draft-ietf-quic-manageability-10.txt   draft-ietf-quic-manageability-11.txt 
Network Working Group M. Kuehlewind Network Working Group M. Kuehlewind
Internet-Draft Ericsson Internet-Draft Ericsson
Intended status: Informational B. Trammell Intended status: Informational B. Trammell
Expires: 26 August 2021 Google Expires: 23 October 2021 Google Switzerland GmbH
22 February 2021 21 April 2021
Manageability of the QUIC Transport Protocol Manageability of the QUIC Transport Protocol
draft-ietf-quic-manageability-10 draft-ietf-quic-manageability-11
Abstract Abstract
This document discusses manageability of the QUIC transport protocol, This document discusses manageability of the QUIC transport protocol,
focusing on caveats impacting network operations involving QUIC focusing on the implications of QUIC's design and wire image on
traffic. Its intended audience is network operators, as well as network operations involving QUIC traffic. Its intended audience is
content providers that rely on the use of QUIC-aware middleboxes, network operators and equipment vendors who rely on the use of
e.g. for load balancing. transport-aware network functions.
Status of This Memo Status of This Memo
This Internet-Draft is submitted in full conformance with the This Internet-Draft is submitted in full conformance with the
provisions of BCP 78 and BCP 79. provisions of BCP 78 and BCP 79.
Internet-Drafts are working documents of the Internet Engineering Internet-Drafts are working documents of the Internet Engineering
Task Force (IETF). Note that other groups may also distribute Task Force (IETF). Note that other groups may also distribute
working documents as Internet-Drafts. The list of current Internet- working documents as Internet-Drafts. The list of current Internet-
Drafts is at https://datatracker.ietf.org/drafts/current/. Drafts is at https://datatracker.ietf.org/drafts/current/.
Internet-Drafts are draft documents valid for a maximum of six months Internet-Drafts are draft documents valid for a maximum of six months
and may be updated, replaced, or obsoleted by other documents at any and may be updated, replaced, or obsoleted by other documents at any
time. It is inappropriate to use Internet-Drafts as reference time. It is inappropriate to use Internet-Drafts as reference
material or to cite them other than as "work in progress." material or to cite them other than as "work in progress."
This Internet-Draft will expire on 26 August 2021. This Internet-Draft will expire on 23 October 2021.
Copyright Notice Copyright Notice
Copyright (c) 2021 IETF Trust and the persons identified as the Copyright (c) 2021 IETF Trust and the persons identified as the
document authors. All rights reserved. document authors. All rights reserved.
This document is subject to BCP 78 and the IETF Trust's Legal This document is subject to BCP 78 and the IETF Trust's Legal
Provisions Relating to IETF Documents (https://trustee.ietf.org/ Provisions Relating to IETF Documents (https://trustee.ietf.org/
license-info) in effect on the date of publication of this document. license-info) in effect on the date of publication of this document.
Please review these documents carefully, as they describe your rights Please review these documents carefully, as they describe your rights
skipping to change at page 2, line 17 skipping to change at page 2, line 17
1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 3 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 3
2. Features of the QUIC Wire Image . . . . . . . . . . . . . . . 4 2. Features of the QUIC Wire Image . . . . . . . . . . . . . . . 4
2.1. QUIC Packet Header Structure . . . . . . . . . . . . . . 4 2.1. QUIC Packet Header Structure . . . . . . . . . . . . . . 4
2.2. Coalesced Packets . . . . . . . . . . . . . . . . . . . . 6 2.2. Coalesced Packets . . . . . . . . . . . . . . . . . . . . 6
2.3. Use of Port Numbers . . . . . . . . . . . . . . . . . . . 6 2.3. Use of Port Numbers . . . . . . . . . . . . . . . . . . . 6
2.4. The QUIC Handshake . . . . . . . . . . . . . . . . . . . 7 2.4. The QUIC Handshake . . . . . . . . . . . . . . . . . . . 7
2.5. Integrity Protection of the Wire Image . . . . . . . . . 11 2.5. Integrity Protection of the Wire Image . . . . . . . . . 11
2.6. Connection ID and Rebinding . . . . . . . . . . . . . . . 11 2.6. Connection ID and Rebinding . . . . . . . . . . . . . . . 11
2.7. Packet Numbers . . . . . . . . . . . . . . . . . . . . . 12 2.7. Packet Numbers . . . . . . . . . . . . . . . . . . . . . 12
2.8. Version Negotiation and Greasing . . . . . . . . . . . . 12 2.8. Version Negotiation and Greasing . . . . . . . . . . . . 12
3. Network-visible Information about QUIC Flows . . . . . . . . 13 3. Network-Visible Information about QUIC Flows . . . . . . . . 12
3.1. Identifying QUIC Traffic . . . . . . . . . . . . . . . . 13 3.1. Identifying QUIC Traffic . . . . . . . . . . . . . . . . 13
3.1.1. Identifying Negotiated Version . . . . . . . . . . . 13 3.1.1. Identifying Negotiated Version . . . . . . . . . . . 13
3.1.2. Rejection of Garbage Traffic . . . . . . . . . . . . 14 3.1.2. First Packet Identification for Garbage Rejection . . 14
3.2. Connection Confirmation . . . . . . . . . . . . . . . . . 14 3.2. Connection Confirmation . . . . . . . . . . . . . . . . . 14
3.3. Distinguishing Acknowledgment traffic . . . . . . . . . . 15 3.3. Distinguishing Acknowledgment Traffic . . . . . . . . . . 15
3.4. Application Identification . . . . . . . . . . . . . . . 15 3.4. Server Name Indication (SNI) . . . . . . . . . . . . . . 15
3.4.1. Extracting Server Name Indication (SNI) 3.4.1. Extracting Server Name Indication (SNI)
Information . . . . . . . . . . . . . . . . . . . . . 15 Information . . . . . . . . . . . . . . . . . . . . . 15
3.5. Flow Association . . . . . . . . . . . . . . . . . . . . 17 3.5. Flow Association . . . . . . . . . . . . . . . . . . . . 17
3.6. Flow teardown . . . . . . . . . . . . . . . . . . . . . . 17 3.6. Flow Teardown . . . . . . . . . . . . . . . . . . . . . . 17
3.7. Flow Symmetry Measurement . . . . . . . . . . . . . . . . 17 3.7. Flow Symmetry Measurement . . . . . . . . . . . . . . . . 17
3.8. Round-Trip Time (RTT) Measurement . . . . . . . . . . . . 17 3.8. Round-Trip Time (RTT) Measurement . . . . . . . . . . . . 18
3.8.1. Measuring Initial RTT . . . . . . . . . . . . . . . . 18 3.8.1. Measuring Initial RTT . . . . . . . . . . . . . . . . 18
3.8.2. Using the Spin Bit for Passive RTT Measurement . . . 18 3.8.2. Using the Spin Bit for Passive RTT Measurement . . . 18
4. Specific Network Management Tasks . . . . . . . . . . . . . . 20 4. Specific Network Management Tasks . . . . . . . . . . . . . . 20
4.1. Stateful Treatment of QUIC Traffic . . . . . . . . . . . 20 4.1. Passive Network Performance Measurement and
4.2. Passive Network Performance Measurement and Troubleshooting . . . . . . . . . . . . . . . . . . . . 20
Troubleshooting . . . . . . . . . . . . . . . . . . . . . 21 4.2. Stateful Treatment of QUIC Traffic . . . . . . . . . . . 20
4.3. Server Cooperation with Load Balancers . . . . . . . . . 21 4.3. Address Rewriting to Ensure Routing Stability . . . . . . 22
4.4. DDoS Detection and Mitigation . . . . . . . . . . . . . . 21 4.4. Server Cooperation with Load Balancers . . . . . . . . . 22
4.5. UDP Policing . . . . . . . . . . . . . . . . . . . . . . 22 4.5. Filtering Behavior . . . . . . . . . . . . . . . . . . . 23
4.6. Handling ICMP Messages . . . . . . . . . . . . . . . . . 22 4.6. UDP Blocking or Throttling . . . . . . . . . . . . . . . 23
4.7. Quality of Service handling and ECMP . . . . . . . . . . 23 4.7. DDoS Detection and Mitigation . . . . . . . . . . . . . . 24
4.8. QUIC and Network Address Translation (NAT) . . . . . . . 23 4.8. Quality of Service handling and ECMP . . . . . . . . . . 25
4.8.1. Resource Conservation . . . . . . . . . . . . . . . . 24 4.9. Handling ICMP Messages . . . . . . . . . . . . . . . . . 26
4.8.2. "Helping" with routing infrastructure issues . . . . 25 4.10. Guiding Path MTU . . . . . . . . . . . . . . . . . . . . 26
4.9. Filtering behavior . . . . . . . . . . . . . . . . . . . 26 5. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 27
5. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 26 6. Security Considerations . . . . . . . . . . . . . . . . . . . 27
6. Security Considerations . . . . . . . . . . . . . . . . . . . 26 7. Contributors . . . . . . . . . . . . . . . . . . . . . . . . 28
7. Contributors . . . . . . . . . . . . . . . . . . . . . . . . 26 8. Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . 28
8. Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . 27 9. References . . . . . . . . . . . . . . . . . . . . . . . . . 28
9. References . . . . . . . . . . . . . . . . . . . . . . . . . 27 9.1. Normative References . . . . . . . . . . . . . . . . . . 28
9.1. Normative References . . . . . . . . . . . . . . . . . . 27 9.2. Informative References . . . . . . . . . . . . . . . . . 29
9.2. Informative References . . . . . . . . . . . . . . . . . 27 Appendix A. Distinguishing IETF QUIC and Google QUIC Versions . 32
Appendix A. Appendix . . . . . . . . . . . . . . . . . . . . . . 30 A.1. Extracting the CRYPTO frame . . . . . . . . . . . . . . . 33
A.1. Distinguishing IETF QUIC and Google QUIC Versions . . . . 30
A.2. Extracting the CRYPTO frame . . . . . . . . . . . . . . . 31 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 34
Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 32
1. Introduction 1. Introduction
QUIC [QUIC-TRANSPORT] is a new transport protocol that is QUIC [QUIC-TRANSPORT] is a new transport protocol that is
encapsulated in UDP. QUIC integrates TLS [QUIC-TLS] to encrypt all encapsulated in UDP. QUIC integrates TLS [QUIC-TLS] to encrypt all
payload data and most control information. QUIC version 1 was payload data and most control information. QUIC version 1 was
designed primarily as a transport for HTTP, with the resulting designed primarily as a transport for HTTP, with the resulting
protocol being known as HTTP/3 [QUIC-HTTP]. protocol being known as HTTP/3 [QUIC-HTTP].
Given that QUIC is an end-to-end transport protocol, all information
in the protocol header, even that which can be inspected, is not
meant to be mutable by the network, and is therefore integrity-
protected. While less information is visible to the network than for
TCP, integrity protection can also simplify troubleshooting, because
none of the nodes on the network path can modify the transport layer
information.
This document provides guidance for network operations that manage This document provides guidance for network operations that manage
QUIC traffic. This includes guidance on how to interpret and utilize QUIC traffic. This includes guidance on how to interpret and utilize
information that is exposed by QUIC to the network, requirements and information that is exposed by QUIC to the network, requirements and
assumptions that the QUIC design with respect to network treatment, assumptions of the QUIC design with respect to network treatment, and
and a description of how common network management practices will be a description of how common network management practices will be
impacted by QUIC. impacted by QUIC.
Since QUIC's wire image [WIRE-IMAGE] is integrity-protected, in- QUIC is an end-to-end transport protocol. No information in the
network operations that depend on modification of data are not protocol header, even that which can be inspected, is meant to be
possible without the cooperation of an endpoint. Network operation mutable by the network. This is achieved through integrity
practices that alter data are only possible if performed as a QUIC protection of the wire image [WIRE-IMAGE]. Encryption of most
endpoint; this might be possible with the introduction of a proxy control signaling means that less information is visible to the
which authenticates as an endpoint. Proxy operations are not in network than is the case with TCP.
scope for this document.
Integrity protection can also simplify troubleshooting, because none
of the nodes on the network path can modify transport layer
information. However, it does imply that in-network operations that
depend on modification of data are not possible without the
cooperation of an QUIC endpoint. This might be possible with the
introduction of a proxy which authenticates as an endpoint. Proxy
operations are not in scope for this document.
Network management is not a one-size-fits-all endeavour: practices Network management is not a one-size-fits-all endeavour: practices
considered necessary or even mandatory within enterprise networks considered necessary or even mandatory within enterprise networks
with certain compliance requirements, for example, would be with certain compliance requirements, for example, would be
impermissible on other networks without those requirements. This impermissible on other networks without those requirements. This
document therefore does not make any specific recommendations as to document therefore does not make any specific recommendations as to
which practices should or should not be applied; for each practice, which practices should or should not be applied; for each practice,
it describes what is and is not possible with the QUIC transport it describes what is and is not possible with the QUIC transport
protocol as defined. protocol as defined.
skipping to change at page 4, line 26 skipping to change at page 4, line 26
Negotiation Packet, are both inspectable and invariant Negotiation Packet, are both inspectable and invariant
[QUIC-INVARIANTS]. [QUIC-INVARIANTS].
This document describes version 1 of the QUIC protocol, whose wire This document describes version 1 of the QUIC protocol, whose wire
image is fully defined in [QUIC-TRANSPORT] and [QUIC-TLS]. Features image is fully defined in [QUIC-TRANSPORT] and [QUIC-TLS]. Features
of the wire image described herein may change in future versions of of the wire image described herein may change in future versions of
the protocol, except when specified as an invariant the protocol, except when specified as an invariant
[QUIC-INVARIANTS], and cannot be used to identify QUIC as a protocol [QUIC-INVARIANTS], and cannot be used to identify QUIC as a protocol
or to infer the behavior of future versions of QUIC. or to infer the behavior of future versions of QUIC.
