QUIC: A UDP-Based Multiplexed and Secure Transport

[x]

Indicates that x is optional

x (A)

Indicates that x is A bits long

x (A/B/C) …

Indicates that x is one of A, B, or C bits long

x (i) …

Indicates that x uses the variable-length encoding in Section 7.1

x (*) …

Indicates that x is variable-length

Header Form:

The most significant bit (0x80) of octet 0 (the first octet) is set to 1 for long headers.

Long Packet Type:

The remaining seven bits of octet 0 contain the packet type. This field can indicate one of 128 packet types. The types specified for this version are listed in Table 1.

Version:

The QUIC Version is a 32-bit field that follows the Type. This field indicates which version of QUIC is in use and determines how the rest of the protocol fields are interpreted.

DCIL and SCIL:

The octet following the version contains the lengths of the two connection ID fields that follow it. These lengths are encoded as two 4-bit unsigned integers. The Destination Connection ID Length (DCIL) field occupies the 4 high bits of the octet and the Source Connection ID Length (SCIL) field occupies the 4 low bits of the octet. An encoded length of 0 indicates that the connection ID is also 0 octets in length. Non-zero encoded lengths are increased by 3 to get the full length of the connection ID, producing a length between 4 and 18 octets inclusive. For example, an octet with the value 0x50 describes an 8-octet Destination Connection ID and a zero-length Source Connection ID.

Destination Connection ID:

The Destination Connection ID field follows the connection ID lengths and is either 0 octets in length or between 4 and 18 octets. Section 4.10 describes the use of this field in more detail.

Source Connection ID:

The Source Connection ID field follows the Destination Connection ID and is either 0 octets in length or between 4 and 18 octets. Section 4.10 describes the use of this field in more detail.

Length:

The length of the remainder of the packet (that is, the Packet Number and Payload fields) in octets, encoded as a variable-length integer (Section 7.1).

Packet Number:

The packet number field is 1, 2, or 4 octets long. The packet number has confidentiality protection separate from packet protection, as described in Section 5.3 of [QUIC-TLS]. The length of the packet number field is encoded in the plaintext packet number. See Section 4.11 for details.

Payload:

The payload of the packet.

The header form, type, connection ID lengths octet, destination and source connection IDs, and version fields of a long header packet are version-independent. The packet number and values for packet types defined in Table 1 are version-specific. See [QUIC-INVARIANTS] for details on how packets from different versions of QUIC are interpreted.¶

The interpretation of the fields and the payload are specific to a version and packet type. Type-specific semantics for this version are described in the following sections.¶

The end of the packet is determined by the Length field. The Length field covers the both the Packet Number and Payload fields, both of which are confidentiality protected and initially of unknown length. The size of the Payload field is learned once the packet number protection is removed.¶

Senders can sometimes coalesce multiple packets into one UDP datagram. See Section 4.9 for more details.¶

Header Form:

The most significant bit (0x80) of octet 0 is set to 0 for the short header.

Key Phase Bit:

The second bit (0x40) of octet 0 indicates the key phase, which allows a recipient of a packet to identify the packet protection keys that are used to protect the packet. See [QUIC-TLS] for details.

[[Editor’s Note: this section should be removed and the bit definitions changed before this draft goes to the IESG.]]¶

Third Bit:

The third bit (0x20) of octet 0 is set to 1.

[[Editor’s Note: this section should be removed and the bit definitions changed before this draft goes to the IESG.]]¶

Fourth Bit:

The fourth bit (0x10) of octet 0 is set to 1.

[[Editor’s Note: this section should be removed and the bit definitions changed before this draft goes to the IESG.]]¶

Google QUIC Demultipexing Bit:

The fifth bit (0x8) of octet 0 is set to 0. This allows implementations of Google QUIC to distinguish Google QUIC packets from short header packets sent by a client because Google QUIC servers expect the connection ID to always be present. The special interpretation of this bit SHOULD be removed from this specification when Google QUIC has finished transitioning to the new header format.

Reserved:

The sixth, seventh, and eighth bits (0x7) of octet 0 are reserved for experimentation. Endpoints MUST ignore these bits on packets they receive unless they are participating in an experiment that uses these bits. An endpoint not actively using these bits SHOULD set the value randomly on packets they send to protect against unwanted inference about particular values.

Destination Connection ID:

The Destination Connection ID is a connection ID that is chosen by the intended recipient of the packet. See Section 6.1 for more details.

Packet Number:

The packet number field is 1, 2, or 4 octets long. The packet number has confidentiality protection separate from packet protection, as described in Section 5.3 of [QUIC-TLS]. The length of the packet number field is encoded in the plaintext packet number. See Section 4.11 for details.

Protected Payload:

Packets with a short header always include a 1-RTT protected payload.

The header form and connection ID field of a short header packet are version-independent. The remaining fields are specific to the selected QUIC version. See [QUIC-INVARIANTS] for details on how packets from different versions of QUIC are interpreted.¶

A Version Negotiation packet is inherently not version-specific, and does not use the long packet header (see Section 4.1. Upon receipt by a client, it will appear to be a packet using the long header, but will be identified as a Version Negotiation packet based on the Version field having a value of 0.¶

The Version Negotiation packet is a response to a client packet that contains a version that is not supported by the server, and is only sent by servers.¶

The value in the Unused field is selected randomly by the server.¶

The Version field of a Version Negotiation packet MUST be set to 0x00000000.¶

The server MUST include the value from the Source Connection ID field of the packet it receives in the Destination Connection ID field. The value for Source Connection ID MUST be copied from the Destination Connection ID of the received packet, which is initially randomly selected by a client. Echoing both connection IDs gives clients some assurance that the server received the packet and that the Version Negotiation packet was not generated by an off-path attacker.¶

The remainder of the Version Negotiation packet is a list of 32-bit versions which the server supports.¶

A Version Negotiation packet cannot be explicitly acknowledged in an ACK frame by a client. Receiving another Initial packet implicitly acknowledges a Version Negotiation packet.¶

The Version Negotiation packet does not include the Packet Number and Length fields present in other packets that use the long header form. Consequently, a Version Negotiation packet consumes an entire UDP datagram.¶

See Section 6.3 for a description of the version negotiation process.¶

A Retry packet uses a long packet header with a type value of 0x7E. It carries an address validation token created by the server. It is used by a server that wishes to perform a stateless retry (see Section 6.7).¶

ODCIL:

The length of the Original Destination Connection ID field. The length is encoded in the least significant 4 bits of the octet, using the same encoding as the DCIL and SCIL fields. The most significant 4 bits of this octet are reserved. Unless a use for these bits has been negotiated, endpoints SHOULD send randomized values and MUST ignore any value that it receives.

Original Destination Connection ID:

The Original Destination Connection ID contains the value of the Destination Connection ID from the Initial packet that this Retry is in response to. The length of this field is given in ODCIL.

Retry Token:

An opaque token that the server can use to validate the client’s address.

