVersion supported_quic_versions<0..2^8-1>; uint32 connection_initial_window; uint32 stream_initial_window; uint32 implicit_shutdown_timeout; QuicTransportParameter transport_parameters<0..2^16-1>; QuicOption options<0..2^8-2>; } QuicParametersExtension; This extension MUST be included if a QUIC version is negotiated. A server MUST NOT negotiate QUIC if this extension is not present. Based on the values offered by a client a server MAY use the values in this extension to determine whether it wants to continue the connection, request source address validation, or reject the connection. Since this extension is initially unencrypted, the server can use the information prior to committing the resources needed to complete a key exchange. If the server decides to use QUIC, this extension MUST be included in the EncryptedExtensions message. The parameters are: connection_id: The 64-bit connection identifier for the connection, as selected by the client. quic_version: The currently selected QUIC version that is used for the connection. This is the version negotiated using the unauthenticated QUIC version negotiation (Section 4.1). supported_quic_versions: This is a list of supported QUIC versions for each peer. A client sends an empty list if the version of QUIC being used is their preferred version; however, a client MUST include their preferred version if this was not negotiated using QUIC version negotiation. A server MUST include all versions that it supports in this list. connection_initial_window: The initial value for the connection flow control window for the endpoint, in octets. connection_initial_window: The initial value for the flow control window of new streams created by the peer endpoint, in octets. implicit_shutdown_timeout: The time, in seconds, that a connection can remain idle before being implicitly shutdown. transport_parameters: A list of parameters for the QUIC connection, expressed as key-value pairs of arbitrary length. The QuicTransportParameterType identifies each parameter; duplicate types are not permitted and MUST be rejected with a fatal illegal_parameter alert. Type values are taken from a single space that is shared by all QUIC versions. ISSUE: There is currently no way to update the value of parameters once the connection has started. QUIC crypto provided a SCFG message that could be sent after the connection was established. options: A list of options that can be negotiated for a given connection. These are set during the initial handshake and are fixed thereafter. These options are used to enable or disable optional features in the protocol. The set of features that are supported across different versions might vary. A client SHOULD include all options that it is willing to use. The server MAY select any subset of those options that apply to the version of QUIC that it selects. Only those options selected by the server are available for use. Note: This sort of optional behaviour seems like it could be accommodated adequately by defining new versions of QUIC for each experiment. However, as an evolving protocol, multiple experiments need to be conducted concurrently and continuously. The options parameter provides a flexible way to regulate which experiments are enabled on a per-connection basis. 6.2.5.3). 5.1. Unprotected Frames Prior to Handshake Completion This section describes the handling of messages that are sent and received prior to the completion of the TLS handshake. Sending and receiving unprotected messages is hazardous. Unless expressly permitted, receipt of an unprotected message of any kind MUST be treated as a fatal error.6.2.1.5.1.1. STREAM Frames "STREAM" frames for stream 1 are permitted. These carry the TLS handshake messages. Receiving unprotected "STREAM" framesthat do not contain TLS handshake messagesfor other streams MUST be treated as a fatal error.6.2.2.5.1.2. ACK Frames "ACK" frames are permitted prior to the handshake being complete.However, an unauthenticatedInformation learned from "ACK"frame can only be used to obtain NACK ranges. Timestamps MUST NOTframes cannot beincluded in an unprotected ACK frame,entirely relied upon, sincethese might be modified byan attackerwith the intent of altering congestion control response. Information on FEC-revived packetsisredundant, since use of FEC in this phase is prohibited.able to inject these packets. Timing and packet retransmission information from "ACK" framesMAY be sent a second time once record protectionisenabled. Once protected, timestamps can be included. Editor's Note: This prohibitioncritical to the functioning of the protocol, but these frames might bea little too strong, but this is thespoofed or altered. Endpoints MUST NOT use an unprotected "ACK" frame to