Appendix A.1 provides non-normative guidance on the identification of Appendix A provides non-normative guidance on the identification of
QUIC version 1 packets compared to some pre-standard versions. QUIC version 1 packets compared to some pre-standard versions.
2.1. QUIC Packet Header Structure 2.1. QUIC Packet Header Structure
QUIC packets may have either a long header, or a short header. The QUIC packets may have either a long header or a short header. The
first bit of the QUIC header is the Header Form bit, and indicates first bit of the QUIC header is the Header Form bit, and indicates
which type of header is present. The purpose of this bit is which type of header is present. The purpose of this bit is
invariant across QUIC versions. invariant across QUIC versions.
The long header exposes more information. It is used during The long header exposes more information. In version 1 of QUIC, it
connection establishment, including version negotiation, retry, and is used during connection establishment, including version
0-RTT data. It contains a version number, as well as source and negotiation, retry, and 0-RTT data. It contains a version number, as
destination connection IDs for grouping packets belonging to the same well as source and destination connection IDs for grouping packets
flow. The definition and location of these fields in the QUIC long belonging to the same flow. The definition and location of these
header are invariant for future versions of QUIC, although future fields in the QUIC long header are invariant for future versions of
versions of QUIC may provide additional fields in the long header QUIC, although future versions of QUIC may provide additional fields
[QUIC-INVARIANTS]. in the long header [QUIC-INVARIANTS].
Short headers are used after connection establishment, and contain
only an optional destination connection ID and the spin bit for RTT
measurement.
The following information is exposed in QUIC packet headers:
* "fixed bit": the second most significant bit of the first octet
most QUIC packets of the current version is currently set to 1,
for endpoints to demultiplex with other UDP-encapsulated
protocols. Even thought this bit is fixed in the QUICv1
specification, endpoints may use a version or extension that
varies the bit. Therefore, observers cannot depend on it as an
identifier for QUIC.
* latency spin bit: the third most significant bit of first octet in Short headers contain only an optional destination connection ID and
the short packet header. The spin bit is set by endpoints such the spin bit for RTT measurement. In version 1 of QUIC, they are
that tracking edge transitions can be used to passively observe used after connection establishment.
end-to-end RTT. See Section 3.8.2 for further details.
* header type: the long header has a 2 bit packet type field The following information is exposed in QUIC packet headers in all
following the Header Form and fixed bits. Header types correspond versions of QUIC:
to stages of the handshake; see Section 17.2 of [QUIC-TRANSPORT]
for details.
* version number: the version number is present in the long header, * version number: the version number is present in the long header,
and identifies the version used for that packet. During Version and identifies the version used for that packet. During Version
Negotiation (see Section 2.8 and Section 17.2.1 of Negotiation (see Section 17.2.1 of [QUIC-TRANSPORT] and
[QUIC-TRANSPORT]), the version number field has a special value Section 2.8), the version number field has a special value
(0x00000000) that identifies the packet as a Version Negotiation (0x00000000) that identifies the packet as a Version Negotiation
packet. Upon time of publishing of this document, QUIC versions packet. QUIC version 1 uses version 0x00000001. Operators should
that start with 0xff implement IETF drafts. QUIC version 1 uses expect to observe packets with other version numbers as a result
version 0x00000001. Operators should expect to observe packets of various Internet experiments, future standards, and greasing.
with other version numbers as a result of various Internet All deployed versions are maintained in an IANA registry (see
experiments and future standards. Section 22.2 of [QUIC-TRANSPORT]).
* source and destination connection ID: short and long packet * source and destination connection ID: short and long packet
headers carry a destination connection ID, a variable-length field headers carry a destination connection ID, a variable-length field
that can be used to identify the connection associated with a QUIC that can be used to identify the connection associated with a QUIC
packet, for load-balancing and NAT rebinding purposes; see packet, for load-balancing and NAT rebinding purposes; see
Section 4.3 and Section 2.6. Long packet headers additionally Section 4.4 and Section 2.6. Long packet headers additionally
carry a source connection ID. The source connection ID carry a source connection ID. The source connection ID
corresponds to the destination connection ID the source would like corresponds to the destination connection ID the source would like
to have on packets sent to it, and is only present on long packet to have on packets sent to it, and is only present on long packet
headers. On long header packets, the length of the connection IDs headers. On long header packets, the length of the connection IDs
is also present; on short header packets, the length of the is also present; on short header packets, the length of the
destination connection ID is implicit. destination connection ID is implicit.
* length: the length of the remaining QUIC packet after the length In version 1 of QUIC, the following additional information is
exposed:
* "fixed bit": The second-most-significant bit of the first octet of
most QUIC packets of the current version is set to 1, enabling
endpoints to demultiplex with other UDP-encapsulated protocols.
Even though this bit is fixed in the version 1 specification,
endpoints might use an extension that varies the bit. Therefore,
observers cannot reliably use it as an identifier for QUIC.
* latency spin bit: The third-most-significant bit of the first
octet in the short packet header for version 1. The spin bit is
set by endpoints such that tracking edge transitions can be used
to passively observe end-to-end RTT. See Section 3.8.2 for
further details.
* header type: The long header has a 2 bit packet type field
following the Header Form and fixed bits. Header types correspond
to stages of the handshake; see Section 17.2 of [QUIC-TRANSPORT]
for details.
* length: The length of the remaining QUIC packet after the length
field, present on long headers. This field is used to implement field, present on long headers. This field is used to implement
coalesced packets during the handshake (see Section 2.2). coalesced packets during the handshake (see Section 2.2).
* token: Initial packets may contain a token, a variable-length * token: Initial packets may contain a token, a variable-length
opaque value optionally sent from client to server, used for opaque value optionally sent from client to server, used for
validating the client's address. Retry packets also contain a validating the client's address. Retry packets also contain a
token, which can be used by the client in an Initial packet on a token, which can be used by the client in an Initial packet on a
subsequent connection attempt. The length of the token is subsequent connection attempt. The length of the token is
explicit in both cases. explicit in both cases.
Retry (Section 17.2.5 of [QUIC-TRANSPORT]) and Version Negotiation Retry (Section 17.2.5 of [QUIC-TRANSPORT]) and Version Negotiation
(Section 17.2.1 of [QUIC-TRANSPORT]) packets are not encrypted or (Section 17.2.1 of [QUIC-TRANSPORT]) packets are not encrypted or
obfuscated in any way. For other kinds of packets, other information obfuscated in any way. For other kinds of packets, version 1 of QUIC
in the packet headers is cryptographically obfuscated: cryptographically obfuscates other information in the packet headers:
* packet number: All packets except Version Negotiation and Retry * packet number: All packets except Version Negotiation and Retry
packets have an associated packet number; however, this packet packets have an associated packet number; however, this packet
number is encrypted, and therefore not of use to on-path number is encrypted, and therefore not of use to on-path
observers. The offset of the packet number is encoded in long observers. The offset of the packet number is encoded in long
headers, while it is implicit (depending on destination connection headers, while it is implicit (depending on destination connection
ID length) in short headers. The length of the packet number is ID length) in short headers. The length of the packet number is
cryptographically obfuscated. cryptographically obfuscated.
* key phase: The Key Phase bit, present in short headers, specifies * key phase: The Key Phase bit, present in short headers, specifies
skipping to change at page 6, line 49 skipping to change at page 7, line 7
to use the same port number for both services. However, as for all to use the same port number for both services. However, as for all
other IETF transports [RFC7605], there is no guarantee that a other IETF transports [RFC7605], there is no guarantee that a
specific application will use a given registered port, or that a specific application will use a given registered port, or that a
given port carries traffic belonging to the respective registered given port carries traffic belonging to the respective registered
service, especially when application layer information is encrypted. service, especially when application layer information is encrypted.
For example, [QUIC-HTTP] specifies the use of Alt-Svc for discovery For example, [QUIC-HTTP] specifies the use of Alt-Svc for discovery
of HTTP/3 services on other ports. of HTTP/3 services on other ports.
Further, as QUIC has a connection ID, it is also possible to maintain Further, as QUIC has a connection ID, it is also possible to maintain
multiple QUIC connections over one 5-tuple. However, if the multiple QUIC connections over one 5-tuple. However, if the
connection ID is not present in the packet header, all packets of the connection ID is zero-length, all packets of the 5-tuple belong to
5-tuple belong to the same QUIC connection. the same QUIC connection.
2.4. The QUIC Handshake 2.4. The QUIC Handshake
New QUIC connections are established using a handshake, which is New QUIC connections are established using a handshake, which is
distinguishable on the wire and contains some information that can be distinguishable on the wire and contains some information that can be
passively observed. passively observed.
To illustrate the information visible in the QUIC wire image during To illustrate the information visible in the QUIC wire image during
the handshake, we first show the general communication pattern the handshake, we first show the general communication pattern
visible in the UDP datagrams containing the QUIC handshake, then visible in the UDP datagrams containing the QUIC handshake, then
examine each of the datagrams in detail. examine each of the datagrams in detail.
The QUIC handshake can normally be recognized on the wire through at The QUIC handshake can normally be recognized on the wire through at
least four datagrams we'll call "QUIC Client Hello", "QUIC Server least four datagrams we'll call "Client Initial", "Server Initial",
Hello", and "Initial Completion", and "Handshake Completion", for and "Client Completion", and "Server Completion", for purposes of
purposes of this illustration, as shown in Figure 1. this illustration, as shown in Figure 1.
Packets in the handshake belong to three separate cryptographic and Packets in the handshake belong to three separate cryptographic and
transport contexts ("Initial", which contains observable payload, and transport contexts ("Initial", which contains observable payload, and
"Handshake" and "1-RTT", which do not). QUIC packets in separate "Handshake" and "1-RTT", which do not). QUIC packets in separate
contexts during the handshake are generally coalesced (see contexts during the handshake are generally coalesced (see
Section 2.2) in order to reduce the number of UDP datagrams sent Section 2.2) in order to reduce the number of UDP datagrams sent
during the handshake. during the handshake.
As shown here, the client can send 0-RTT data as soon as it has sent As shown here, the client can send 0-RTT data as soon as it has sent
its Client Hello, and the server can send 1-RTT data as soon as it its Client Hello, and the server can send 1-RTT data as soon as it
has sent its Server Hello. has sent its Server Hello.
Client Server Client Server
| | | |
+----QUIC Client Hello-------------------->| +----Client Initial----------------------->|
+----(zero or more 0RTT)------------------>| +----(zero or more 0RTT)------------------>|
| | | |
|<--------------------QUIC Server Hello----+ |<-----------------------Server Initial----+
|<---------(1RTT encrypted data starts)----+ |<---------(1RTT encrypted data starts)----+
| | | |
+----Initial Completion------------------->| +----Client Completion-------------------->|
+----(1RTT encrypted data starts)--------->| +----(1RTT encrypted data starts)--------->|
| | | |
|<-----------------Handshake Completion----+ |<--------------------Server Completion----+
| | | |
Figure 1: General communication pattern visible in the QUIC handshake Figure 1: General communication pattern visible in the QUIC handshake
A typical handshake starts with the client sending of a QUIC Client A typical handshake starts with the client sending of a Client
Hello datagram as shown in Figure 2, which elicits a QUIC Server Initial datagram as shown in Figure 2, which elicits a Server Initial
Hello datagram as shown in Figure 3 typically containing three datagram as shown in Figure 3 typically containing three packets: an
packets: an Initial packet with the Server Hello, a Handshake packet Initial packet with the Server Initial, a Handshake packet with the
with the rest of the server's side of the TLS handshake, and initial rest of the server's side of the TLS handshake, and initial 1-RTT
1-RTT data, if present. data, if present.
The Initial Completion datagram contains at least one Handshake The Client Completion datagram contains at least one Handshake packet
packet and some also include an Initial packet. and some also include an Initial packet.
Datagrams that contain a QUIC Initial Packet (Client Hello, Server Datagrams that contain a Client Initial Packet (Client Initial,
Hello, and some Initial Completion) contain at least 1200 octets of Server Initial, and some Client Completion) contain at least 1200
UDP payload. This protects against amplification attacks and octets of UDP payload. This protects against amplification attacks
verifies that the network path meets the requirements for the minimum and verifies that the network path meets the requirements for the
QUIC IP packet size, see Section 14 of [QUIC-TRANSPORT]. This is minimum QUIC IP packet size; see Section 14 of [QUIC-TRANSPORT].
accomplished by either adding PADDING frames within the Initial This is accomplished by either adding PADDING frames within the
packet, coalescing other packets with the Initial packet, or leaving Initial packet, coalescing other packets with the Initial packet, or
unused payload in the UDP packet after the Initial packet. A network leaving unused payload in the UDP packet after the Initial packet. A
path needs to be able to forward at least this size of packet for network path needs to be able to forward at least this size of packet
QUIC to be used. for QUIC to be used.
The content of QUIC Initial packets are encrypted using Initial The content of Client Initial packets are encrypted using Initial
Secrets, which are derived from a per-version constant and the Secrets, which are derived from a per-version constant and the
client's destination connection ID; they are therefore observable by client's destination connection ID; they are therefore observable by
any on-path device that knows the per-version constant. They are any on-path device that knows the per-version constant. They are
therefore considered visible in this illustration. The content of therefore considered visible in this illustration. The content of
QUIC Handshake packets are encrypted using keys established during QUIC Handshake packets are encrypted using keys established during
the initial handshake exchange, and are therefore not visible. the initial handshake exchange, and are therefore not visible.