The server populates the Destination Connection ID with the connection ID that the client included in the Source Connection ID of the Initial packet.¶

The server includes a connection ID of its choice in the Source Connection ID field. The client MUST use this connection ID in the Destination Connection ID of subsequent packets that it sends.¶

A Retry packet does not include a packet number and cannot be explicitly acknowledged by a client.¶

A server MUST NOT send a Retry in response to packets other than Initial or 0-RTT packets. A server MAY choose to only send Retry in response to Initial packets and discard or buffer 0-RTT packets corresponding to unvalidated client addresses.¶

If the Original Destination Connection ID field does not match the Destination Connection ID from the most recent Initial packet it sent, clients MUST discard the packet. This prevents an off-path attacker from injecting a Retry packet.¶

The client responds to a Retry packet with an Initial packet that includes the provided Retry Token to continue connection establishment.¶

A client MAY attempt 0-RTT after receiving a Retry packet by sending 0-RTT packets to the connection ID provided by the server. A client that sends additional 0-RTT packets MUST NOT reset the packet number to 0 after a Retry packet, see Section 4.6.4.¶

A server that might send another Retry packet in response to a subsequent Initial packet MUST set the Source Connection ID to a new value of at least 8 octets in length. This allows clients to distinguish between Retry packets when the server sends multiple rounds of Retry packets. Consequently, a valid Retry packet will always have an Original Destination Connection ID that is at least 8 octets long; clients MUST discard Retry packets that include a shorter value. A server that will not send additional Retry packets can set the Source Connection ID to any value.¶

Once version negotiation is complete, the cryptographic handshake is used to agree on cryptographic keys. The cryptographic handshake is carried in Initial (Section 4.6) and Handshake (Section 4.7) packets.¶

All these packets use the long header and contain the current QUIC version in the version field.¶

In order to prevent tampering by version-unaware middleboxes, Initial packets are protected with connection- and version-specific keys (Initial keys) as described in [QUIC-TLS]. This protection does not provide confidentiality or integrity against on-path attackers, but provides some level of protection against off-path attackers.¶

The Initial packet uses long headers with a type value of 0x7F. It carries the first CRYPTO frames sent by the client and server to perform key exchange, and carries ACKs in either direction. The Initial packet is protected by Initial keys as described in [QUIC-TLS].¶

The Initial packet (shown in Figure 5) has two additional header fields that are added to the Long Header before the Length field.¶

Token Length:

A variable-length integer specifying the length of the Token field, in bytes. This value is zero if no token is present. Initial packets sent by the server MUST set the Token Length field to zero; clients that receive an Initial packet with a non-zero Token Length field MUST either discard the packet or generate a connection error of type PROTOCOL_VIOLATION.

Token:

The value of the token.

The client and server use the Initial packet type for any packet that contains an initial cryptographic handshake message. This includes all cases where a new packet containing the initial cryptographic message needs to be created, such as the packets sent after receiving a Version Negotiation (Section 4.3) or Retry packet (Section 4.4).¶

A server sends its first Initial packet in response to a client Initial. A server may send multiple Initial packets. The cryptographic key exchange could require multiple round trips or retransmissions of this data.¶

The payload of an Initial packet includes a CRYPTO frame (or frames) containing a cryptographic handshake message, ACK frames, or both. PADDING frames are also permitted. The first CRYPTO frame sent always begins at an offset of 0 (see Section 6.4).¶

The first packet sent by a client always includes a CRYPTO frame that contains the entirety of the first cryptographic handshake message. This packet, and the cryptographic handshake message, MUST fit in a single UDP datagram (see Section 6.4).¶

Note that if the server sends a HelloRetryRequest, the client will send a second Initial packet. This Initial packet will continue the cryptographic handshake and will contain a CRYPTO frame with an offset matching the size of the CRYPTO frame sent in the first Initial packet. Cryptographic handshake messages subsequent to the first do not need to fit within a single UDP datagram.¶

A Handshake packet uses long headers with a type value of 0x7D. It is used to carry acknowledgments and cryptographic handshake messages from the server and client.¶

A server sends its cryptographic handshake in one or more Handshake packets in response to an Initial packet if it does not send a Retry packet. Once a client has received a Handshake packet from a server, it uses Handshake packets to send subsequent cryptographic handshake messages and acknowledgments to the server.¶

The Destination Connection ID field in a Handshake packet contains a connection ID that is chosen by the recipient of the packet; the Source Connection ID includes the connection ID that the sender of the packet wishes to use (see Section 4.10).¶

The first Handshake packet sent by a server contains a packet number of 0. Handshake packets are their own packet number space. Packet numbers are incremented normally for other Handshake packets.¶

Servers MUST NOT send more than three datagrams including Initial and Handshake packets without receiving a packet from a verified source address. Source addresses can be verified through an address validation token (delivered via a Retry packet or a NEW_TOKEN frame) or by receiving any message from the client encrypted using the Handshake keys.¶

The payload of this packet contains CRYPTO frames and could contain PADDING, or ACK frames. Handshake packets MAY contain CONNECTION_CLOSE frames if the handshake is unsuccessful.¶

All QUIC packets use packet protection. Packets that are protected with the static handshake keys or the 0-RTT keys are sent with long headers; all packets protected with 1-RTT keys are sent with short headers. The different packet types explicitly indicate the encryption level and therefore the keys that are used to remove packet protection. 0-RTT and 1-RTT protected packets share a single packet number space.¶

Packets protected with handshake keys only use packet protection to ensure that the sender of the packet is on the network path. This packet protection is not effective confidentiality protection; any entity that receives the Initial packet from a client can recover the keys necessary to remove packet protection or to generate packets that will be successfully authenticated.¶

Packets protected with 0-RTT and 1-RTT keys are expected to have confidentiality and data origin authentication; the cryptographic handshake ensures that only the communicating endpoints receive the corresponding keys.¶

Packets protected with 0-RTT keys use a type value of 0x7C. The connection ID fields for a 0-RTT packet MUST match the values used in the Initial packet (Section 4.6).¶

The client can send 0-RTT packets after receiving an Initial Section 4.6 or Handshake (Section 4.7) packet, if that packet does not complete the handshake. Even if the client receives a different connection ID in the Handshake packet, it MUST continue to use the same Destination Connection ID for 0-RTT packets, see Section 4.10.¶

The version field for protected packets is the current QUIC version.¶

The packet number field contains a packet number, which has additional confidentiality protection that is applied after packet protection is applied (see [QUIC-TLS] for details). The underlying packet number increases with each packet sent, see Section 4.11 for details.¶

The payload is protected using authenticated encryption. [QUIC-TLS] describes packet protection in detail. After decryption, the plaintext consists of a sequence of frames, as described in Section 5.¶

A sender can coalesce multiple QUIC packets (typically a Cryptographic Handshake packet and a Protected packet) into one UDP datagram. This can reduce the number of UDP datagrams needed to send application data during the handshake and immediately afterwards. It is not necessary for senders to coalesce packets, though failing to do so will require sending a significantly larger number of datagrams during the handshake. Receivers MUST be able to process coalesced packets.¶