acknowledge data that was protected by 0-RTT or 1-RTT keys. An endpoint MUST ignore an unprotected "ACK" frame if it claims to acknowledge data that was protected data. Such an acknowledgement can onlyobviously safe option. If the amountserve as a denial ofdamage thatservice, since anattackerendpoint that cando by modifying timestampsread protected data islimited, then it might be OKalways permitted topermitsend protected data. An endpoint SHOULD use data from unprotected or 0-RTT-protected "ACK" frames only during theinclusion of timestamps. Note that an attacker need notinitial handshake and while they have insufficient information from 1-RTT-protected "ACK" frames. Once sufficient information has been obtained from protected messages, information obtained from less reliable sources can beon-path to inject an ACK. 6.2.3.discarded. 5.1.3. WINDOW_UPDATE FramesSending a"WINDOW_UPDATE" frames MUST NOT be sent unprotected. Though data is exchanged on stream1 might be necessary1, the initial flow control window is is sufficiently large topermitallow thecompletionTLS handshake to complete. This limits the maximum size of the TLShandshake, particularly in cases where the certification path is lengthy. To avoid stalling due to flow control exhaustion, "WINDOW_UPDATE" frames with stream 1 are permitted. Receiving a "WINDOW_UPDATE" frame on streams other than 1 MUST be treated ashandshake and would prevent afatal error.server or client from using an abnormally large certificate chain. Stream 1 is exempt from the connection-level flow control window.The position5.1.4. Denial of Service with Unprotected Packets Accepting unprotected - specifically unauthenticated - packets presents a denial of service risk to endpoints. An attacker that is able to inject unprotected packets can cause a recipient to drop even protected packets with a matching sequence number. The spurious packet shadows theflow control windowgenuine packet, causing the genuine packet to be ignored as redundant. Once the TLS handshake is complete, both peers MUST ignore unprotected packets. The handshake is complete when the server receives a client's Finished message and when a client receives an acknowledgement that their Finished message was received. From that point onward, unprotected messages can beresetsafely dropped. Note that the client could retransmit its Finished message todefaults oncethe server, so the server cannot reject such a message. Since only TLS handshake packets and acknowledgments are sent in the clear, an attacker iscomplete.able to force implementations to rely on retransmission for packets that are lost or shadowed. Thus, an attacker that intends to deny service to an endpoint has to drop or shadow protected packets in order to ensure that their victim continues to accept unprotected packets. The ability to shadow packets means that an attacker does not need to be on path. ISSUE: Thismight resultwould not be an issue if QUIC had a randomized starting sequence number. If we choose to randomize, we fix this problem and reduce the denial of service exposure to on-path attackers. The only possible problem is in authenticating thewindow positioninitial value, so that peers can be sure that they haven't missed an initial message. In addition to denying endpoints messages, an attacker to generate packets that cause no state change in a recipient. See Section 7.2 foreithera discussion of these risks. To avoid receiving TLS packets that contain no useful data, a TLS implementation MUST reject empty TLS handshake records and any record that is not permitted by theconnectionTLS state machine. Any TLS application data orstream 1 being smalleralerts - other than a single end_of_early_data at thenumberappropriate time - that is received prior to the end ofoctetsthe handshake MUST be treated as a fatal error. 5.2. Use of 0-RTT Keys If 0-RTT keys are available, the lack of replay protection means thathave been sentrestrictions onthose streams. A "WINDOW_UPDATE" frame might therefore betheir use are necessary topreventavoid replay attacks on theconnection from being stalled. Note: This isprotocol. A client MUST onlypotentially problematic for servers, who might needuse 0-RTT keys tosend large certificate chains. In other cases, this is unlikely givenprotect data thatQUIC - like HTTP [RFC7230] -isa protocol where the server is unableidempotent. A client MAY wish to apply additional restrictions on what data it sends prior toexercisetheopportunitycompletion of the TLSpresentshandshake. A client otherwise treats 0-RTT keys as equivalent tosend first. If1-RTT keys. A client that receives an indication that its 0-RTT data has been accepted by a server can send 0-RTT data until it receives all of the server's handshake