Initial, Handshake, and the Short Header packets transmitted after Initial, Handshake, and the Short Header packets transmitted after
the handshake belong to cryptographic and transport contexts. The the handshake belong to cryptographic and transport contexts. The
Initial Completion Figure 4 and the Handshake Completion Figure 5 Client Completion Figure 4 and the Server Completion Figure 5
datagrams finish these first two contexts, by sending the final datagrams finish these first two contexts, by sending the final
acknowledgment and finishing the transmission of CRYPTO frames. acknowledgment and finishing the transmission of CRYPTO frames.
+----------------------------------------------------------+ +----------------------------------------------------------+
| UDP header (source and destination UDP ports) | | UDP header (source and destination UDP ports) |
+----------------------------------------------------------+ +----------------------------------------------------------+
| QUIC long header (type = Initial, Version, DCID, SCID) (Length) | QUIC long header (type = Initial, Version, DCID, SCID) (Length)
+----------------------------------------------------------+ | +----------------------------------------------------------+ |
| QUIC CRYPTO frame header | | | QUIC CRYPTO frame header | |
+----------------------------------------------------------+ | +----------------------------------------------------------+ |
| TLS Client Hello (incl. TLS SNI) | | | TLS Client Hello (incl. TLS SNI) | |
+----------------------------------------------------------+ | +----------------------------------------------------------+ |
| QUIC PADDING frames | | | QUIC PADDING frames | |
+----------------------------------------------------------+<-+ +----------------------------------------------------------+<-+
Figure 2: Typical QUIC Client Hello datagram pattern with no 0-RTT Figure 2: Typical Client Initial datagram pattern without 0-RTT
The Client Hello datagram exposes version number, source and The Client Initial datagram exposes version number, source and
destination connection IDs without encryption. Information in the destination connection IDs without encryption. Information in the
TLS Client Hello frame, including any TLS Server Name Indication TLS Client Hello frame, including any TLS Server Name Indication
(SNI) present, is obfuscated using the Initial secret. Note that the (SNI) present, is obfuscated using the Initial secret. Note that the
location of PADDING is implementation-dependent, and PADDING frames location of PADDING is implementation-dependent, and PADDING frames
may not appear in a coalesced Initial packet. might not appear in a coalesced Initial packet.
+------------------------------------------------------------+ +------------------------------------------------------------+
| UDP header (source and destination UDP ports) | | UDP header (source and destination UDP ports) |
+------------------------------------------------------------+ +------------------------------------------------------------+
| QUIC long header (type = Initial, Version, DCID, SCID) (Length) | QUIC long header (type = Initial, Version, DCID, SCID) (Length)
+------------------------------------------------------------+ | +------------------------------------------------------------+ |
| QUIC CRYPTO frame header | | | QUIC CRYPTO frame header | |
+------------------------------------------------------------+ | +------------------------------------------------------------+ |
| TLS Server Hello | | | TLS Server Hello | |
+------------------------------------------------------------+ | +------------------------------------------------------------+ |
skipping to change at page 9, line 32 skipping to change at page 9, line 32
+------------------------------------------------------------+<-+ +------------------------------------------------------------+<-+
| QUIC long header (type = Handshake, Version, DCID, SCID) (Length) | QUIC long header (type = Handshake, Version, DCID, SCID) (Length)
+------------------------------------------------------------+ | +------------------------------------------------------------+ |
| encrypted payload (presumably CRYPTO frames) | | | encrypted payload (presumably CRYPTO frames) | |
+------------------------------------------------------------+<-+ +------------------------------------------------------------+<-+
| QUIC short header | | QUIC short header |
+------------------------------------------------------------+ +------------------------------------------------------------+
| 1-RTT encrypted payload | | 1-RTT encrypted payload |
+------------------------------------------------------------+ +------------------------------------------------------------+
Figure 3: Typical QUIC Server Hello datagram pattern Figure 3: Typical Server Initial datagram pattern
The Server Hello datagram also exposes version number, source and The Server Initial datagram also exposes version number, source and
destination connection IDs and information in the TLS Server Hello destination connection IDs in the clear; information in the TLS
message which is obfuscated using the Initial secret. Server Hello message is obfuscated using the Initial secret.
+------------------------------------------------------------+ +------------------------------------------------------------+
| UDP header (source and destination UDP ports) | | UDP header (source and destination UDP ports) |
+------------------------------------------------------------+ +------------------------------------------------------------+
| QUIC long header (type = Initial, Version, DCID, SCID) (Length) | QUIC long header (type = Initial, Version, DCID, SCID) (Length)
+------------------------------------------------------------+ | +------------------------------------------------------------+ |
| QUIC ACK frame (acknowledging Server Hello Initial) | | | QUIC ACK frame (acknowledging Server Initial Initial) | |
+------------------------------------------------------------+<-+ +------------------------------------------------------------+<-+
| QUIC long header (type = Handshake, Version, DCID, SCID) (Length) | QUIC long header (type = Handshake, Version, DCID, SCID) (Length)
+------------------------------------------------------------+ | +------------------------------------------------------------+ |
| encrypted payload (presumably CRYPTO/ACK frames) | | | encrypted payload (presumably CRYPTO/ACK frames) | |
+------------------------------------------------------------+<-+ +------------------------------------------------------------+<-+
| QUIC short header | | QUIC short header |
+------------------------------------------------------------+ +------------------------------------------------------------+
| 1-RTT encrypted payload | | 1-RTT encrypted payload |
+------------------------------------------------------------+ +------------------------------------------------------------+
Figure 4: Typical QUIC Initial Completion datagram pattern Figure 4: Typical Client Completion datagram pattern
The Initial Completion datagram does not expose any additional The Client Completion datagram does not expose any additional
information; however, recognizing it can be used to determine that a information; however, recognizing it can be used to determine that a
handshake has completed (see Section 3.2), and for three-way handshake has completed (see Section 3.2), and for three-way
handshake RTT estimation as in Section 3.8. handshake RTT estimation as in Section 3.8.
+------------------------------------------------------------+ +------------------------------------------------------------+
| UDP header (source and destination UDP ports) | | UDP header (source and destination UDP ports) |
+------------------------------------------------------------+ +------------------------------------------------------------+
| QUIC long header (type = Handshake, Version, DCID, SCID) (Length) | QUIC long header (type = Handshake, Version, DCID, SCID) (Length)
+------------------------------------------------------------+ | +------------------------------------------------------------+ |
| encrypted payload (presumably ACK frame) | | | encrypted payload (presumably ACK frame) | |
+------------------------------------------------------------+<-+ +------------------------------------------------------------+<-+
| QUIC short header | | QUIC short header |
+------------------------------------------------------------+ +------------------------------------------------------------+
| 1-RTT encrypted payload | | 1-RTT encrypted payload |
+------------------------------------------------------------+ +------------------------------------------------------------+
Figure 5: Typical QUIC Handshake Completion datagram pattern Figure 5: Typical Server Completion datagram pattern
Similar to Initial Completion, Handshake Completion also exposes no Similar to Client Completion, Server Completion also exposes no
additional information; observing it serves only to determine that additional information; observing it serves only to determine that
the handshake has completed. the handshake has completed.
When the client uses 0-RTT connection resumption, 0-RTT data may also When the client uses 0-RTT connection resumption, 0-RTT data may also
be seen in the QUIC Client Hello datagram, as shown in Figure 6. be seen in the Client Initial datagram, as shown in Figure 6.
+----------------------------------------------------------+ +----------------------------------------------------------+
| UDP header (source and destination UDP ports) | | UDP header (source and destination UDP ports) |
+----------------------------------------------------------+ +----------------------------------------------------------+
| QUIC long header (type = Initial, Version, DCID, SCID) (Length) | QUIC long header (type = Initial, Version, DCID, SCID) (Length)
+----------------------------------------------------------+ | +----------------------------------------------------------+ |
| QUIC CRYPTO frame header | | | QUIC CRYPTO frame header | |
+----------------------------------------------------------+ | +----------------------------------------------------------+ |
| TLS Client Hello (incl. TLS SNI) | | | TLS Client Hello (incl. TLS SNI) | |
+----------------------------------------------------------+<-+ +----------------------------------------------------------+<-+
| QUIC long header (type = 0RTT, Version, DCID, SCID) (Length) | QUIC long header (type = 0RTT, Version, DCID, SCID) (Length)
+----------------------------------------------------------+ | +----------------------------------------------------------+ |
| 0-rtt encrypted payload | | | 0-rtt encrypted payload | |
+----------------------------------------------------------+<-+ +----------------------------------------------------------+<-+
Figure 6: Typical 0-RTT QUIC Client Hello datagram pattern Figure 6: Typical 0-RTT Client Initial datagram pattern
In a 0-RTT QUIC Client Hello datagram, the PADDING frame is only In a 0-RTT Client Initial datagram, the PADDING frame is only present
present if necessary to increase the size of the datagram with 0RTT if necessary to increase the size of the datagram with 0RTT data to
data to at least 1200 bytes. Additional datagrams containing only at least 1200 bytes. Additional datagrams containing only 0-RTT
0-RTT protected long header packets may be sent from the client to protected long header packets may be sent from the client to the
the server after the Client Hello datagram, containing the rest of server after the Client Initial datagram, containing the rest of the
the 0-RTT data. The amount of 0-RTT protected data that can be sent 0-RTT data. The amount of 0-RTT protected data that can be sent in
in the first round is limited by the initial congestion window, the first round is limited by the initial congestion window,
typically around 10 packets (see Section 7.2 of [QUIC-RECOVERY]). typically around 10 packets (see Section 7.2 of [QUIC-RECOVERY]).
2.5. Integrity Protection of the Wire Image 2.5. Integrity Protection of the Wire Image
As soon as the cryptographic context is established, all information As soon as the cryptographic context is established, all information
in the QUIC header, including exposed information, is integrity in the QUIC header, including exposed information, is integrity-
protected. Further, information that was sent and exposed in protected. Further, information that was exposed in packets sent
handshake packets sent before the cryptographic context was before the cryptographic context was established is validated during
established are validated later during the cryptographic handshake. the cryptographic handshake. Therefore, devices on path cannot alter
Therefore, devices on path cannot alter any information or bits in any information or bits in QUIC packets. Such alterations would
QUIC packets. Such alterations would cause the integrity check to cause the integrity check to fail, which results in the receiver
fail, which results in the receiver discarding the packet. Some discarding the packet. Some parts of Initial packets could be
parts of Initial packets could be altered by removing and re-applying altered by removing and re-applying the authenticated encryption
the authenticated encryption without immediate discard at the without immediate discard at the receiver. However, the
receiver. However, the cryptographic handshake validates most fields cryptographic handshake validates most fields and any modifications
and any modifications in those fields will result in connection in those fields will result in connection establishment failing later
establishment failing later on. on.
2.6. Connection ID and Rebinding 2.6. Connection ID and Rebinding
The connection ID in the QUIC packet headers allows routing of QUIC The connection ID in the QUIC packet headers allows association of
packets at load balancers on other than five-tuple information, QUIC packets using information independent of the five-tuple. This
ensuring that related flows are appropriately balanced together; and allows rebinding of a connection after one of one endpoint
to allow rebinding of a connection after one of the endpoint's experienced an address change - usually the client. Further it can
addresses changes - usually the client's. Client and server be used by in-network devices to ensure that related 5-tuple flows
negotiate connection IDs during the handshake; typically, however, are appropriately balanced together.
only the server will request a connection ID for the lifetime of the
connection. Connection IDs for either endpoint may change during the
lifetime of a connection, with the new connection ID being negotiated
via encrypted frames. See Section 5.1 of [QUIC-TRANSPORT].
Therefore, observing a new connection ID does not necessary indicate
a new connection.
Server-generated connection IDs should seek to obscure any encoding,
of routing identities or any other information. Exposing the server
mapping would allow linkage of multiple IP addresses to the same host
if the server also supports migration. Furthermore, this opens an
attack vector on specific servers or pools.
The best way to obscure an encoding is to appear random to observers, Client and server negotiate connection IDs during the handshake;
which is most rigorously achieved with encryption. Even when typically, however, only the server will request a connection ID for
encrypted, a scheme could embed the unencrypted length of the the lifetime of the connection. Connection IDs for either endpoint
connection ID in the connection ID itself, instead of remembering it. may change during the lifetime of a connection, with the new
connection ID being supplied via encrypted frames (see Section 5.1 of
[QUIC-TRANSPORT]). Therefore, observing a new connection ID does not
necessary indicate a new connection.
[QUIC_LB] further specified possible algorithms to generate [QUIC_LB] specifies algorithms for encoding the server mapping in a
connection IDs at load balancers. connection ID in order to share this information with selected on-
path devices such as load balancers. Server mappings should only be
exposed to selected entities. Uncontrolled exposure would allow
linkage of multiple IP addresses to the same host if the server also
supports migration which opens an attack vector on specific servers
or pools. The best way to obscure an encoding is to appear random to
any other observers, which is most rigorously achieved with
encryption. As a result any attempt to infer information from
specific parts of a connection ID is unlikely to be useful.