Senders SHOULD coalesce packets in order of increasing encryption levels (Initial, Handshake, 0-RTT, 1-RTT), as this makes it more likely the receiver will be able to process all the packets in a single pass. A packet with a short header does not include a length, so it will always be the last packet included in a UDP datagram.¶

Senders MUST NOT coalesce QUIC packets with different Destination Connection IDs into a single UDP datagram. Receivers SHOULD ignore any subsequent packets with a different Destination Connection ID than the first packet in the datagram.¶

Every QUIC packet that is coalesced into a single UDP datagram is separate and complete. Though the values of some fields in the packet header might be redundant, no fields are omitted. The receiver of coalesced QUIC packets MUST individually process each QUIC packet and separately acknowledge them, as if they were received as the payload of different UDP datagrams. If one or more packets in a datagram cannot be processed yet (because the keys are not yet available) or processing fails (decryption failure, unknown type, etc.), the receiver MUST still attempt to process the remaining packets. The skipped packets MAY either be discarded or buffered for later processing, just as if the packets were received out-of-order in separate datagrams.¶

Retry (Section 4.4) and Version Negotiation (Section 4.3) packets cannot be coalesced.¶

A connection ID is used to ensure consistent routing of packets, as described in Section 6.1. The long header contains two connection IDs: the Destination Connection ID is chosen by the recipient of the packet and is used to provide consistent routing; the Source Connection ID is used to set the Destination Connection ID used by the peer.¶

During the handshake, packets with the long header are used to establish the connection ID that each endpoint uses. Each endpoint uses the Source Connection ID field to specify the connection ID that is used in the Destination Connection ID field of packets being sent to them. Upon receiving a packet, each endpoint sets the Destination Connection ID it sends to match the value of the Source Connection ID that they receive.¶

During the handshake, an endpoint might receive multiple packets with the long header, and thus be given multiple opportunities to update the Destination Connection ID it sends. A client MUST only change the value it sends in the Destination Connection ID in response to the first packet of each type it receives from the server (Retry or Initial); a server MUST set its value based on the Initial packet. Any additional changes are not permitted; if subsequent packets of those types include a different Source Connection ID, they MUST be discarded. This avoids problems that might arise from stateless processing of multiple Initial packets producing different connection IDs.¶

Short headers only include the Destination Connection ID and omit the explicit length. The length of the Destination Connection ID field is expected to be known to endpoints.¶

Endpoints using a connection-ID based load balancer could agree with the load balancer on a fixed or minimum length and on an encoding for connection IDs. This fixed portion could encode an explicit length, which allows the entire connection ID to vary in length and still be used by the load balancer.¶

The very first packet sent by a client includes a random value for Destination Connection ID. The same value MUST be used for all 0-RTT packets sent on that connection (Section 4.8). This randomized value is used to determine the packet protection keys for Initial packets (see Section 5.1.1 of [QUIC-TLS]).¶

A Version Negotiation (Section 4.3) packet MUST use both connection IDs selected by the client, swapped to ensure correct routing toward the client.¶

The connection ID can change over the lifetime of a connection, especially in response to connection migration (Section 6.11). NEW_CONNECTION_ID frames (Section 7.13) are used to provide new connection ID values.¶

The packet number is an integer in the range 0 to 2^62-1. The value is used in determining the cryptographic nonce for packet encryption. Each endpoint maintains a separate packet number for sending and receiving.¶

• Initial space: All Initial packets Section 4.6 are in this space.
• Handshake space: All Handshake packets Section 4.7 are in this space.
• Application data space: All 0-RTT and 1-RTT encrypted packets Section 4.8 are in this space.

As described in [QUIC-TLS], each packet type uses different encryption keys.¶

Conceptually, a packet number space is the encryption context in which a packet can be processed and ACKed. Initial packets can only be sent with Initial encryption keys and ACKed in packets which are also Initial packets. Similarly, Handshake packets can only be sent and acknowledged in Handshake packets.¶

This enforces cryptographic separation between the data sent in the different packet sequence number spaces. Each packet number space starts at packet number 0. Subsequent packets sent in the same packet number space MUST increase the packet number by at least one.¶

0-RTT and 1-RTT data exist in the same packet number space to make loss recovery algorithms easier to implement between the two packet types.¶

A QUIC endpoint MUST NOT reuse a packet number within the same packet number space in one connection (that is, under the same cryptographic keys). If the packet number for sending reaches 2^62 - 1, the sender MUST close the connection without sending a CONNECTION_CLOSE frame or any further packets; an endpoint MAY send a Stateless Reset (Section 6.13.4) in response to further packets that it receives.¶

In the QUIC long and short packet headers, the number of bits required to represent the packet number is reduced by including only a variable number of the least significant bits of the packet number. One or two of the most significant bits of the first octet determine how many bits of the packet number are provided, as shown in Table 2.¶

Note that these encodings are similar to those in Section 7.1, but use different values.¶

The encoded packet number is protected as described in Section 5.3 [QUIC-TLS]. Protection of the packet number is removed prior to recovering the full packet number. The full packet number is reconstructed at the receiver based on the number of significant bits present, the content of those bits, and the largest packet number received on a successfully authenticated packet. Recovering the full packet number is necessary to successfully remove packet protection.¶

Once packet number protection is removed, the packet number is decoded by finding the packet number value that is closest to the next expected packet. The next expected packet is the highest received packet number plus one. For example, if the highest successfully authenticated packet had a packet number of 0xaa82f30e, then a packet containing a 14-bit value of 0x9b3 will be decoded as 0xaa8309b3. Example pseudo-code for packet number decoding can be found in Appendix A.¶

The sender MUST use a packet number size able to represent more than twice as large a range than the difference between the largest acknowledged packet and packet number being sent. A peer receiving the packet will then correctly decode the packet number, unless the packet is delayed in transit such that it arrives after many higher-numbered packets have been received. An endpoint SHOULD use a large enough packet number encoding to allow the packet number to be recovered even if the packet arrives after packets that are sent afterwards.¶

As a result, the size of the packet number encoding is at least one more than the base 2 logarithm of the number of contiguous unacknowledged packet numbers, including the new packet.¶

For example, if an endpoint has received an acknowledgment for packet 0x6afa2f, sending a packet with a number of 0x6b2d79 requires a packet number encoding with 14 bits or more; whereas the 30-bit packet number encoding is needed to send a packet with a number of 0x6bc107.¶

A receiver MUST discard a newly unprotected packet unless it is certain that it has not processed another packet with the same packet number from the same packet number space. Duplicate suppression MUST happen after removing packet protection for the reasons described in Section 9.3 of [QUIC-TLS]. An efficient algorithm for duplicate suppression can be found in Section 3.4.3 of [RFC2406].¶

A Version Negotiation packet (Section 4.3) does not include a packet number. The Retry packet (Section 4.4) has special rules for populating the packet number field.¶

QUIC frames do not use a self-describing encoding. An endpoint therefore needs to understand the syntax of all frames before it can successfully process a packet. This allows for efficient encoding of frames, but it means that an endpoint cannot send a frame of a type that is unknown to its peer.¶