messages. A client SHOULD stop sending 0-RTT data if it receives an indication that 0-RTT data hasa large certificate chain, or later modifications or extensionsbeen rejected. In addition toQUIC permit the servera ServerHello without an early_data extension, an unprotected handshake message with a KEY_PHASE bit set to 0 indicates that 0-RTT data has been rejected. A client SHOULD sendfirst,its end_of_early_data alert only after it has received all of the server's handshake messages. Alternatively phrased, a clientmight reduce the chance of stalling dueis encouraged toflow control in this first round trip by setting larger values foruse 0-RTT keys until 1-RTT keys become available. This prevents stalling of theinitial stream andconnectionflow control windowsand allows the client to send continuously. A server MUST NOT use 0-RTT keys to protect anything other than TLS handshake messages. Servers therefore treat packets protected with 0-RTT keys as equivalent to unprotected packets in determining what is permissible to send. A server protects handshake messages using the"quic_parameters" extension (Section 6.1). Editor's Note: Unlike "ACK",0-RTT key if it decides to accept a 0-RTT key. A server MUST still include theprohibition on "WINDOW_UPDATE" is much less of an imposition on implementations. And, given thatearly_data extension in its ServerHello message. This restriction prevents aspurious "WINDOW_UPDATE" might be usedserver from responding tocreateagreat deal of memory pressure on an endpoint,request using frames protected by therestriction seems justifiable. Besides, I understand0-RTT keys. This ensures that all application data from the server are always protected with keys that have forward secrecy. However, thisone a lot better. 6.2.4. FEC Packets FEC packets MUST NOTresults in head-of-line blocking at the client because server responses cannot besent prior to completingdecrypted until all theTLS handshake. Endpoints MUST treat receipt of an unprotected FEC packet as a fatal error. 6.3.server's handshake messages are received by the client. 5.3. Protected Frames Prior to Handshake Completion Due to reordering and loss, protected packets might be received by an endpoint before the final handshake messages are received. If these can be decrypted successfully, such packets MAY be stored and used once the handshake is complete. Unless expressly permitted below, encrypted packets MUST NOT be used prior to completing the TLS handshake, in particular the receipt of a valid Finished message and any authentication of the peer. If packets are processed prior to completion of the handshake, an attacker might use the willingness of an implementation to use these packets to mount attacks. TLS handshake messages are covered by record protection during the handshake, once key agreement has completed. This means that protected messages need to be decrypted to determine if they are TLS handshake messages or not. Similarly, "ACK" and "WINDOW_UPDATE" frames might be needed to successfully complete the TLS handshake. Any timestamps present in "ACK" frames MUST be ignored rather than causing a fatal error. Timestamps on protected frames MAY be saved and used once the TLS handshake completes successfully. An endpointMUSTMAY save the last protected "WINDOW_UPDATE" frame it receives for each stream and apply the values once the TLS handshake completes.Editor's Note: Ugh. This last one is pretty ugly. Maybe we should just makeFailing to do this might result in temporary stalling of affected streams. 6. QUIC-Specific Additions to the TLS Handshake QUIC uses the TLS handshakeexempt from flow control upfor more than just negotiation of cryptographic parameters. The TLS handshake validates protocol version selection, provides preliminary values for QUIC transport parameters, and allows a server to perform return routeability checks on clients. 6.1. Protocol and Version Negotiation The QUIC version negotiation mechanism is used to negotiate theFinished message. Then we can prohibit unauthenticated "WINDOW_UPDATE" messages. We would still likely wantversion of QUIC that is used prior toaccount forthepackets sent and received, sincecompletion of the handshake. However, this packet is not authenticated, enabling an active attacker todo otherwise would create some hairy special cases. That meansforce a version downgrade. To ensure thatstallinga QUIC version downgrade ispossible, butnot forced by an attacker, version information is copied into the TLS handshake, which provides integrity protection for the QUIC negotiation. This does not prevent version downgrade during the handshake, though it means thatwe can avoid ugly rules likesuch a downgrade causes a handshake failure. Protocols that use theabove. 