2.7. Packet Numbers 2.7. Packet Numbers
The packet number field is always present in the QUIC packet header; The packet number field is always present in the QUIC packet header
however, it is always encrypted. The encryption key for packet in version 1; however, it is always encrypted. The encryption key
number protection on handshake packets sent before cryptographic for packet number protection on handshake packets sent before
context establishment is specific to the QUIC version, while packet cryptographic context establishment is specific to the QUIC version,
number protection on subsequent packets uses secrets derived from the while packet number protection on subsequent packets uses secrets
end-to-end cryptographic context. Packet numbers are therefore not derived from the end-to-end cryptographic context. Packet numbers
part of the wire image that is visible to on-path observers. are therefore not part of the wire image that is visible to on-path
observers.
2.8. Version Negotiation and Greasing 2.8. Version Negotiation and Greasing
Version Negotiation packets are used by the server to indicate that a Version Negotiation packets are used by the server to indicate that a
requested version from the client is not supported (see section 6 of requested version from the client is not supported (see Section 6 of
[QUIC-TRANSPORT]. Version Negotiation packets are not intrinsically [QUIC-TRANSPORT]. Version Negotiation packets are not intrinsically
protected, but QUIC versions can use later encrypted messages to protected, but future QUIC versions will use later encrypted messages
verify that they were authentic. Therefore any modification of this to verify that they were authentic. Therefore any modification of
list will be detected and may cause the endpoints to terminate the this list will be detected and may cause the endpoints to terminate
connection attempt. the connection attempt.
Also note that the list of versions in the Version Negotiation packet Also note that the list of versions in the Version Negotiation packet
may contain reserved versions. This mechanism is used to avoid may contain reserved versions. This mechanism is used to avoid
ossification in the implementation on the selection mechanism. ossification in the implementation on the selection mechanism.
Further, a client may send a Initial Client packet with a reserved Further, a client may send a Initial Client packet with a reserved
version number to trigger version negotiation. In the Version version number to trigger version negotiation. In the Version
Negotiation packet the connection ID and packet number of the Client Negotiation packet, the connection IDs of the Client Initial packet
Initial packet are reflected to provide a proof of return- are reflected to provide a proof of return-routability. Therefore,
routability. Therefore changing this information will also cause the changing this information will also cause the connection to fail.
connection to fail.
QUIC is expected to evolve rapidly, so new versions, both QUIC is expected to evolve rapidly, so new versions, both
experimental and IETF standard versions, will be deployed in the experimental and IETF standard versions, will be deployed in the
Internet more often than with traditional Internet- and transport- Internet more often than with traditional Internet- and transport-
layer protocols. Using a particular version number to recognize layer protocols. Using a particular version number to recognize
valid QUIC traffic is likely to persistently miss a fraction of QUIC valid QUIC traffic is likely to persistently miss a fraction of QUIC
flows and completely fail in the near future, and is therefore not flows and completely fail in the near future, and is therefore not
recommended. In addition, due to the speed of evolution of the recommended. In addition, due to the speed of evolution of the
protocol, devices that attempt to distinguish QUIC traffic from non- protocol, devices that attempt to distinguish QUIC traffic from non-
QUIC traffic for purposes of network admission control should admit QUIC traffic for purposes of network admission control should admit
all QUIC traffic regardless of version. all QUIC traffic regardless of version.
3. Network-visible Information about QUIC Flows 3. Network-Visible Information about QUIC Flows
This section addresses the different kinds of observations and This section addresses the different kinds of observations and
inferences that can be made about QUIC flows by a passive observer in inferences that can be made about QUIC flows by a passive observer in
the network based on the wire image in Section 2. Here we assume a the network based on the wire image in Section 2. Here we assume a
bidirectional observer (one that can see packets in both directions bidirectional observer (one that can see packets in both directions
in the sequence in which they are carried on the wire) unless noted. in the sequence in which they are carried on the wire) unless noted.
3.1. Identifying QUIC Traffic 3.1. Identifying QUIC Traffic
The QUIC wire image is not specifically designed to be The QUIC wire image is not specifically designed to be
distinguishable from other UDP traffic. distinguishable from other UDP traffic.
The only application binding defined by the IETF QUIC WG is HTTP/3 The only application binding defined by the IETF QUIC WG is HTTP/3
[QUIC-HTTP] at the time of this writing; however, many other [QUIC-HTTP] at the time of this writing; however, many other
applications are currently being defined and deployed over QUIC, so applications are currently being defined and deployed over QUIC, so
an assumption that all QUIC traffic is HTTP/3 is not valid. HTTP an assumption that all QUIC traffic is HTTP/3 is not valid. HTTP/3
over QUIC uses UDP port 443 by default, although URLs referring to uses UDP port 443 by default, although URLs referring to resources
resources available over HTTP/3 may specify alternate port numbers. available over HTTP/3 may specify alternate port numbers. Simple
Simple assumptions about whether a given flow is using QUIC based assumptions about whether a given flow is using QUIC based upon a UDP
upon a UDP port number may therefore not hold; see also [RFC7605] port number may therefore not hold; see also Section 5 of [RFC7605].
section 5.
While the second most significant bit (0x40) of the first octet is While the second-most-significant bit (0x40) of the first octet is
set to 1 in most QUIC packets of the current version (see Section 2.1 set to 1 in most QUIC packets of the current version (see Section 2.1
and section 17 of [QUIC-TRANSPORT]), this method of recognizing QUIC and Section 17 of [QUIC-TRANSPORT]), this method of recognizing QUIC
traffic is not reliable. First, it only provides one bit of traffic is not reliable. First, it only provides one bit of
information and is prone to collision with UDP-based protocols other information and is prone to collision with UDP-based protocols other
than those that this static bit is meant to allow multiplexing with. than those considered in [RFC7983]. Second, this feature of the wire
Second, this feature of the wire image is not invariant image is not invariant [QUIC-INVARIANTS] and may change in future
[QUIC-INVARIANTS] and may change in future versions of the protocol, versions of the protocol, or even be negotiated during the handshake
or even be negotiated during the handshake via the use of transport via the use of an extension.
parameters.
Even though transport parameters transmitted in the client initial Even though transport parameters transmitted in the client's Initial
are obserable by the network, they cannot be modified by the network packet are observable by the network, they cannot be modified by the
without risking connection failure. Further, the negotiated reply network without risking connection failure. Further, the reply from
from the server cannot be observed, so observers on the network the server cannot be observed, so observers on the network cannot
cannot know which parameters are actually in use. know which parameters are actually in use.
3.1.1. Identifying Negotiated Version 3.1.1. Identifying Negotiated Version
An in-network observer assuming that a set of packets belongs to a An in-network observer assuming that a set of packets belongs to a
QUIC flow can infer the version number in use by observing the QUIC flow can infer the version number in use by observing the
handshake: for QUIC version 1 if the version number in the Initial handshake: for QUIC version 1, if the version number in the Initial
packet from a client is the same as the version number in Initial packet from a client is the same as the version number in the Initial
packet of the server response, that version has been accepted by both packet of the server response, that version has been accepted by both
endpoints to be used for the rest of the connection. endpoints to be used for the rest of the connection.
Negotiated version cannot be identified for flows for which a The negotiated version cannot be identified for flows for which a
handshake is not observed, such as in the case of connection handshake is not observed, such as in the case of connection
migration; however, it might be possible to associate a flow with a migration; however, it might be possible to associate a flow with a
flow for which a version has been identified; see Section 3.5. flow for which a version has been identified; see Section 3.5.
This document focuses on QUIC Version 1, and this section applies 3.1.2. First Packet Identification for Garbage Rejection
only to packets belonging to Version 1 QUIC flows; for purposes of
on-path observation, it assumes that these packets have been
identified as such through the observation of a version number
exchange as described above.
3.1.2. Rejection of Garbage Traffic
A related question is whether a first packet of a given flow on a A related question is whether the first packet of a given flow on a
known QUIC-associated port is a valid QUIC packet, to support in- port known to be associated with QUIC is a valid QUIC packet. This
network filtering of garbage UDP packets (reflection attacks, random determination supports in-network filtering of garbage UDP packets
backscatter). While heuristics based on the first byte of the packet (reflection attacks, random backscatter, etc.). While heuristics
(packet type) could be used to separate valid from invalid first based on the first byte of the packet (packet type) could be used to
packet types, the deployment of such heuristics is not recommended, separate valid from invalid first packet types, the deployment of
as packet types may have different meanings in future versions of the such heuristics is not recommended, as bits in the first byte may
protocol. have different meanings in future versions of the protocol.
3.2. Connection Confirmation 3.2. Connection Confirmation
This document focuses on QUIC version 1, and this section applies
only to packets belonging to QUIC version 1 flows; for purposes of
on-path observation, it assumes that these packets have been
identified as such through the observation of a version number
exchange as described above.
Connection establishment uses Initial and Handshake packets Connection establishment uses Initial and Handshake packets
containing a TLS handshake, and Retry packets that do not contain containing a TLS handshake, and Retry packets that do not contain
parts of the handshake. Connection establishment can therefore be parts of the handshake. Connection establishment can therefore be
detected using heuristics similar to those used to detect TLS over detected using heuristics similar to those used to detect TLS over
TCP. A client initiating a 0-RTT connection may also send data TCP. A client initiating a connection may also send data in 0-RTT
packets in 0-RTT Protected packets directly after the Initial packet packets directly after the Initial packet containing the TLS Client
containing the TLS Client Hello. Since these packets may be Hello. Since these packets may be reordered in the network, 0-RTT
reordered in the network, 0-RTT Protected data packets could be seen packets could be seen before the Initial packet.
before the Initial packet.
Note that clients send Initial packets before servers do, servers Note that in this version of QUIC, clients send Initial packets
send Handshake packets before clients do, and only clients send before servers do, servers send Handshake packets before clients do,
Initial packets with tokens. Therefore, the role as a client or and only clients send Initial packets with tokens. Therefore, an
server can generally be confirmed by an on- path observer. An endpoint can be identified as a client or server by an on-path
attempted connection after Retry can be detected by correlating the observer. An attempted connection after Retry can be detected by
token on the Retry with the token on the subsequent Initial packet correlating the contents of the Retry packet with the Token and the
and the destination connection ID of the new Initial packet. Destination Connection ID fields of the new Initial packet.
3.3. Distinguishing Acknowledgment traffic 3.3. Distinguishing Acknowledgment Traffic
Some deployed in-network functions distinguish pure-acknowledgment Some deployed in-network functions distinguish pure-acknowledgment
(ACK) packets from packets carrying upper-layer data in order to (ACK) packets from packets carrying upper-layer data in order to
attempt to enhance performance, for example by queueing ACKs attempt to enhance performance, for example by queueing ACKs
differently or manipulating ACK signaling. Distinguishing ACK differently or manipulating ACK signaling. Distinguishing ACK
packets is trivial in TCP, but not supported by QUIC, since packets is trivial in TCP, but not supported by QUIC, since
acknowledgment signaling is carried inside QUIC's encrypted payload, acknowledgment signaling is carried inside QUIC's encrypted payload,
and ACK manipulation is impossible. Specifically, heuristics and ACK manipulation is impossible. Specifically, heuristics
attempting to distinguish ACK-only packets from payload-carrying attempting to distinguish ACK-only packets from payload-carrying
packets based on packet size are likely to fail, and are not packets based on packet size are likely to fail, and are not
recommended to use as a way to construe internals of QUIC's operation recommended to use as a way to construe internals of QUIC's operation
as those mechanisms can change, e.g., due to the use of extensions. as those mechanisms can change, e.g., due to the use of extensions.
3.4. Application Identification 3.4. Server Name Indication (SNI)
The cleartext TLS handshake may contain Server Name Indication (SNI) The client's TLS ClientHello may contain a Server Name Indication
[RFC6066], by which the client reveals the name of the server it (SNI) [RFC6066] extension, by which the client reveals the name of
intends to connect to, in order to allow the server to present a the server it intends to connect to, in order to allow the server to
certificate based on that name. It may also contain information from present a certificate based on that name. It may also contain an
Application-Layer Protocol Negotiation (ALPN) [RFC7301], by which the Application-Layer Protocol Negotiation (ALPN) [RFC7301] extension, by
client exposes the names of application-layer protocols it supports; which the client exposes the names of application-layer protocols it
an observer can deduce that one of those protocols will be used if supports; an observer can deduce that one of those protocols will be
the connection continues. used if the connection continues.
Work is currently underway in the TLS working group to encrypt the Work is currently underway in the TLS working group to encrypt the
SNI in TLS 1.3 [TLS-ESNI]. This would make SNI-based application contents of the ClientHello in TLS 1.3 [TLS-ECH]. This would make
identification impossible by on-path observation for QUIC and other SNI-based application identification impossible by on-path
protocols that use TLS. observation for QUIC and other protocols that use TLS.
3.4.1. Extracting Server Name Indication (SNI) Information 3.4.1. Extracting Server Name Indication (SNI) Information
If the SNI is not encrypted it can be derived from the QUIC Initial If the ClientHello is not encrypted, it can be derived from the
packet by calculating the Initial Secret to decrypt the packet client's Initial packet by calculating the Initial secret to decrypt
payload and parse the QUIC CRYPTO Frame containing the TLS the packet payload and parsing the QUIC CRYPTO Frame containing the
ClientHello. TLS ClientHello.
As both the initial salt for the Initial Secret as well as CRYPTO As both the derivation of the Initial secret and the structure of the
frame itself are version-specific, the first step is always to parse Initial packet itself are version-specific, the first step is always
the version number (second to sixth byte of the long header). Note to parse the version number (second to sixth bytes of the long
that only long header packets carry the version number, so it is header). Note that only long header packets carry the version
necessary to also check the if first bit of the QUIC packet is set to number, so it is necessary to also check if the first bit of the QUIC
1, indicating a long header. packet is set to 1, indicating a long header.