An extension to QUIC that wishes to use a new type of frame MUST first ensure that a peer is able to understand the frame. An endpoint can use a transport parameter to signal its willingness to receive one or more extension frame types with the one transport parameter.¶

An IANA registry is used to manage the assignment of frame types, see Section 13.2.¶

Each connection possesses a set of identifiers, any of which could be used to distinguish it from other connections. A connection ID can be either 0 octets in length, or between 4 and 18 octets (inclusive). Connection IDs are selected independently in each direction.¶

The primary function of a connection ID is to ensure that changes in addressing at lower protocol layers (UDP, IP, and below) don’t cause packets for a QUIC connection to be delivered to the wrong endpoint. Each endpoint selects connection IDs using an implementation-specific (and perhaps deployment-specific) method which will allow packets with that connection ID to be routed back to the endpoint and identified by the endpoint upon receipt.¶

A zero-length connection ID MAY be used when the connection ID is not needed for routing and the address/port tuple of packets is sufficient to associate them to a connection. An endpoint whose peer has selected a zero-length connection ID MUST continue to use a zero-length connection ID for the lifetime of the connection and MUST NOT send packets from any other local address.¶

When an endpoint has requested a non-zero-length connection ID, it will issue a series of connection IDs over the lifetime of a connection. The series of connection IDs issued by an endpoint is ordered, with the final connection ID selected during the handshake coming first. Additional connection IDs are provided using the NEW_CONNECTION_ID frame (Section 7.13), each with a specified sequence number. The series of connection IDs issued SHOULD be contiguous, but might not appear to be upon receipt due to reordering or loss.¶

Each connection ID MUST be used on only one local address. When packets are sent for the first time on a new local address, a new connection ID MUST be used with a higher sequence number than any connection ID previously used on any local address. At any time, an endpoint MAY change to a new connection ID on a local address already in use.¶

An endpoint MUST NOT send packets with a connection ID which has a lower sequence number than the highest sequence number of any connection ID ever sent or received on that local address. This ensures that when an endpoint migrates to a new path or changes connection ID on an existing path, the packets will use a new connection ID in both directions.¶

Implementations SHOULD ensure that peers have a connection ID with a matching sequence number available when changing to a new connection ID. An implementation could do this by always supplying a corresponding connection ID to a peer for each connection ID received from that peer.¶

While endpoints select connection IDs as appropriate for their implementation, the connection ID MUST NOT include the unprotected sequence number. Endpoints need to be able to recover the sequence number associated with each connection ID they generate without relying on information available to unaffiliated parties. A connection ID that encodes an unencrypted sequence number could be used to correlate connection IDs across network paths.¶

Incoming packets are classified on receipt. Packets can either be associated with an existing connection, or - for servers - potentially create a new connection.¶

Hosts try to associate a packet with an existing connection. If the packet has a Destination Connection ID corresponding to an existing connection, QUIC processes that packet accordingly. Note that more than one connection ID can be associated with a connection; see Section 6.1.¶

If the Destination Connection ID is zero length and the packet matches the address/port tuple of a connection where the host did not require connection IDs, QUIC processes the packet as part of that connection. Endpoints MUST drop packets with zero-length Destination Connection ID fields if they do not correspond to a single connection.¶

Version negotiation ensures that client and server agree to a QUIC version that is mutually supported. A server sends a Version Negotiation packet in response to each packet that might initiate a new connection, see Section 6.2 for details.¶

The size of the first packet sent by a client will determine whether a server sends a Version Negotiation packet. Clients that support multiple QUIC versions SHOULD pad their Initial packets to reflect the largest minimum Initial packet size of all their versions. This ensures that the server responds if there are any mutually supported versions.¶

QUIC relies on a combined cryptographic and transport handshake to minimize connection establishment latency. QUIC uses the CRYPTO frame Section 7.21 to transmit the cryptographic handshake. Version 0x00000001 of QUIC uses TLS 1.3 as described in [QUIC-TLS]; a different QUIC version number could indicate that a different cryptographic handshake protocol is in use.¶

• authenticated key exchange, where
  • a server is always authenticated,
  • a client is optionally authenticated,
  • every connection produces distinct and unrelated keys,
  • keying material is usable for packet protection for both 0-RTT and 1-RTT packets, and
  • 1-RTT keys have forward secrecy
• authenticated values for the transport parameters of the peer (see Section 6.6)
• authenticated confirmation of version negotiation (see Section 6.6.4)
• authenticated negotiation of an application protocol (TLS uses ALPN [RFC7301] for this purpose)
• for the server, the ability to carry data that provides assurance that the client can receive packets that are addressed with the transport address that is claimed by the client (see Section 6.9)

The initial CRYPTO frame MUST be sent in a single packet. Any second attempt that is triggered by address validation MUST also be sent within a single packet. This avoids having to reassemble a message from multiple packets.¶

The first client packet of the cryptographic handshake protocol MUST fit within a 1232 octet QUIC packet payload. This includes overheads that reduce the space available to the cryptographic handshake protocol.¶

The CRYPTO frame can be sent in different packet number spaces. CRYPTO frames in each packet number space carry a separate sequence of handshake data starting from an offset of 0.¶

Details of how TLS is integrated with QUIC are provided in [QUIC-TLS], but some examples are provided here.¶

Figure 8 provides an overview of the 1-RTT handshake. Each line shows a QUIC packet with the packet type and packet number shown first, followed by the contents. So, for instance the first packet is of type Initial, with packet number 0, and contains a CRYPTO frame carrying the ClientHello.¶

Note that multiple QUIC packets – even of different encryption levels – may be coalesced into a single UDP datagram (see Section 4.9), and so this handshake may consist of as few as 4 UDP datagrams, or any number more. For instance, the server’s first flight contains packets from the Initial encryption level (obfuscation), the Handshake level, and “0.5-RTT data” from the server at the 1-RTT encryption level.¶

Figure 9 shows an example of a connection with a 0-RTT handshake and a single packet of 0-RTT data. Note that as described in Section 4.11, the server ACKs the 0-RTT data at the 1-RTT encryption level, and the client’s sequence numbers at the 1-RTT encryption level continue to increment from its 0-RTT packets.¶

During connection establishment, both endpoints make authenticated declarations of their transport parameters. These declarations are made unilaterally by each endpoint. Endpoints are required to comply with the restrictions implied by these parameters; the description of each parameter includes rules for its handling.¶

The format of the transport parameters is the TransportParameters struct from Figure 10. This is described using the presentation language from Section 3 of [TLS13].¶

The extension_data field of the quic_transport_parameters extension defined in [QUIC-TLS] contains a TransportParameters value. TLS encoding rules are therefore used to encode the transport parameters.¶

QUIC encodes transport parameters into a sequence of octets, which are then included in the cryptographic handshake. Once the handshake completes, the transport parameters declared by the peer are available. Each endpoint validates the value provided by its peer. In particular, version negotiation MUST be validated (see Section 6.6.4) before the connection establishment is considered properly complete.¶