7. Connection ID TheQUICconnectiontransport MUST use Application Layer Protocol Negotiation (ALPN) [RFC7301]. The ALPN identifierservesfor the protocol MUST be specific toidentifythe QUIC version that it operates over. When constructing aconnection and to allowClientHello, clients MUST include aserver to resumelist of all the ALPN identifiers that they support, regardless of whether the QUIC version that they have currently selected supports that protocol. Servers SHOULD select anexisting connection from a newapplication protocol based solely on the information in the ClientHello, not using the QUIC version that the clientaddresshas selected. If the protocol that is selected is not supported with the QUIC version that is incase of mobility events. However,use, the server MUST either send a QUIC version negotiation packet if thiscreatesis possible, or fail the connection otherwise. 6.2. QUIC Extension QUIC defines anidentifier that a passive observer [RFC7258] canextension for useto correlate connections.with TLS. That extension defines transport-related parameters. This provides integrity protection for these values. Including these in the TLS1.3 offers connection resumption using pre-shared keys, whichhandshake alsoallowsmake the values that a client sets available tosend 0-RTT application data. This mode could be used to continueaconnection ratherserver one-round trip earlier thanrely on a publicly visible correlator.parameters that are carried in QUIC frames. Thisonly requiresdocument does not define thatservers produceextension. 6.3. Source Address Validation QUIC implementations describe anew ticket on every connection andsource address token. This is an opaque blob that a server might provide to clientsdo not resume from the same ticket more than once.when they first use a given source address. Theadvantage of relying on 0-RTT modes for mobility eventsclient returns this token in subsequent messages as a return routeability check. That is, the client returns this token to prove that it is able to receive packets at the source address thatthisit claims. This prevents the server from being used in packet reflection attacks (see Section 7.1). A source address token isalso more robust. Ifopaque and consumed only by the server. Therefore it can be included in the TLS 1.3 pre-shared key identifier for 0-RTT handshakes. Servers that use 0-RTT are advised to provide newpointpre-shared key identifiers after every handshake to avoid linkability ofattachment results in contactingconnections by passive observers. Clients MUST use a new pre-shared key identifier for every connection that they initiate; if no pre-shared key identifier is available, then resumption is not possible. A serverinstance - onethatlacksis under load might include a source address token in thesession state - thencookie extension of afallbackHelloRetryRequest. (Note: the current version of TLS 1.3 does not include the ability to include a cookie in HelloRetryRequest.) 6.4. Priming 0-RTT QUIC uses TLS without modification. Therefore, it iseasy. The main drawback withpossible to use aclean restart or anything resemblingpre-shared key that was obtained in arestart isTLS connection over TCP to enable 0-RTT in QUIC. Similarly, QUIC can provide a pre-shared key thataccumulated statecan belost. Aside from progressused to enable 0-RTT in TCP. All the restrictions onincomplete requests,thestateuse of 0-RTT apply, and theHPACK header compression table couldcertificate MUST bequite valuable. Existing QUIC implementationsconsidered valid for both connections, which will use different protocol stacks and could use different port numbers. For instance, HTTP/1.1 and HTTP/2 operate over TLS and TCP, whereas QUIC operates over UDP. Source address validation is not completely portable between different protocol stacks. Even if theconnection ID to route packets tosource IP address remains constant, theserver thatport number ishandling the connection, which avoidslikely to be different. Packet reflection attacks are still possible in thissortsituation, though the set ofproblem.hosts that can initiate these attacks is greatly reduced. Alightweight state resurrection extensionserver mightbe usedchoose to avoidhavingsource address validation for such a connection, or allow an increase torecreate any expensive state. 8.the amount of data that it sends toward the client without source validation. 