Note that proprietary QUIC versions, that have been deployed before Note that proprietary QUIC versions, that have been deployed before
standardization, might not set the first bit in a QUIC long header standardization, might not set the first bit in a QUIC long header
packets to 1. To parse these versions, example code is provided in packet to 1. To parse these versions, example code is provided in
the appendix (see Appendix A.1), however, it is expected that these the appendix (see Appendix A). However, it is expected that these
versions will gradually disappear over time. versions will gradually disappear over time.
When the version has been identified as QUIC version 1, the packet When the version has been identified as QUIC version 1, the packet
type needs to be verified as an Initial packet by checking that the type needs to be verified as an Initial packet by checking that the
third and fourth bit of the header are both set to 0. Then the third and fourth bits of the header are both set to 0. Then the
client destination connection ID needs to be extracted to calculate Destination Connection ID needs to be extracted to calculate the
the Initial Secret together with the version specific initial salt, Initial secret using the version-specific Initial salt, as described
as described in [QUIC-TLS]. The length of the connection ID is in Section 5.2 of [QUIC-TLS]. The length of the connection ID is
indicated in the 6th byte of the header followed by the connection ID indicated in the 6th byte of the header followed by the connection ID
itself. itself.
To determine the end of the header and find the start of the payload, To determine the end of the header and find the start of the payload,
the packet number length, the source connection ID length, and the the packet number length, the source connection ID length, and the
token length need to be extracted. The packet number length is token length need to be extracted. The packet number length is
defined by the seventh and eight bits of the header as described in defined by the seventh and eight bits of the header as described in
section 17.2. of [QUIC-TRANSPORT], but is obfuscated as described in Section 17.2 of [QUIC-TRANSPORT], but is obfuscated as described in
[QUIC-TLS]. The source connection ID length is specified in the byte Section 5.4 of [QUIC-TLS]. The source connection ID length is
after the destination connection ID. And the token length, which specified in the byte after the destination connection ID. The token
follows the source connection ID, is a variable length integer as length, which follows the source connection ID, is a variable-length
specified in Section 16 of [QUIC-TRANSPORT]. integer as specified in Section 16 of [QUIC-TRANSPORT].
After decryption, the Initial Client packet can be parsed to detect After decryption, the client's Initial packet can be parsed to detect
the CRYPTO frame that contains the TLS Client Hello, which then can the CRYPTO frame that contains the TLS ClientHello, which then can be
be parsed similarly to TLS over TCP connections. The Initial client parsed similarly to TLS over TCP connections. The client's Initial
packet may contain other frames, so the first bytes of each frame packet may contain other frames, so the first bytes of each frame
need to be checked to identify the frame type, and if needed skip need to be checked to identify the frame type, and if needed skip
over it. Note that the length of the frames is dependent on the over it. Note that the length of the frames is dependent on the
frame type. In QUIC version 1, the packet is expected to only carry frame type. In QUIC version 1, the packet is expected to contain
the CRYPTO frame and optionally padding frames. However, PADDING only CRYPTO frames and optionally PADDING frames. PADDING frames,
frames, each consisting of a single zero byte, may also occur before each consisting of a single zero byte, may occur before, after, or
or after the CRYPTO frame. between CRYPTO frames. There might be multiple CRYPTO frames.
Finally, an extension might define additional frame types which could
be present.
Note that client Initial packets after the first do not always use Note that subsequent Initial packets might contain a Destination
the destination connection ID that was used to generate the Initial Connection ID other than the one used to generate the Initial secret.
keys. Therefore, attempts to decrypt these packets using the Therefore, attempts to decrypt these packets using the procedure
procedure above might fail. above might fail unless the Initial secret is retained by the
observer.
3.5. Flow Association 3.5. Flow Association
The QUIC connection ID (see Section 2.6) is designed to allow an on- The QUIC connection ID (see Section 2.6) is designed to allow a
path device such as a load-balancer to associate two flows as coordinating on-path device, such as a load-balancer, to associate
identified by five-tuple when the address and port of one of the two flows when one of the endpoints changes address or port. This
endpoints changes; e.g. due to NAT rebinding or server IP address change can be due to NAT rebinding or address migration.
migration. An observer keeping flow state can associate a connection
ID with a given flow, and can associate a known flow with a new flow
when when observing a packet sharing a connection ID and one endpoint
address (IP address and port) with the known flow.
However, since the connection ID may change multiple times during the The connection ID must change upon intentional address change by an
lifetime of a flow, and the negotiation of connection ID changes is endpoint, and connection ID negotiation is encrypted, so it is not
encrypted, packets with the same 5-tuple but different connection IDs possible for a passive observer to link intended changes of address
may or may not belong to the same connection. using the connection ID.
The connection ID value should be treated as opaque; see Section 4.3 When one endpoint unintentionally changes its address, as is the case
for caveats regarding connection ID selection at servers. with NAT rebinding, an on-path observer may be able to use the
connection ID to associate the flow on the new address with the flow
on the old address.
3.6. Flow teardown A network function that attempts to use the connection ID to
associate flows must be robust to the failure of this technique.
Since the connection ID may change multiple times during the lifetime
of a connection, packets with the same five-tuple but different
connection IDs might or might not belong to the same connection.
Likewise, packets with the same connection ID but different five-
tuples might not belong to the same connection, either.
Connection IDs should be treated as opaque; see Section 4.4 for
caveats regarding connection ID selection at servers.
3.6. Flow Teardown
QUIC does not expose the end of a connection; the only indication to QUIC does not expose the end of a connection; the only indication to
on-path devices that a flow has ended is that packets are no longer on-path devices that a flow has ended is that packets are no longer
observed. Stateful devices on path such as NATs and firewalls must observed. Stateful devices on path such as NATs and firewalls must
therefore use idle timeouts to determine when to drop state for QUIC therefore use idle timeouts to determine when to drop state for QUIC
flows, see further section Section 4.1. flows; see Section 4.2.
3.7. Flow Symmetry Measurement 3.7. Flow Symmetry Measurement
QUIC explicitly exposes which side of a connection is a client and QUIC explicitly exposes which side of a connection is a client and
which side is a server during the handshake. In addition, the which side is a server during the handshake. In addition, the
symmetry of a flow (whether primarily client-to-server, primarily symmetry of a flow (whether primarily client-to-server, primarily
server-to-client, or roughly bidirectional, as input to basic traffic server-to-client, or roughly bidirectional, as input to basic traffic
classification techniques) can be inferred through the measurement of classification techniques) can be inferred through the measurement of
data rate in each direction. While QUIC traffic is protected and data rate in each direction. While QUIC traffic is protected and
ACKs may be padded, padding is not required. ACKs may be padded, padding is not required.
3.8. Round-Trip Time (RTT) Measurement 3.8. Round-Trip Time (RTT) Measurement
Round-trip time of QUIC flows can be inferred by observation once per The round-trip time of QUIC flows can be inferred by observation once
flow, during the handshake, as in passive TCP measurement; this per flow, during the handshake, as in passive TCP measurement; this
requires parsing of the QUIC packet header and recognition of the requires parsing of the QUIC packet header and recognition of the
handshake, as illustrated in Section 2.4. It can also be inferred handshake, as illustrated in Section 2.4. It can also be inferred
during the flow's lifetime, if the endpoints use the spin bit during the flow's lifetime, if the endpoints use the spin bit
facility described below and in [QUIC-TRANSPORT], section 17.3.1. facility described below and in Section 17.3.1 of [QUIC-TRANSPORT].
3.8.1. Measuring Initial RTT 3.8.1. Measuring Initial RTT
In the common case, the delay between the Initial packet containing In the common case, the delay between the client's Initial packet
the TLS Client Hello and the Handshake packet containing the TLS (containing the TLS ClientHello) and the server's Initial packet
Server Hello represents the RTT component on the path between the (containing the TLS ServerHello) represents the RTT component on the
observer and the server. The delay between the TLS Server Hello and path between the observer and the server. The delay between the
the Handshake packet containing the TLS Finished message sent by the server's first Handshake packet and the Handshake packet sent by the
client represents the RTT component on the path between the observer client represents the RTT component on the path between the observer
and the client. While the client may send 0-RTT Protected packets and the client. While the client may send 0-RTT packets after the
after the Initial packet during 0-RTT connection re-establishment, Initial packet during connection re-establishment, these can be
these can be ignored for RTT measurement purposes. ignored for RTT measurement purposes.
Handshake RTT can be measured by adding the client-to-observer and Handshake RTT can be measured by adding the client-to-observer and
observer-to-server RTT components together. This measurement observer-to-server RTT components together. This measurement
necessarily includes any transport and application layer delay (the necessarily includes any transport- and application-layer delay (the
latter mainly caused by the asymmetric crypto operations associated latter mainly caused by the asymmetric crypto operations associated
with the TLS handshake) at both sides. with the TLS handshake) at both sides.
3.8.2. Using the Spin Bit for Passive RTT Measurement 3.8.2. Using the Spin Bit for Passive RTT Measurement
The spin bit provides a version-specific method to measure per-flow The spin bit provides a version-specific method to measure per-flow
RTT from observation points on the network path throughout the RTT from observation points on the network path throughout the
duration of a connection. See section 17.4 of [QUIC-TRANSPORT] for duration of a connection. See Section 17.4 of [QUIC-TRANSPORT] for
the definition of the spin bit in Version 1 of QUIC. Endpoint the definition of the spin bit in Version 1 of QUIC. Endpoint
participation in spin bit signaling is optional. That is, while its participation in spin bit signaling is optional. That is, while its
location is fixed in this version of QUIC, an endpoint can location is fixed in this version of QUIC, an endpoint can
unilaterally choose to not support "spinning" the bit. unilaterally choose to not support "spinning" the bit.
Use of the spin bit for RTT measurement by devices on path is only Use of the spin bit for RTT measurement by devices on path is only
possible when both endpoints enable it. Some endpoints may disable possible when both endpoints enable it. Some endpoints may disable
use of the spin bit by default, others only in specific deployment use of the spin bit by default, others only in specific deployment
scenarios, e.g. for servers and clients where the RTT would reveal scenarios, e.g. for servers and clients where the RTT would reveal
the presence of a VPN or proxy. To avoid making these connections the presence of a VPN or proxy. To avoid making these connections
skipping to change at page 20, line 20 skipping to change at page 20, line 20
can also be used to generate RTT distribution information, including can also be used to generate RTT distribution information, including
minimum RTT (which approximates network RTT over longer time windows) minimum RTT (which approximates network RTT over longer time windows)
and RTT variance (which approximates jitter as seen by the and RTT variance (which approximates jitter as seen by the
application). application).
4. Specific Network Management Tasks 4. Specific Network Management Tasks
In this section, we review specific network management and In this section, we review specific network management and
measurement techniques and how QUIC's design impacts them. measurement techniques and how QUIC's design impacts them.
4.1. Stateful Treatment of QUIC Traffic 4.1. Passive Network Performance Measurement and Troubleshooting
Limited RTT measurement is possible by passive observation of QUIC
traffic; see Section 3.8. No passive measurement of loss is possible
with the present wire image. Extremely limited observation of
upstream congestion may be possible via the observation of CE
markings on ECN-enabled QUIC traffic.
4.2. Stateful Treatment of QUIC Traffic
Stateful treatment of QUIC traffic (e.g., at a firewall or NAT Stateful treatment of QUIC traffic (e.g., at a firewall or NAT
middlebox) is possible through QUIC traffic and version middlebox) is possible through QUIC traffic and version
identification (Section 3.1) and observation of the handshake for identification (Section 3.1) and observation of the handshake for
connection confirmation (Section 3.2). The lack of any visible end- connection confirmation (Section 3.2). The lack of any visible end-
of-flow signal (Section 3.6) means that this state must be purged of-flow signal (Section 3.6) means that this state must be purged
either through timers or through least-recently-used eviction, either through timers or through least-recently-used eviction,
depending on application requirements. depending on application requirements.
While QUIC has no clear network-visible end-of-connection signal and
therefore does require timer-based state removal, the QUIC handshake
indicates confirmation by both ends of a valid bidirectional
transmission. As soon as the handshake completed, timers should be
set long enough to also allow for short idle time during a valid
transmission.
[RFC4787] requires a timeout that is not less than 2 minutes for most [RFC4787] requires a timeout that is not less than 2 minutes for most
UDP traffic. However, in pratice, timers are often lower, in the UDP traffic. However, in practice, timers are sometimes lower, in
range of 15 to 30 seconds. In contrast, [RFC5382] recommends a the range of 30 to 60 seconds. In contrast, [RFC5382] recommends a
timeout of more than 2 hours for TCP, given that TCP is a connection- timeout of more than 2 hours for TCP, given that TCP is a connection-
oriented protocol with well-defined closure semantics. For network oriented protocol with well- defined closure semantics.
devices that are QUIC-aware, it is recommended to also use longer
timeouts for QUIC traffic, as QUIC is connection-oriented. As such,
a handshake packet from the server indicates the willingness of the
server to communicate with the client.
The QUIC header optionally contains a connection ID which can be used Even though QUIC has explicitly been designed tolerate NAT
as additional entropy beyond the 5-tuple, if needed. The QUIC rebindings, decreasing the NAT timeout is not recommended, as it may
handshake needs to be observed in order to understand whether the negatively impact application performance or incentivize endpoints to
connection ID is present and what length it has. However, connection send very frequent keep-alive packets. Instead it is recommended,
IDs may be renegotiated during after the handshake, and this even when lower timers are used for other UDP traffic, to use a timer
renegotiation is not visible to the path. Using the connection ID as of at least two minutes for QUIC traffic.
a flow key field for stateful treatment of flows may therefore cause
undetectable and unrecoverable loss of state in the middle of a
connection. Use of connection IDs is specifically discouraged for
NAT applications.