Definitions for each of the defined transport parameters are included in Section 6.6.1. Any given parameter MUST appear at most once in a given transport parameters extension. An endpoint MUST treat receipt of duplicate transport parameters as a connection error of type TRANSPORT_PARAMETER_ERROR.¶

A server can process an initial cryptographic handshake messages from a client without committing any state. This allows a server to perform address validation (Section 6.9), or to defer connection establishment costs.¶

A server that generates a response to an Initial packet without retaining connection state MUST use the Retry packet (Section 4.4). This packet causes a client to restart the connection attempt and includes the token in the new Initial packet (Section 4.6) to prove source address ownership.¶

QUIC endpoints use Explicit Congestion Notification (ECN) [RFC3168] to detect and respond to network congestion. ECN allows a network node to indicate congestion in the network by setting a codepoint in the IP header of a packet instead of dropping it. Endpoints react to congestion by reducing their sending rate in response, as described in [QUIC-RECOVERY].¶

To use ECN, QUIC endpoints first determine whether a path and peer support ECN marking. Verifying the path occurs at the beginning of a connection and when the connection migrates to a new path (see Section 6.11).¶

Each endpoint independently verifies and enables ECN for the path from it to the peer.¶

To verify that both a path and the peer support ECN, an endpoint MUST set one of the ECN Capable Transport (ECT) codepoints – ECT(0) or ECT(1) – in the IP header [RFC8311] of all outgoing packets.¶

If an ECT codepoint set in the IP header is not corrupted by a network device, then a received packet contains either the codepoint sent by the peer or the Congestion Experienced (CE) codepoint set by a network device that is experiencing congestion.¶

On receiving a packet with an ECT or CE codepoint, an endpoint that supports ECN increases the corresponding ECT(0), ECT(1), or CE count, and includes these counters in subsequent (see Section 8.1) ACK_ECN frames (see Section 7.16).¶

A packet detected by a receiver as a duplicate does not affect the receiver’s local ECN codepoint counts; see (Section 12.7) for relevant security concerns.¶

If an endpoint receives a packet without an ECT or CE codepoint, it responds per Section 8.1 with an ACK frame.¶

If an endpoint does not support ECN or does not have access to received ECN codepoints, it acknowledges received packets per Section 8.1 with an ACK frame.¶

If a packet sent with an ECT codepoint is newly acknowledged by the peer in an ACK frame, the endpoint stops setting ECT codepoints in subsequent packets, with the expectation that either the network or the peer no longer supports ECN.¶

• The increase in ECT(0) and ECT(1) counters MUST be at least the number of packets newly acknowledged that were sent with the corresponding codepoint.
• The total increase in ECT(0), ECT(1), and CE counters reported in the ACK_ECN frame MUST be at least the total number of packets newly acknowledged in this ACK_ECN frame.

An endpoint could miss acknowledgements for a packet when ACK frames are lost. It is therefore possible for the total increase in ECT(0), ECT(1), and CE counters to be greater than the number of packets acknowledged in an ACK frame. When this happens, the local reference counts MUST be increased to match the counters in the ACK frame.¶

Upon successful verification, an endpoint continues to set ECT codepoints in subsequent packets with the expectation that the path is ECN-capable.¶

If verification fails, then the endpoint ceases setting ECT codepoints in subsequent packets with the expectation that either the network or the peer does not support ECN.¶

If an endpoint sets ECT codepoints on outgoing packets and encounters a retransmission timeout due to the absence of acknowledgments from the peer (see [QUIC-RECOVERY]), or if an endpoint has reason to believe that a network element might be corrupting ECN codepoints, the endpoint MAY cease setting ECT codepoints in subsequent packets. Doing so allows the connection to traverse network elements that drop or corrupt ECN codepoints in the IP header.¶

Transport protocols commonly spend a round trip checking that a client owns the transport address (IP and port) that it claims. Verifying that a client can receive packets sent to its claimed transport address protects against spoofing of this information by malicious clients.¶

This technique is used primarily to avoid QUIC from being used for traffic amplification attack. In such an attack, a packet is sent to a server with spoofed source address information that identifies a victim. If a server generates more or larger packets in response to that packet, the attacker can use the server to send more data toward the victim than it would be able to send on its own.¶

Several methods are used in QUIC to mitigate this attack. Firstly, the initial handshake packet is sent in a UDP datagram that contains at least 1200 octets of UDP payload. This allows a server to send a similar amount of data without risking causing an amplification attack toward an unproven remote address.¶

A server eventually confirms that a client has received its messages when the first Handshake-level message is received. This might be insufficient, either because the server wishes to avoid the computational cost of completing the handshake, or it might be that the size of the packets that are sent during the handshake is too large. This is especially important for 0-RTT, where the server might wish to provide application data traffic - such as a response to a request - in response to the data carried in the early data from the client.¶

To send additional data prior to completing the cryptographic handshake, the server then needs to validate that the client owns the address that it claims.¶

Source address validation is therefore performed by the core transport protocol during the establishment of a connection.¶

A different type of source address validation is performed after a connection migration, see Section 6.10.¶

Path validation is used by an endpoint to verify reachability of a peer over a specific path. That is, it tests reachability between a specific local address and a specific peer address, where an address is the two-tuple of IP address and port. Path validation tests that packets can be both sent to and received from a peer.¶

Path validation is used during connection migration (see Section 6.11 and Section 6.12) by the migrating endpoint to verify reachability of a peer from a new local address. Path validation is also used by the peer to verify that the migrating endpoint is able to receive packets sent to the its new address. That is, that the packets received from the migrating endpoint do not carry a spoofed source address.¶

Path validation can be used at any time by either endpoint. For instance, an endpoint might check that a peer is still in possession of its address after a period of quiescence.¶

Path validation is not designed as a NAT traversal mechanism. Though the mechanism described here might be effective for the creation of NAT bindings that support NAT traversal, the expectation is that one or other peer is able to receive packets without first having sent a packet on that path. Effective NAT traversal needs additional synchronization mechanisms that are not provided here.¶

An endpoint MAY bundle PATH_CHALLENGE and PATH_RESPONSE frames that are used for path validation with other frames. For instance, an endpoint may pad a packet carrying a PATH_CHALLENGE for PMTU discovery, or an endpoint may bundle a PATH_RESPONSE with its own PATH_CHALLENGE.¶

QUIC allows connections to survive changes to endpoint addresses (that is, IP address and/or port), such as those caused by a endpoint migrating to a new network. This section describes the process by which an endpoint migrates to a new address.¶

An endpoint MUST NOT initiate connection migration before the handshake is finished and the endpoint has 1-RTT keys. The design of QUIC relies on endpoints retaining a stable address for the duration of the handshake.¶

An endpoint also MUST NOT initiate connection migration if the peer sent the disable_migration transport parameter during the handshake. An endpoint which has sent this transport parameter, but detects that a peer has nonetheless migrated to a different network MAY treat this as a connection error of type INVALID_MIGRATION.¶