7. Security Considerations There are likely to be some real clangers here eventually, but the current set of issues is well captured in the relevant sections of the main text. Never assume that because it isn't in the security considerations section it doesn't affect security. Most of this document does.9.7.1. Packet Reflection Attack Mitigation A small ClientHello that results in a large block of handshake messages from a server can be used in packet reflection attacks to amplify the traffic generated by an attacker. Certificate caching [RFC7924] can reduce the size of the server's handshake messages significantly. A client SHOULD also pad [RFC7685] its ClientHello to at least 1024 octets (TODO: tune this value). A server is less likely to generate a packet reflection attack if the data it sends is a small multiple of the data it receives. A server SHOULD use a HelloRetryRequest if the size of the handshake messages it sends is likely to exceed the size of the ClientHello. 7.2. Peer Denial of Service QUIC, TLS and HTTP/2 all contain a messages that have legitimate uses in some contexts, but that can be abused to cause a peer to expend processing resources without having any observable impact on the state of the connection. If processing is disproportionately large in comparison to the observable effects on bandwidth or state, then this could allow a malicious peer to exhaust processing capacity without consequence. QUIC prohibits the sending of empty "STREAM" frames unless they are marked with the FIN bit. This prevents "STREAM" frames from being sent that only waste effort. TLS records SHOULD always contain at least one octet of a handshake messages or alert. Records containing only padding are permitted during the handshake, but an excessive number might be used to generate unnecessary work. Once the TLS handshake is complete, endpoints SHOULD NOT send TLS application data records unless it is to hide the length of QUIC records. QUIC packet protection does not include any allowance for padding; padded TLS application data records can be used to mask the length of QUIC frames. While there are legitimate uses for some redundant packets, implementations SHOULD track redundant packets and treat excessive volumes of any non-productive packets as indicative of an attack. 8. IANA Considerations This document has no IANA actions. Yet.10.9. References10.1.9.1. Normative References[I-D.ietf-tls-tls13] Rescorla, E., "The Transport Layer Security (TLS) Protocol Version 1.3", draft-ietf-tls-tls13-11 (work in progress), December 2015. [I-D.tsvwg-quic-protocol][I-D.hamilton-quic-transport-protocol] Hamilton, R., Iyengar, J., Swett, I., and A. Wilk, "QUIC: A UDP-BasedSecureMultiplexed andReliableSecure Transport", draft- hamilton-quic-transport-protocol-00 (work in progress), July 2016. [I-D.ietf-tls-tls13] Rescorla, E., "The Transportfor HTTP/2", draft-tsvwg-quic-protocol-02Layer Security (TLS) Protocol Version 1.3", draft-ietf-tls-tls13-17 (work in progress),JanuaryOctober 2016. [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate Requirement Levels", BCP 14, RFC 2119, DOI 10.17487/RFC2119, March 1997, <http://www.rfc-editor.org/info/rfc2119>. [RFC5116] McGrew, D., "An Interface and Algorithms for Authenticated Encryption", RFC 5116, DOI 10.17487/RFC5116, January 2008, <http://www.rfc-editor.org/info/rfc5116>. [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, <http://www.rfc-editor.org/info/rfc7301>.10.2.9.2. Informative References [RFC0793] Postel, J., "Transmission Control Protocol", STD 7, RFC 793, DOI 10.17487/RFC0793, September 1981, <http://www.rfc-editor.org/info/rfc793>. [RFC7230] Fielding, R., Ed. and J. Reschke, Ed., "Hypertext Transfer Protocol (HTTP/1.1): Message Syntax and Routing", RFC 7230, DOI 10.17487/RFC7230, June 2014, <http://www.rfc-editor.org/info/rfc7230>. [RFC7258] Farrell, S. and H. Tschofenig, "Pervasive Monitoring Is an Attack", BCP 188, RFC 7258, DOI 10.17487/RFC7258, May 2014, <http://www.rfc-editor.org/info/rfc7258>. [RFC7540] Belshe, M., Peon, R., and M. Thomson, Ed., "Hypertext Transfer Protocol Version 2 (HTTP/2)", RFC 7540, DOI 10.17487/RFC7540, May 2015, <http://www.rfc-editor.org/info/rfc7540>. [RFC7685] Langley, A., "A Transport Layer Security (TLS) ClientHello Padding Extension", RFC 7685, DOI 10.17487/RFC7685, October 2015, <http://www.rfc-editor.org/info/rfc7685>. [RFC7924] Santesson, S. and H. Tschofenig, "Transport Layer Security (TLS) Cached Information Extension", RFC 7924, DOI 10.17487/RFC7924, July 2016, <http://www.rfc-editor.org/info/rfc7924>. Appendix A. Acknowledgments Christian Huitema's knowledge of QUIC is far better than my own. This would be even more inaccurate and useless if not for his assistance. This document has variously benefited from a long series of discussions withRyan Hamilton,Jana Iyengar, Adam Langley, Roberto Peon, Eric Rescorla, Ian Swett, and likely many others who are merely forgotten by a faulty meat computer. Authors' Addresses Martin Thomson Mozilla Email: martin.thomson@gmail.com Ryan Hamilton Google Email: rch@google.com