4.2. Passive Network Performance Measurement and Troubleshooting If state is removed too early, this could lead to black-holing of
incoming packets after a short idle period. To detect this
situation, a timer at the client needs to expire before a re-
establishment can happen (if at all), which would lead to unnecessary
long delays in an otherwise working connection.
Limited RTT measurement is possible by passive observation of QUIC Furthermore, not all endpoints use routing architectures where
traffic; see Section 3.8. No passive measurement of loss is possible connections will survive a port or address change. So even when the
with the present wire image. Extremely limited observation of client revives the connection, a NAT rebinding can cause a routing
upstream congestion may be possible via the observation of CE mismatch where a packet is not even delivered to the server that
markings on ECN-enabled QUIC traffic. might support address migration. For these reasons, the limits in
[RFC4787] are important to avoid black-holing of packets (and hence
avoid interrupting the flow of data to the client), especially where
devices are able to distinguish QUIC traffic from other UDP payloads.
4.3. Server Cooperation with Load Balancers The QUIC header optionally contains a connection ID which could
provide additional entropy beyond the 5-tuple. The QUIC handshake
needs to be observed in order to understand whether the connection ID
is present and what length it has. However, connection IDs may be
renegotiated after the handshake, and this renegotiation is not
visible to the path. Therefore using the connection ID as a flow key
field for stateful treatment of flows is not recommended as
connection ID changes will cause undetectable and unrecoverable loss
of state in the middle of a connection. Specially, the use of the
connection ID for functions that require state to make a forwarding
decison is not viable as it will break connectivity or at minimum
cause long timeout-based delays before this problem is detected by
the endpoints and the connection can potentially be re-established.
In the case of content distribution networking architectures Use of connection IDs is specifically discouraged for NAT
including load balancers, the connection ID provides a way for the applications. If a NAT hits an operational limit, it is recommended
server to signal information about the desired treatment of a flow to to rather drop the initial packets of a flow (see also Section 4.5),
the load balancers. Guidance on assigning connection IDs is given in which potentially triggers a fallback to TCP. Use of the connection
[QUIC-APPLICABILITY]. ID to multiplex multiple connections on the same IP address/port pair
is not a viable solution as it risks connectivity breakage, in case
the connection ID changes.
4.4. DDoS Detection and Mitigation 4.3. Address Rewriting to Ensure Routing Stability
Current practices in detection and mitigation of Distributed Denial While QUIC's migration capability makes it possible for an server to
of Service (DDoS) attacks generally involve classification of survive address changes, this does not work if the routers or
incoming traffic (as packets, flows, or some other aggregate) into switches in the server infrastructure route using the address-port
"good" (productive) and "bad" (DDoS) traffic, and then differential 4-tuple. If infrastructure routes on addresses only, NAT rebinding
treatment of this traffic to forward only good traffic. This or address migration will cause packets to be delivered to the wrong
operation is often done in a separate specialized mitigation server. [QUIC_LB] describes a way to addresses this problem by
environment through which all traffic is filtered; a generalized coordinating the selection and use of connection IDs between load-
architecture for separation of concerns in mitigation is given in balancers and servers.
[DOTS-ARCH].
Key to successful DDoS mitigation is efficient classification of this Applying address translation at a middlebox to maintain a stable
traffic in the mitigation environment. Limited first-packet garbage address-port mapping for flows based on connection ID might seem like
detection as in Section 3.1.2 and stateful tracking of QUIC traffic a solution to this problem. However, hiding information about the
as in Section 4.1 above may be useful during classification. change of the IP address or port conceals important and security-
relevant information from QUIC endpoints and as such would facilitate
amplification attacks (see Section 9 of [QUIC-TRANSPORT]). A NAT
function that hides peer address changes prevents the other end from
detecting and mitigating attacks as the endpoint cannot verify
connectivity to the new address using QUIC PATH_CHALLENGE and
PATH_RESPONSE frames.
Note that the use of a connection ID to support connection migration In addition, a change of IP address or port is also an input signal
renders 5-tuple based filtering insufficient and requires more state to other internal mechanisms in QUIC. When a path change is
to be maintained by DDoS defense systems. For the common case of NAT detected, path-dependent variables like congestion control parameters
rebinding, DDoS defense systems can detect a change in the client's will be reset protecting the new path from overload.
endpoint address by linking flows based on the server's connection
IDs. QUIC's linkability resistance ensures that a deliberate
connection migration is accompanied by a change in the connection ID.
It is questionable whether connection migrations must be supported Therefore, the use of address rewriting to ensure routing stability
during a DDoS attack. If the connection migration is not visible to can open QUIC up to various attacks, as it conceals client address
the network that performs the DDoS detection, an active, migrated changes, and as such masks important signals that drive security
QUIC connection may be blocked by such a system under attack. As mechanisms.
soon as the connection blocking is detected by the client, the client
may rely on the fast resumption mechanism provided by QUIC. When
clients migrate to a new path, they should be prepared for the
migration to fail and attempt to reconnect quickly.
TCP syncookies [RFC4937] are a well-established method of mitigating 4.4. Server Cooperation with Load Balancers
some kinds of TCP DDoS attacks. QUIC Retry packets are the
functional analogue to syncookies, forcing clients to prove
possession of their IP address before committing server state.
However, there are safeguards in QUIC against unsolicited injection
of these packets by intermediaries who do not have consent of the end
server. See [QUIC_LB] for standard ways for intermediaries to send
Retry packets on behalf of consenting servers.
4.5. UDP Policing In the case of networking architectures that include load balancers,
the connection ID can be used as a way for the server to signal
information about the desired treatment of a flow to the load
balancers. Guidance on assigning connection IDs is given in
[QUIC-APPLICABILITY]. [QUIC_LB] describes a system for coordinating
selection and use of connection IDs between load-balancers and
servers.
4.5. Filtering Behavior
[RFC4787] describes possible packet filtering behaviors that relate
to NATs but is often also used is other scenarios where packet
filtering is desired. Though the guidance there holds, a
particularly unwise behavior is to admit a handful of UDP packets and
then make a decision as to whether or not to filter it. QUIC
applications are encouraged to fail over to TCP if early packets do
not arrive at their destination [I-D.ietf-quic-applicability], as
QUIC is based on UDP and there are known blocks of UDP traffic (see
Section 4.6). Admitting a few packets allows the QUIC endpoint to
determine that the path accepts QUIC. Sudden drops afterwards will
result in slow and costly timeouts before abandoning the connection.
4.6. UDP Blocking or Throttling
Today, UDP is the most prevalent DDoS vector, since it is easy for Today, UDP is the most prevalent DDoS vector, since it is easy for
compromised non-admin applications to send a flood of large UDP compromised non-admin applications to send a flood of large UDP
packets (while with TCP the attacker gets throttled by the congestion packets (while with TCP the attacker gets throttled by the congestion
controller) or to craft reflection and amplification attacks. controller) or to craft reflection and amplification attacks. Some
Networks should therefore be prepared for UDP flood attacks on ports networks therefore block UDP traffic. With increased deployment of
used for QUIC traffic. One possible response to this threat is to QUIC, there is also an increased need to allow UDP traffic on ports
police UDP traffic on the network, allocating a fixed portion of the used for QUIC. However, if UDP is generally enabled on these ports,
network capacity to UDP and blocking UDP datagram over that cap. UDP flood attacks may also use the same ports. One possible response
to this threat is to throttle UDP traffic on the network, allocating
a fixed portion of the network capacity to UDP and blocking UDP
datagrams over that cap. As the portion of QUIC traffic compared to
TCP is also expected to increase over time, using such a limit is not
recommended but if done, limits might need to be adapted dynamically.
The recommended way to police QUIC packets is to either drop them all Further, if UDP traffic is desired to be throttled, it is recommended
or to throttle them based on the hash of the UDP datagram's source to block individual QUIC flows entirely rather than dropping packets
and destination addresses, blocking a portion of the hash space that randomly. When the handshake is blocked, QUIC-capable applications
corresponds to the fraction of UDP traffic one wishes to drop. When may failover to TCP However, blocking a random fraction of QUIC
the handshake is blocked, QUIC-capable applications may failover to packets across 4-tuples will allow many QUIC handshakes to complete,
TCP (at least applications using well-known UDP ports). However, preventing a TCP failover, but the connections will suffer from
blindly blocking a significant fraction of QUIC packets will allow severe packet loss (see also Section 4.5). Therefore UDP throttling
many QUIC handshakes to complete, preventing a TCP failover, but the should be realized by per-flow policing as opposed to per-packet
connections will suffer from severe packet loss. policing. Note that this per-flow policing should be stateless to
avoid problems with stateful treatment of QUIC flows (see
Section 4.2), for example blocking a portion of the space of values
of a hash function over the addresses and ports in the UDP datagram.
While QUIC endpoints are often able to survive address changes, e.g.
by NAT rebindings, blocking a portion of the traffic based on 5-tuple
hashing increases the risk of black-holing an active connection when
the address changes.
4.6. Handling ICMP Messages 4.7. DDoS Detection and Mitigation
Datagram Packetization Layer PMTU Discovery (PLPMTUD) can be used by On-path observation of the transport headers of packets can be used
QUIC to probe for the supported PMTU. PLPMTUD optionally uses ICMP for various security functions. For example, Denial of Service (DOS)
messages (e.g., IPv6 Packet Too Big messages). Given known attacks and Distributed DOS (DDOS) attacks against the infrastructure or
with the use of ICMP messages, the use of PLPMTUD in QUIC has been against an endpoint can be detected and mitigated by characterising
designed to safely use but not rely on receiving ICMP feedback (see anomalous traffic. Other uses include support for security audits
Section 14.2.1. of [QUIC-TRANSPORT]). (e.g., verifying the compliance with ciphersuites); client and
application fingerprinting for inventory; and to provide alerts for
network intrusion detection and other next generation firewall
functions.
Networks are recommended to forward these ICMP messages and retain as Current practices in detection and mitigation of DDoS attacks
much of the original packet as possible without exceeding the minimum generally involve classification of incoming traffic (as packets,
MTU for the IP version when generating ICMP messages as recommended flows, or some other aggregate) into "good" (productive) and "bad"
in [RFC1812] and [RFC4443]. (DDoS) traffic, and then differential treatment of this traffic to
forward only good traffic. This operation is often done in a
separate specialized mitigation environment through which all traffic
is filtered; a generalized architecture for separation of concerns in
mitigation is given in [DOTS-ARCH].
4.7. Quality of Service handling and ECMP Efficient classification of this DDoS traffic in the mitigation
environment is key to the success of this approach. Limited first-
packet garbage detection as in Section 3.1.2 and stateful tracking of
QUIC traffic as in Section 4.2 above may be useful during
classification.
Note that the use of a connection ID to support connection migration
renders 5-tuple based filtering insufficient to detect active flows
and requires more state to be maintained by DDoS defense systems if
support of migration of QUIC flows is desired. For the common case
of NAT rebinding, where the client's address changes without the
client's intent or knowledge, DDoS defense systems can detect a
change in the client's endpoint address by linking flows based on the
server's connection IDs. However, QUIC's linkability resistance
ensures that a deliberate connection migration is accompanied by a
change in the connection ID. In this case, the connection ID can not
be used to distinguish valid, active traffic from new attack traffic.
It is also possible for endpoints to directly support security
functions such as DoS classification and mitigation. Endpoints can
cooperate with an in-network device directly by e.g. sharing
information about connection IDs.
Another potential method could use an on-path network device that
relies on pattern inferences in the traffic and heuristics or machine
learning instead of processing observed header information.
However, it is questionable whether connection migrations must be
supported during a DDoS attack. While unintended migration without a
connection ID change can be more easily supported, it might be
acceptable to not support migrations of active QUIC connections that
are not visible to the network functions performing the DDoS
detection. As soon as the connection blocking is detected by the
client, the client may be able to rely on the fast resumption
mechanism provided by QUIC. When clients migrate to a new path, they
should be prepared for the migration to fail and attempt to reconnect
quickly.
Beyond in-network DDoS protection mechanisms, TCP syncookies
[RFC4937] are a well-established method of mitigating some kinds of
TCP DDoS attacks. QUIC Retry packets are the functional analogue to
syncookies, forcing clients to prove possession of their IP address
before committing server state. However, there are safeguards in
QUIC against unsolicited injection of these packets by intermediaries
who do not have consent of the end server. See [QUIC_LB] for
standard ways for intermediaries to send Retry packets on behalf of
consenting servers.
4.8. Quality of Service handling and ECMP
It is expected that any QoS handling in the network, e.g. based on It is expected that any QoS handling in the network, e.g. based on
use of DiffServ Code Points (DSCPs) [RFC2475] as well as Equal-Cost use of DiffServ Code Points (DSCPs) [RFC2475] as well as Equal-Cost
Multi-Path (ECMP) routing, is applied on a per flow-basis (and not Multi-Path (ECMP) routing, is applied on a per flow-basis (and not
per-packet) and as such that all packets belonging to the same QUIC per-packet) and as such that all packets belonging to the same QUIC
connection get uniform treatment. Using ECMP to distribute packets connection get uniform treatment. Using ECMP to distribute packets
from a single flow across multiple network paths or any other non- from a single flow across multiple network paths or any other non-
uniform treatment of packets belong to the same connection could uniform treatment of packets belong to the same connection could
result in variations in order, delivery rate, and drop rate. As result in variations in order, delivery rate, and drop rate. As
feedback about loss or delay of each packet is used as input to the feedback about loss or delay of each packet is used as input to the
congestion controller, these variations could adversely affect congestion controller, these variations could adversely affect
performance. performance.