Not all changes of peer address are intentional migrations. The peer could experience NAT rebinding: a change of address due to a middlebox, usually a NAT, allocating a new outgoing port or even a new outgoing IP address for a flow. Endpoints SHOULD perform path validation (Section 6.10) if a NAT rebinding does not cause the connection to fail.¶

This document limits migration of connections to new client addresses, except as described in Section 6.12. Clients are responsible for initiating all migrations. Servers do not send non-probing packets (see Section 6.11.1) toward a client address until it sees a non-probing packet from that address. If a client receives packets from an unknown server address, the client MAY discard these packets.¶

QUIC allows servers to accept connections on one IP address and attempt to transfer these connections to a more preferred address shortly after the handshake. This is particularly useful when clients initially connect to an address shared by multiple servers but would prefer to use a unicast address to ensure connection stability. This section describes the protocol for migrating a connection to a preferred server address.¶

Migrating a connection to a new server address mid-connection is left for future work. If a client receives packets from a new server address not indicated by the preferred_address transport parameter, the client SHOULD discard these packets.¶

• idle timeout (Section 6.13.2)
• immediate close (Section 6.13.3)
• stateless reset (Section 6.13.4)

QUIC frames commonly use a variable-length encoding for non-negative integer values. This encoding ensures that smaller integer values need fewer octets to encode.¶

The QUIC variable-length integer encoding reserves the two most significant bits of the first octet to encode the base 2 logarithm of the integer encoding length in octets. The integer value is encoded on the remaining bits, in network byte order.¶

This means that integers are encoded on 1, 2, 4, or 8 octets and can encode 6, 14, 30, or 62 bit values respectively. Table 4 summarizes the encoding properties.¶

For example, the eight octet sequence c2 19 7c 5e ff 14 e8 8c (in hexadecimal) decodes to the decimal value 151288809941952652; the four octet sequence 9d 7f 3e 7d decodes to 494878333; the two octet sequence 7b bd decodes to 15293; and the single octet 25 decodes to 37 (as does the two octet sequence 40 25).¶

Error codes (Section 11.3) are described using integers, but do not use this encoding.¶

The PADDING frame (type=0x00) has no semantic value. PADDING frames can be used to increase the size of a packet. Padding can be used to increase an initial client packet to the minimum required size, or to provide protection against traffic analysis for protected packets.¶

A PADDING frame has no content. That is, a PADDING frame consists of the single octet that identifies the frame as a PADDING frame.¶

An endpoint may use a RST_STREAM frame (type=0x01) to abruptly terminate a stream.¶

After sending a RST_STREAM, an endpoint ceases transmission and retransmission of STREAM frames on the identified stream. A receiver of RST_STREAM can discard any data that it already received on that stream.¶

An endpoint that receives a RST_STREAM frame for a send-only stream MUST terminate the connection with error PROTOCOL_VIOLATION.¶

Stream ID:

A variable-length integer encoding of the Stream ID of the stream being terminated.

Application Protocol Error Code:

A 16-bit application protocol error code (see Section 11.4) which indicates why the stream is being closed.

Final Offset:

A variable-length integer indicating the absolute byte offset of the end of data written on this stream by the RST_STREAM sender.

An endpoint sends a CONNECTION_CLOSE frame (type=0x02) to notify its peer that the connection is being closed. CONNECTION_CLOSE is used to signal errors at the QUIC layer, or the absence of errors (with the NO_ERROR code).¶

If there are open streams that haven’t been explicitly closed, they are implicitly closed when the connection is closed.¶

Error Code:

A 16-bit error code which indicates the reason for closing this connection. CONNECTION_CLOSE uses codes from the space defined in Section 11.3.

Frame Type:

A variable-length integer encoding the type of frame that triggered the error. A value of 0 (equivalent to the mention of the PADDING frame) is used when the frame type is unknown.

Reason Phrase Length:

A variable-length integer specifying the length of the reason phrase in bytes. Note that a CONNECTION_CLOSE frame cannot be split between packets, so in practice any limits on packet size will also limit the space available for a reason phrase.

Reason Phrase:

A human-readable explanation for why the connection was closed. This can be zero length if the sender chooses to not give details beyond the Error Code. This SHOULD be a UTF-8 encoded string [RFC3629].

An APPLICATION_CLOSE frame (type=0x03) is used to signal an error with the protocol that uses QUIC.¶

Error Code:

A 16-bit error code which indicates the reason for closing this connection. APPLICATION_CLOSE uses codes from the application protocol error code space, see Section 11.4.

Reason Phrase Length:

This field is identical in format and semantics to the Reason Phrase Length field from CONNECTION_CLOSE.

Reason Phrase:

This field is identical in format and semantics to the Reason Phrase field from CONNECTION_CLOSE.

APPLICATION_CLOSE has similar format and semantics to the CONNECTION_CLOSE frame (Section 7.4). Aside from the semantics of the Error Code field and the omission of the Frame Type field, both frames are used to close the connection.¶

The MAX_DATA frame (type=0x04) is used in flow control to inform the peer of the maximum amount of data that can be sent on the connection as a whole.¶

Maximum Data:

A variable-length integer indicating the maximum amount of data that can be sent on the entire connection, in units of octets.

All data sent in STREAM frames counts toward this limit. The sum of the largest received offsets on all streams - including streams in terminal states - MUST NOT exceed the value advertised by a receiver. An endpoint MUST terminate a connection with a QUIC_FLOW_CONTROL_RECEIVED_TOO_MUCH_DATA error if it receives more data than the maximum data value that it has sent, unless this is a result of a change in the initial limits (see Section 6.6.2).¶

The MAX_STREAM_DATA frame (type=0x05) is used in flow control to inform a peer of the maximum amount of data that can be sent on a stream.¶

An endpoint that receives a MAX_STREAM_DATA frame for a receive-only stream MUST terminate the connection with error PROTOCOL_VIOLATION.¶

An endpoint that receives a MAX_STREAM_DATA frame for a send-only stream it has not opened MUST terminate the connection with error PROTOCOL_VIOLATION.¶

Note that an endpoint may legally receive a MAX_STREAM_DATA frame on a bidirectional stream it has not opened.¶

Stream ID:

The stream ID of the stream that is affected encoded as a variable-length integer.

Maximum Stream Data:

A variable-length integer indicating the maximum amount of data that can be sent on the identified stream, in units of octets.

When counting data toward this limit, an endpoint accounts for the largest received offset of data that is sent or received on the stream. Loss or reordering can mean that the largest received offset on a stream can be greater than the total size of data received on that stream. Receiving STREAM frames might not increase the largest received offset.¶

The data sent on a stream MUST NOT exceed the largest maximum stream data value advertised by the receiver. An endpoint MUST terminate a connection with a FLOW_CONTROL_ERROR error if it receives more data than the largest maximum stream data that it has sent for the affected stream, unless this is a result of a change in the initial limits (see Section 6.6.2).¶

The MAX_STREAM_ID frame (type=0x06) informs the peer of the maximum stream ID that they are permitted to open.¶

Maximum Stream ID:

ID of the maximum unidirectional or bidirectional peer-initiated stream ID for the connection encoded as a variable-length integer. The limit applies to unidirectional steams if the second least signification bit of the stream ID is 1, and applies to bidirectional streams if it is 0.

Loss or reordering can mean that a MAX_STREAM_ID frame can be received which states a lower stream limit than the client has previously received. MAX_STREAM_ID frames which do not increase the maximum stream ID MUST be ignored.¶

A peer MUST NOT initiate a stream with a higher stream ID than the greatest maximum stream ID it has received. An endpoint MUST terminate a connection with a STREAM_ID_ERROR error if a peer initiates a stream with a higher stream ID than it has sent, unless this is a result of a change in the initial limits (see Section 6.6.2).¶

Endpoints can use PING frames (type=0x07) to verify that their peers are still alive or to check reachability to the peer. The PING frame contains no additional fields.¶

The receiver of a PING frame simply needs to acknowledge the packet containing this frame.¶

The PING frame can be used to keep a connection alive when an application or application protocol wishes to prevent the connection from timing out. An application protocol SHOULD provide guidance about the conditions under which generating a PING is recommended. This guidance SHOULD indicate whether it is the client or the server that is expected to send the PING. Having both endpoints send PING frames without coordination can produce an excessive number of packets and poor performance.¶

A connection will time out if no packets are sent or received for a period longer than the time specified in the idle_timeout transport parameter (see Section 6.13). However, state in middleboxes might time out earlier than that. Though REQ-5 in [RFC4787] recommends a 2 minute timeout interval, experience shows that sending packets every 15 to 30 seconds is necessary to prevent the majority of middleboxes from losing state for UDP flows.¶

A sender SHOULD send a BLOCKED frame (type=0x08) when it wishes to send data, but is unable to due to connection-level flow control (see Section 10.2.1). BLOCKED frames can be used as input to tuning of flow control algorithms (see Section 10.1.2).¶

The BLOCKED frame contains a single field.¶

Offset:

A variable-length integer indicating the connection-level offset at which the blocking occurred.

A sender SHOULD send a STREAM_BLOCKED frame (type=0x09) when it wishes to send data, but is unable to due to stream-level flow control. This frame is analogous to BLOCKED (Section 7.10).¶

An endpoint that receives a STREAM_BLOCKED frame for a send-only stream MUST terminate the connection with error PROTOCOL_VIOLATION.¶

Stream ID:

A variable-length integer indicating the stream which is flow control blocked.

Offset:

A variable-length integer indicating the offset of the stream at which the blocking occurred.

A sender MAY send a STREAM_ID_BLOCKED frame (type=0x0a) when it wishes to open a stream, but is unable to due to the maximum stream ID limit set by its peer (see Section 7.8). This does not open the stream, but informs the peer that a new stream was needed, but the stream limit prevented the creation of the stream.¶

The STREAM_ID_BLOCKED frame contains a single field.¶

Stream ID:

A variable-length integer indicating the highest stream ID that the sender was permitted to open.

An endpoint sends a NEW_CONNECTION_ID frame (type=0x0b) to provide its peer with alternative connection IDs that can be used to break linkability when migrating connections (see Section 6.11.5).¶

Sequence:

A variable-length integer. This value starts at 0 and increases by 1 for each connection ID that is provided by the server. The connection ID that is assigned during the handshake is assumed to have a sequence of -1. That is, the value selected during the handshake comes immediately before the first value that a server can send.

Length:

An 8-bit unsigned integer containing the length of the connection ID. Values less than 4 and greater than 18 are invalid and MUST be treated as a connection error of type PROTOCOL_VIOLATION.

Connection ID:

A connection ID of the specified length.

Stateless Reset Token:

A 128-bit value that will be used to for a stateless reset when the associated connection ID is used (see Section 6.13.4).

An endpoint MUST NOT send this frame if it currently requires that its peer send packets with a zero-length Destination Connection ID. Changing the length of a connection ID to or from zero-length makes it difficult to identify when the value of the connection ID changed. An endpoint that is sending packets with a zero-length Destination Connection ID MUST treat receipt of a NEW_CONNECTION_ID frame as a connection error of type PROTOCOL_VIOLATION.¶

An endpoint may use a STOP_SENDING frame (type=0x0c) to communicate that incoming data is being discarded on receipt at application request. This signals a peer to abruptly terminate transmission on a stream.¶

Receipt of a STOP_SENDING frame is only valid for a send stream that exists and is not in the “Ready” state (see Section 9.2.1). Receiving a STOP_SENDING frame for a send stream that is “Ready” or non-existent MUST be treated as a connection error of type PROTOCOL_VIOLATION. An endpoint that receives a STOP_SENDING frame for a receive-only stream MUST terminate the connection with error PROTOCOL_VIOLATION.¶

Stream ID:

A variable-length integer carrying the Stream ID of the stream being ignored.

Application Error Code:

A 16-bit, application-specified reason the sender is ignoring the stream (see Section 11.4).

Receivers send ACK frames (type=0x0d) to inform senders which packets they have received and processed. The ACK frame contains any number of ACK blocks. ACK blocks are ranges of acknowledged packets.¶

QUIC acknowledgements are irrevocable. Once acknowledged, a packet remains acknowledged, even if it does not appear in a future ACK frame. This is unlike TCP SACKs ([RFC2018]).¶

It is expected that a sender will reuse the same packet number across different packet number spaces. ACK frames only acknowledge the packet numbers that were transmitted by the sender in the same packet number space of the packet that the ACK was received in.¶

A client MUST NOT acknowledge Retry packets. Retry packets include the packet number from the Initial packet it responds to. Version Negotiation packets cannot be acknowledged because they do not contain a packet number. Rather than relying on ACK frames, these packets are implicitly acknowledged by the next Initial packet sent by the client.¶

An ACK frame is shown below.¶

Largest Acknowledged:

A variable-length integer representing the largest packet number the peer is acknowledging; this is usually the largest packet number that the peer has received prior to generating the ACK frame. Unlike the packet number in the QUIC long or short header, the value in an ACK frame is not truncated.

ACK Delay:

A variable-length integer including the time in microseconds that the largest acknowledged packet, as indicated in the Largest Acknowledged field, was received by this peer to when this ACK was sent. The value of the ACK Delay field is scaled by multiplying the encoded value by the 2 to the power of the value of the ack_delay_exponent transport parameter set by the sender of the ACK frame. The ack_delay_exponent defaults to 3, or a multiplier of 8 (see Section 6.6.1). Scaling in this fashion allows for a larger range of values with a shorter encoding at the cost of lower resolution.

ACK Block Count:

A variable-length integer specifying the number of Additional ACK Block (and Gap) fields after the First ACK Block.

ACK Blocks:

Contains one or more blocks of packet numbers which have been successfully received, see Section 7.15.1.