Depending of the loss recovery mechanism implemented, QUIC may be Depending of the loss recovery mechanism implemented, QUIC may be
more tolerant of packet re-ordering than traditional TCP traffic (see more tolerant of packet re-ordering than traditional TCP traffic (see
Section 2.7). However, it cannot be known by the network which exact Section 2.7). However, it cannot be known by the network which exact
recovery mechanism is used and therefore reordering tolerance should recovery mechanism is used and therefore reordering tolerance should
be considered as unknown. be considered as unknown.
4.8. QUIC and Network Address Translation (NAT) 4.9. Handling ICMP Messages
QUIC Connection IDs are opaque byte fields that are expressed
consistently across all QUIC versions [QUIC-INVARIANTS], see
Section 2.6. This feature may appear to present opportunities to
optimize NAT port usage and simplify the work of the QUIC server. In
fact, NAT behavior that relies on CID may instead cause connection
failure when endpoints change Connection ID, and disable important
protocol security features. NATs should retain their existing 4-
tuple-based operation and refrain from parsing or otherwise using
QUIC connection IDs.
This section uses the colloquial term NAT to mean NAPT (section 2.2
of [RFC3022]), which overloads several IP addresses to one IP address
or to an IP address pool, as commonly deployed in carrier-grade NATs
or residential NATs.
The remainder of this section explains how QUIC supports NATs better
than other connection-oriented protocols, why NAT use of Connection
ID might appear attractive, and how NAT use of CID can create serious
problems for the endpoints.
[RFC4787] contains some guidance on building NATs to interact
constructively with a wide range of applications. This section
extends the discussion to QUIC.
By using the CID, QUIC connections can survive NAT rebindings as long
as no routing function in the path is dependent on client IP address
and port to deliver packets between server and NAT. Reducing the
timeout on UDP NATs might be tempting in light of this property, but
not all QUIC server deployments will be robust to rebinding.
4.8.1. Resource Conservation
NATs sometimes hit an operational limit where they exhaust available
public IP addresses and ports, and must evict flows from their
address/port mapping. CIDs might appear to offer a way to multiplex
many connections over a single address and port.
However, QUIC endpoints may negotiate new connection IDs inside
cryptographically protected packets, and begin using them at will.
Imagine two clients behind a NAT that are sharing the same public IP
address and port. The NAT is differentiating them using the incoming
Connection ID. If one client secretly changes its connection ID,
there will be no mapping for the NAT, and the connection will
suddenly break.
QUIC is deliberately designed to fail rather than persist when the
network cannot support its operation. For HTTP/3, this extends to
recommending a fallback to TCP-based versions of HTTP rather than
persisting with a QUIC connection that might be unstable. And
[QUIC-APPLICABILITY] recommends TCP fallback for other protocols on
the basis that this is preferable to sudden connection errors and
time outs. Furthermore, wide deployment of NATs with this behavior
hinders the use of QUIC's migration function, which relies on the
ability to change the connection ID any time during the lifetime of a
QUIC connection.
It is possible, in principle, to encode the client's identity in a
connection ID using the techniques described in [QUIC_LB] and
explicit coordination with the NAT. However, this implies that the
client shares configuration with the NAT, which might be logistically
difficult. This adds administrative overhead while not resolving the
case where a client migrates to a point behind the NAT.
Note that multiplexing connection IDs over a single port anyway Datagram Packetization Layer PMTU Discovery (PLPMTUD) can be used by
violates the best common practice to avoid "port overloading" as QUIC to probe for the supported PMTU. PLPMTUD optionally uses ICMP
described in [RFC4787]. messages (e.g., IPv6 Packet Too Big messages). Given known attacks
with the use of ICMP messages, the use of PLPMTUD in QUIC has been
designed to safely use but not rely on receiving ICMP feedback (see
Section 14.2.1. of [QUIC-TRANSPORT]).
4.8.2. "Helping" with routing infrastructure issues Networks are recommended to forward these ICMP messages and retain as
much of the original packet as possible without exceeding the minimum
MTU for the IP version when generating ICMP messages as recommended
in [RFC1812] and [RFC4443].
Concealing client address changes in order to simplify operational 4.10. Guiding Path MTU
routing issues will mask important signals that drive security
mechanisms, and therefore opens QUIC up to various attacks.
One challenge in QUIC deployments that want to benefit from QUIC's Some networks support 1500-byte packets, but can only do so by
migration capability is server infrastructures with routers and fragmenting at a lower layer before traversing a smaller MTU segment,
switches that direct traffic based on address-port 4-tuple rather and then reassembling. This is permissible even when the IP layer is
than connection ID. The use of source IP address means that a NAT IPv6 or IPv4 with the DF bit set, because it occurs below the IP
rebinding or address migration will deliver packets to the wrong layer. However, this process can add to compute and memory costs,
server. As all QUIC payloads are encrypted, routers and switches leading to a bottleneck that limits network capacity. In such
will not have access to negotiated but not-yet-in-use CIDs. This is networks this generates a desire to influence a majority of senders
a particular problem for low-state load balancers. [QUIC_LB] to use smaller packets, so that the limited reassembly capacity is
addresses this problem proposing a QUIC extension to allow some not exceeded.
server-load balancer coordination for routable CIDs.
It seems that a NAT anywhere in the front of such an infrastructure For TCP, MSS clamping (Section 3.2 of [RFC4459]) is often used to
setup could save the effort of converting all these devices by change the sender's maximum TCP segment size, but QUIC requires a
decoding routable connection IDs and rewriting the packet IP different approach. Section 14 of [QUIC-TRANSPORT] advises senders
addresses to allow consistent routing by legacy devices. to probe larger sizes using Datagram Packetization Layer PMTU
Discovery ([DPLPMTUD]) or Path Maximum Transmission Unit Discovery
(PMTUD: [RFC1191] and [RFC8201]). This mechanism will encourage
senders to approach the maximum size, which could cause fragmentation
with a network segment that they may not be aware of.
Unfortunately, the change of IP address or port is an important If path performance is limited when sending larger packets, an on-
signal to QUIC endpoints. It requires a review of path-dependent path device should support a maximum packet size for a specific
variables like congestion control parameters. It can also signify transport flow and then consistently drop all packets that exceed the
various attacks that mislead one endpoint about the best peer address configured size when the inner IPv4 packet has DF set, or IPv6 is
for the connection (see section 9 of [QUIC-TRANSPORT]). The QUIC used. Endpoints can cache PMTU information between IP flows, in the
PATH_CHALLENGE and PATH_RESPONSE frames are intended to detect and IP-layer cache, so short-term consistency between the PMTU for flows
mitigate these attacks and verify connectivity to the new address. can help avoid an endpoint using a PMTU that is inefficient.
This mechanism cannot work if the NAT is bleaching peer address
changes.
For example, an attacker might copy a legitimate QUIC packet and Networks with configurations that would lead to fragmentation of
change the source address to match its own. In the absence of a large packets should drop such packets rather than fragmenting them.
bleaching NAT, the receiving endpoint would interpret this as a Network operators who plan to implement a more selective policy may
potential NAT rebinding and use a PATH_CHALLENGE frame to prove that start by focussing on QUIC. QUIC flows cannot always be easily
the peer endpoint is not truly at the new address, thus thwarting the distinguished from other UDP traffic, but we assume at least some
attack. A bleaching NAT has no means of sending an encrypted portion of QUIC traffic can be identified (see Section 3.1). For
PATH_CHALLENGE frame, so it might start redirecting all QUIC traffic QUIC endpoints using DPLPMTUD it is recommended for the path to drop
to the attacker address and thus allow an observer to break the a packet larger than the supported size. A QUIC probe packet is used
connection. to discover the PMTU. If lost, this does not impact the flow of QUIC
data.
4.9. Filtering behavior IPv4 routers generate an ICMP message when a packet is dropped
because the link MTU was exceeded. [RFC8504] specifies how an IPv6
node generates an ICMPv6 Packet Too Big message (PTB) in this case.
PMTUD relies upon an endpoint receiving such PTB messages [RFC8201],
whereas DPLPMTUD does not reply upon these messages, but still can
optionally use these to improve performance Section 4.6 of
[DPLPMTUD].
[RFC4787] describes possible packet filtering behaviors that relate Since a network cannot know in advance which discovery method a QUIC
to NATs. Though the guidance there holds, a particularly unwise endpoint is using, it should always send a PTB message in addition to
behavior is to admit a handful of UDP packets and then make a dropping the oversized packet. A generated PTB message should be
decision as to whether or not to filter it. QUIC applications are compliant with the validation requirements of Section 14.2.1 of
encouraged to fail over to TCP if early packets do not arrive at [QUIC-TRANSPORT], otherwise it will be ignored by DPLPMTUD. This
their destination. Admitting a few packets allows the QUIC endpoint will likely provide the right signal for the endpoint to keep the
to determine that the path accepts QUIC. Sudden drops afterwards packet size small and thereby avoid network fragmentation for that
will result in slow and costly timeouts before abandoning the flow entirely.
connection.
5. IANA Considerations 5. IANA Considerations
This document has no actions for IANA. This document has no actions for IANA.
6. Security Considerations 6. Security Considerations
QUIC is an encrypted and authenticated transport. That means, once QUIC is an encrypted and authenticated transport. That means, once
the cryptographic handshake is complete, QUIC endpoints discard most the cryptographic handshake is complete, QUIC endpoints discard most
packets that are not authenticated, greatly limiting the ability of packets that are not authenticated, greatly limiting the ability of
skipping to change at page 27, line 4 skipping to change at page 28, line 20
may result in connection failure. may result in connection failure.
7. Contributors 7. Contributors
The following people have contributed text to sections of this The following people have contributed text to sections of this
document: document:
* Dan Druta * Dan Druta
* Martin Duke * Martin Duke
* Marcus Ilhar
* Igor Lubashev * Igor Lubashev
* David Schinazi * David Schinazi
* Gorry Fairhurst
* Chris Box
8. Acknowledgments 8. Acknowledgments
Special thanks to Martin Thomson and Martin Duke for the detailed Thanks to Thomas Fossati, Jana Iygengar, Marcus Ihlar for their early
reviews and feedback. reviews and feedback. Special thanks also to Martin Thomson and
Martin Duke for their detailed reviews and input. And thanks to Sean
Turner, Mike Bishop, Ian Swett, and Nick Banks for their last call
reviews.
This work is partially supported by the European Commission under This work is partially supported by the European Commission under
Horizon 2020 grant agreement no. 688421 Measurement and Architecture Horizon 2020 grant agreement no. 688421 Measurement and Architecture
for a Middleboxed Internet (MAMI), and by the Swiss State Secretariat for a Middleboxed Internet (MAMI), and by the Swiss State Secretariat
for Education, Research, and Innovation under contract no. 15.0268. for Education, Research, and Innovation under contract no. 15.0268.
This support does not imply endorsement. This support does not imply endorsement.
9. References 9. References
9.1. Normative References 9.1. Normative References
skipping to change at page 27, line 47 skipping to change at page 29, line 22
9.2. Informative References 9.2. Informative References
[DOTS-ARCH] [DOTS-ARCH]
Mortensen, A., Reddy, T., Andreasen, F., Teague, N., and Mortensen, A., Reddy, T., Andreasen, F., Teague, N., and
R. Compton, "DDoS Open Threat Signaling (DOTS) R. Compton, "DDoS Open Threat Signaling (DOTS)
Architecture", Work in Progress, Internet-Draft, draft- Architecture", Work in Progress, Internet-Draft, draft-
ietf-dots-architecture-18, 6 March 2020, ietf-dots-architecture-18, 6 March 2020,
<https://tools.ietf.org/html/draft-ietf-dots-architecture- <https://tools.ietf.org/html/draft-ietf-dots-architecture-
18>. 18>.
[DPLPMTUD] Fairhurst, G., Jones, T., Tüxen, M., Rüngeler, I., and T.
Völker, "Packetization Layer Path MTU Discovery for
Datagram Transports", RFC 8899, DOI 10.17487/RFC8899,
September 2020, <https://www.rfc-editor.org/rfc/rfc8899>.
[I-D.ietf-quic-applicability]
Kuehlewind, M. and B. Trammell, "Applicability of the QUIC
Transport Protocol", Work in Progress, Internet-Draft,
draft-ietf-quic-applicability-11, 21 April 2021,
<https://tools.ietf.org/html/draft-ietf-quic-
applicability-11>.
[IPIM] Allman, M., Beverly, R., and B. Trammell, "In-Protocol [IPIM] Allman, M., Beverly, R., and B. Trammell, "In-Protocol
Internet Measurement (arXiv preprint 1612.02902)", 9 Internet Measurement (arXiv preprint 1612.02902)", 9
December 2016, <https://arxiv.org/abs/1612.02902>. December 2016, <https://arxiv.org/abs/1612.02902>.
[QUIC-APPLICABILITY] [QUIC-APPLICABILITY]
Kuehlewind, M. and B. Trammell, "Applicability of the QUIC Kuehlewind, M. and B. Trammell, "Applicability of the QUIC
Transport Protocol", Work in Progress, Internet-Draft, Transport Protocol", Work in Progress, Internet-Draft,
draft-ietf-quic-applicability-09, 22 January 2021, draft-ietf-quic-applicability-11, 21 April 2021,
<https://tools.ietf.org/html/draft-ietf-quic- <https://tools.ietf.org/html/draft-ietf-quic-
applicability-09>. applicability-11>.
[QUIC-HTTP] [QUIC-HTTP]
Bishop, M., "Hypertext Transfer Protocol Version 3 Bishop, M., "Hypertext Transfer Protocol Version 3
(HTTP/3)", Work in Progress, Internet-Draft, draft-ietf- (HTTP/3)", Work in Progress, Internet-Draft, draft-ietf-
quic-http-34, 2 February 2021, quic-http-34, 2 February 2021,
<https://tools.ietf.org/html/draft-ietf-quic-http-34>. <https://tools.ietf.org/html/draft-ietf-quic-http-34>.
[QUIC-INVARIANTS] [QUIC-INVARIANTS]
Thomson, M., "Version-Independent Properties of QUIC", Thomson, M., "Version-Independent Properties of QUIC",
Work in Progress, Internet-Draft, draft-ietf-quic- Work in Progress, Internet-Draft, draft-ietf-quic-
skipping to change at page 28, line 37 skipping to change at page 30, line 24
Congestion Control", Work in Progress, Internet-Draft, Congestion Control", Work in Progress, Internet-Draft,
draft-ietf-quic-recovery-34, 14 January 2021, draft-ietf-quic-recovery-34, 14 January 2021,
<https://tools.ietf.org/html/draft-ietf-quic-recovery-34>. <https://tools.ietf.org/html/draft-ietf-quic-recovery-34>.
[QUIC_LB] Duke, M. and N. Banks, "QUIC-LB: Generating Routable QUIC [QUIC_LB] Duke, M. and N. Banks, "QUIC-LB: Generating Routable QUIC
Connection IDs", Work in Progress, Internet-Draft, draft- Connection IDs", Work in Progress, Internet-Draft, draft-
ietf-quic-load-balancers-06, 4 February 2021, ietf-quic-load-balancers-06, 4 February 2021,
<https://tools.ietf.org/html/draft-ietf-quic-load- <https://tools.ietf.org/html/draft-ietf-quic-load-
balancers-06>. balancers-06>.
[RFC1191] Mogul, J. and S. Deering, "Path MTU discovery", RFC 1191,
DOI 10.17487/RFC1191, November 1990,
<https://www.rfc-editor.org/rfc/rfc1191>.
[RFC1812] Baker, F., Ed., "Requirements for IP Version 4 Routers", [RFC1812] Baker, F., Ed., "Requirements for IP Version 4 Routers",
RFC 1812, DOI 10.17487/RFC1812, June 1995, RFC 1812, DOI 10.17487/RFC1812, June 1995,
<https://www.rfc-editor.org/rfc/rfc1812>. <https://www.rfc-editor.org/rfc/rfc1812>.
[RFC2475] Blake, S., Black, D., Carlson, M., Davies, E., Wang, Z., [RFC2475] Blake, S., Black, D., Carlson, M., Davies, E., Wang, Z.,
and W. Weiss, "An Architecture for Differentiated and W. Weiss, "An Architecture for Differentiated
Services", RFC 2475, DOI 10.17487/RFC2475, December 1998, Services", RFC 2475, DOI 10.17487/RFC2475, December 1998,
<https://www.rfc-editor.org/rfc/rfc2475>. <https://www.rfc-editor.org/rfc/rfc2475>.
[RFC3022] Srisuresh, P. and K. Egevang, "Traditional IP Network
Address Translator (Traditional NAT)", RFC 3022,
DOI 10.17487/RFC3022, January 2001,
<https://www.rfc-editor.org/rfc/rfc3022>.
[RFC4443] Conta, A., Deering, S., and M. Gupta, Ed., "Internet [RFC4443] Conta, A., Deering, S., and M. Gupta, Ed., "Internet
Control Message Protocol (ICMPv6) for the Internet Control Message Protocol (ICMPv6) for the Internet
Protocol Version 6 (IPv6) Specification", STD 89, Protocol Version 6 (IPv6) Specification", STD 89,
RFC 4443, DOI 10.17487/RFC4443, March 2006, RFC 4443, DOI 10.17487/RFC4443, March 2006,
<https://www.rfc-editor.org/rfc/rfc4443>. <https://www.rfc-editor.org/rfc/rfc4443>.
[RFC4459] Savola, P., "MTU and Fragmentation Issues with In-the-
Network Tunneling", RFC 4459, DOI 10.17487/RFC4459, April
2006, <https://www.rfc-editor.org/rfc/rfc4459>.
[RFC4787] Audet, F., Ed. and C. Jennings, "Network Address [RFC4787] Audet, F., Ed. and C. Jennings, "Network Address
Translation (NAT) Behavioral Requirements for Unicast Translation (NAT) Behavioral Requirements for Unicast
UDP", BCP 127, RFC 4787, DOI 10.17487/RFC4787, January UDP", BCP 127, RFC 4787, DOI 10.17487/RFC4787, January
2007, <https://www.rfc-editor.org/rfc/rfc4787>. 2007, <https://www.rfc-editor.org/rfc/rfc4787>.
[RFC4937] Arberg, P. and V. Mammoliti, "IANA Considerations for PPP [RFC4937] Arberg, P. and V. Mammoliti, "IANA Considerations for PPP
over Ethernet (PPPoE)", RFC 4937, DOI 10.17487/RFC4937, over Ethernet (PPPoE)", RFC 4937, DOI 10.17487/RFC4937,
June 2007, <https://www.rfc-editor.org/rfc/rfc4937>. June 2007, <https://www.rfc-editor.org/rfc/rfc4937>.
[RFC5382] Guha, S., Ed., Biswas, K., Ford, B., Sivakumar, S., and P. [RFC5382] Guha, S., Ed., Biswas, K., Ford, B., Sivakumar, S., and P.
skipping to change at page 29, line 39 skipping to change at page 31, line 28
[RFC7301] Friedl, S., Popov, A., Langley, A., and E. Stephan, [RFC7301] Friedl, S., Popov, A., Langley, A., and E. Stephan,
"Transport Layer Security (TLS) Application-Layer Protocol "Transport Layer Security (TLS) Application-Layer Protocol
Negotiation Extension", RFC 7301, DOI 10.17487/RFC7301, Negotiation Extension", RFC 7301, DOI 10.17487/RFC7301,
July 2014, <https://www.rfc-editor.org/rfc/rfc7301>. July 2014, <https://www.rfc-editor.org/rfc/rfc7301>.
[RFC7605] Touch, J., "Recommendations on Using Assigned Transport [RFC7605] Touch, J., "Recommendations on Using Assigned Transport
Port Numbers", BCP 165, RFC 7605, DOI 10.17487/RFC7605, Port Numbers", BCP 165, RFC 7605, DOI 10.17487/RFC7605,
August 2015, <https://www.rfc-editor.org/rfc/rfc7605>. August 2015, <https://www.rfc-editor.org/rfc/rfc7605>.
[TLS-ESNI] Rescorla, E., Oku, K., Sullivan, N., and C. A. Wood, "TLS [RFC7983] Petit-Huguenin, M. and G. Salgueiro, "Multiplexing Scheme
Updates for Secure Real-time Transport Protocol (SRTP)
Extension for Datagram Transport Layer Security (DTLS)",
RFC 7983, DOI 10.17487/RFC7983, September 2016,
<https://www.rfc-editor.org/rfc/rfc7983>.
[RFC8201] McCann, J., Deering, S., Mogul, J., and R. Hinden, Ed.,
"Path MTU Discovery for IP version 6", STD 87, RFC 8201,
DOI 10.17487/RFC8201, July 2017,
<https://www.rfc-editor.org/rfc/rfc8201>.
[RFC8504] Chown, T., Loughney, J., and T. Winters, "IPv6 Node
Requirements", BCP 220, RFC 8504, DOI 10.17487/RFC8504,
January 2019, <https://www.rfc-editor.org/rfc/rfc8504>.
[TLS-ECH] Rescorla, E., Oku, K., Sullivan, N., and C. A. Wood, "TLS
Encrypted Client Hello", Work in Progress, Internet-Draft, Encrypted Client Hello", Work in Progress, Internet-Draft,
draft-ietf-tls-esni-09, 16 December 2020, draft-ietf-tls-esni-10, 8 March 2021,
<https://tools.ietf.org/html/draft-ietf-tls-esni-09>. <https://tools.ietf.org/html/draft-ietf-tls-esni-10>.
[TMA-QOF] Trammell, B., Gugelmann, D., and N. Brownlee, "Inline Data [TMA-QOF] Trammell, B., Gugelmann, D., and N. Brownlee, "Inline Data
Integrity Signals for Passive Measurement (in Proc. TMA Integrity Signals for Passive Measurement (in Proc. TMA
2014)", April 2014. 2014)", April 2014.
[WIRE-IMAGE] [WIRE-IMAGE]
Trammell, B. and M. Kuehlewind, "The Wire Image of a Trammell, B. and M. Kuehlewind, "The Wire Image of a
Network Protocol", RFC 8546, DOI 10.17487/RFC8546, April Network Protocol", RFC 8546, DOI 10.17487/RFC8546, April
2019, <https://www.rfc-editor.org/rfc/rfc8546>. 2019, <https://www.rfc-editor.org/rfc/rfc8546>.
Appendix A. Appendix Appendix A. Distinguishing IETF QUIC and Google QUIC Versions
This appendix uses the following conventions: array[i] - one byte at
index i of array array[i:j] - subset of array starting with index i
(inclusive) up to j-1 (inclusive) array[i:] - subset of array
starting with index i (inclusive) up to the end of the array
A.1. Distinguishing IETF QUIC and Google QUIC Versions
This section contains algorithms that allows parsing versions from This section contains algorithms that allows parsing versions from
both Google QUIC and IETF QUIC. These mechanisms will become both Google QUIC and IETF QUIC. These mechanisms will become
irrelevant when IETF QUIC is fully deployed and Google QUIC is irrelevant when IETF QUIC is fully deployed and Google QUIC is
deprecated. deprecated.
Note that other than this appendix, nothing in this document applies Note that other than this appendix, nothing in this document applies
to Google QUIC. And the purpose of this appendix is merely to to Google QUIC. And the purpose of this appendix is merely to
distinguish IETF QUIC from any versions of Google QUIC. distinguish IETF QUIC from any versions of Google QUIC.
This appendix uses the following conventions: * array[i] - one byte
at index i of array * array[i:j] - subset of array starting with
index i (inclusive) up to j-1 (inclusive) * array[i:] - subset of
array starting with index i (inclusive) up to the end of the array
Conceptually, a Google QUIC version is an opaque 32bit field. When Conceptually, a Google QUIC version is an opaque 32bit field. When
we refer to a version with four printable characters, we use its we refer to a version with four printable characters, we use its
ASCII representation: for example, Q050 refers to {'Q', '0', '5', ASCII representation: for example, Q050 refers to {'Q', '0', '5',
'0'} which is equal to {0x51, 0x30, 0x35, 0x30}. Otherwise, we use '0'} which is equal to {0x51, 0x30, 0x35, 0x30}. Otherwise, we use
its hexadecimal representation: for example, 0xff00001d refers to its hexadecimal representation: for example, 0xff00001d refers to
{0xff, 0x00, 0x00, 0x1d}. {0xff, 0x00, 0x00, 0x1d}.
QUIC versions that start with 'Q' or 'T' followed by three digits are QUIC versions that start with 'Q' or 'T' followed by three digits are
Google QUIC versions. Versions up to and including 43 are documented Google QUIC versions. Versions up to and including 43 are documented
by <https://docs.google.com/document/d/ by <https://docs.google.com/document/d/
skipping to change at page 31, line 29 skipping to change at page 33, line 29
} }
if (first_byte_bit5) { if (first_byte_bit5) {
version = packet[9:13] version = packet[9:13]
} else { } else {
version = packet[5:9] version = packet[5:9]
} }
} else { } else {
abort("Packet without version") abort("Packet without version")
} }
A.2. Extracting the CRYPTO frame A.1. Extracting the CRYPTO frame
counter = 0 counter = 0
while (payload[counter] == 0) { while (payload[counter] == 0) {
counter += 1 counter += 1
} }
first_nonzero_payload_byte = payload[counter] first_nonzero_payload_byte = payload[counter]
fnz_payload_byte_bit3 = ((first_nonzero_payload_byte & 0x20) != 0) fnz_payload_byte_bit3 = ((first_nonzero_payload_byte & 0x20) != 0)
if (first_nonzero_payload_byte != 0x06) { if (first_nonzero_payload_byte != 0x06) {
abort("Unexpected frame") abort("Unexpected frame")
} }
skipping to change at page 32, line 36 skipping to change at page 34, line 36
ParseTLS(crypto_data) ParseTLS(crypto_data)
Authors' Addresses Authors' Addresses
Mirja Kuehlewind Mirja Kuehlewind
Ericsson Ericsson
Email: mirja.kuehlewind@ericsson.com Email: mirja.kuehlewind@ericsson.com
Brian Trammell Brian Trammell
Google Google Switzerland GmbH
Gustav-Gull-Platz 1 Gustav-Gull-Platz 1
CH- 8004 Zurich CH- 8004 Zurich
Switzerland Switzerland
Email: ietf@trammell.ch Email: ietf@trammell.ch
 End of changes. 138 change blocks. 
552 lines changed or deleted 652 lines changed or added

This html diff was produced by rfcdiff 1.48. The latest version is available from http://tools.ietf.org/tools/rfcdiff/