The ACK_ECN frame (type=0x1a) is used by an endpoint that supports ECN to acknowledge packets received with ECN codepoints of ECT(0), ECT(1), or CE in the packet’s IP header.¶

An ACK_ECN frame contains all the elements of the ACK frame (Section 7.15) with the addition of three counts following the ACK Delay field.¶

ECT(0) Count:

A variable-length integer representing the total number packets received with the ECT(0) codepoint.

ECT(1) Count:

A variable-length integer representing the total number packets received with the ECT(1) codepoint.

CE Count:

A variable-length integer representing the total number packets received with the CE codepoint.

Endpoints can use PATH_CHALLENGE frames (type=0x0e) to check reachability to the peer and for path validation during connection establishment and connection migration.¶

PATH_CHALLENGE frames contain an 8-byte payload.¶

Data:

This 8-byte field contains arbitrary data.

A PATH_CHALLENGE frame containing 8 octets that are hard to guess is sufficient to ensure that it is easier to receive the packet than it is to guess the value correctly.¶

The recipient of this frame MUST generate a PATH_RESPONSE frame (Section 7.18) containing the same Data.¶

The PATH_RESPONSE frame (type=0x0f) is sent in response to a PATH_CHALLENGE frame. Its format is identical to the PATH_CHALLENGE frame (Section 7.17).¶

If the content of a PATH_RESPONSE frame does not match the content of a PATH_CHALLENGE frame previously sent by the endpoint, the endpoint MAY generate a connection error of type PROTOCOL_VIOLATION.¶

A server sends a NEW_TOKEN frame (type=0x19) to provide the client a token to send in the header of an Initial packet for a future connection.¶

Token Length:

A variable-length integer specifying the length of the token in bytes.

Token:

An opaque blob that the client may use with a future Initial packet.

STREAM frames implicitly create a stream and carry stream data. The STREAM frame takes the form 0b00010XXX (or the set of values from 0x10 to 0x17). The value of the three low-order bits of the frame type determine the fields that are present in the frame.¶

• The OFF bit (0x04) in the frame type is set to indicate that there is an Offset field present. When set to 1, the Offset field is present; when set to 0, the Offset field is absent and the Stream Data starts at an offset of 0 (that is, the frame contains the first octets of the stream, or the end of a stream that includes no data).
• The LEN bit (0x02) in the frame type is set to indicate that there is a Length field present. If this bit is set to 0, the Length field is absent and the Stream Data field extends to the end of the packet. If this bit is set to 1, the Length field is present.
• The FIN bit (0x01) of the frame type is set only on frames that contain the final offset of the stream. Setting this bit indicates that the frame marks the end of the stream.

An endpoint that receives a STREAM frame for a send-only stream MUST terminate the connection with error PROTOCOL_VIOLATION.¶

A STREAM frame is shown below.¶

Stream ID:

A variable-length integer indicating the stream ID of the stream (see Section 9.1).

Offset:

A variable-length integer specifying the byte offset in the stream for the data in this STREAM frame. This field is present when the OFF bit is set to 1. When the Offset field is absent, the offset is 0.

Length:

A variable-length integer specifying the length of the Stream Data field in this STREAM frame. This field is present when the LEN bit is set to 1. When the LEN bit is set to 0, the Stream Data field consumes all the remaining octets in the packet.

Stream Data:

The bytes from the designated stream to be delivered.

When a Stream Data field has a length of 0, the offset in the STREAM frame is the offset of the next byte that would be sent.¶

The first byte in the stream has an offset of 0. The largest offset delivered on a stream - the sum of the re-constructed offset and data length - MUST be less than 2^62.¶

Stream multiplexing is achieved by interleaving STREAM frames from multiple streams into one or more QUIC packets. A single QUIC packet can include multiple STREAM frames from one or more streams.¶

Implementation note: One of the benefits of QUIC is avoidance of head-of-line blocking across multiple streams. When a packet loss occurs, only streams with data in that packet are blocked waiting for a retransmission to be received, while other streams can continue making progress. Note that when data from multiple streams is bundled into a single QUIC packet, loss of that packet blocks all those streams from making progress. An implementation is therefore advised to bundle as few streams as necessary in outgoing packets without losing transmission efficiency to underfilled packets.¶

The CRYPTO frame (type=0x18) is used to transmit cryptographic handshake messages. It can be sent in all packet types. The CRYPTO frame offers the cryptographic protocol an in-order stream of bytes. CRYPTO frames are functionally identical to STREAM frames, except that they do not bear a stream identifier; they are not flow controlled; and they do not carry markers for optional offset, optional length, and the end of the stream.¶

A CRYPTO frame is shown below.¶

Offset:

A variable-length integer specifying the byte offset in the stream for the data in this CRYPTO frame.

Length:

A variable-length integer specifying the length of the Crypto Data field in this CRYPTO frame.

Crypto Data:

The cryptographic message data.

There is a separate flow of cryptographic handshake data in each encryption level, each of which starts at an offset of 0. This implies that each encryption level is treated as a separate CRYPTO stream of data.¶

Unlike STREAM frames, which include a Stream ID indicating to which stream the data belongs, the CRYPTO frame carries data for a single stream per encryption level. The stream does not have an explicit end, so CRYPTO frames do not have a FIN bit.¶

A packet MUST NOT be acknowledged until packet protection has been successfully removed and all frames contained in the packet have been processed. Any stream state transitions triggered by the frame MUST have occurred. For STREAM frames, this means the data has been enqueued in preparation to be received by the application protocol, but it does not require that data is delivered and consumed.¶

Once the packet has been fully processed, a receiver acknowledges receipt by sending one or more ACK frames containing the packet number of the received packet. To avoid creating an indefinite feedback loop, an endpoint MUST NOT send an ACK frame in response to a packet containing only ACK or PADDING frames, even if there are packet gaps which precede the received packet. The endpoint MUST acknowledge packets containing only ACK or PADDING frames in the next ACK frame that it sends.¶

While PADDING frames do not elicit an ACK frame from a receiver, they are considered to be in flight for congestion control purposes [QUIC-RECOVERY]. Sending only PADDING frames might cause the sender to become limited by the congestion controller (as described in [QUIC-RECOVERY]) with no acknowledgments forthcoming from the receiver. Therefore, a sender should ensure that other frames are sent in addition to PADDING frames to elicit acknowledgments from the receiver.¶

Strategies and implications of the frequency of generating acknowledgments are discussed in more detail in [QUIC-RECOVERY].¶

QUIC packets that are determined to be lost are not retransmitted whole. The same applies to the frames that are contained within lost packets. Instead, the information that might be carried in frames is sent again in new frames as needed.¶

New frames and packets are used to carry information that is determined to have been lost. In general, information is sent again when a packet containing that information is determined to be lost and sending ceases when a packet containing that information is acknowledged.¶

• Data sent in CRYPTO frames are retransmitted according to the rules in [QUIC-RECOVERY], until either all data has been acknowledged or the crypto state machine implicitly knows that the peer received the data.
• Application data sent in STREAM frames is retransmitted in new STREAM frames unless the endpoint has sent a RST_STREAM for that stream. Once an endpoint sends a RST_STREAM frame, no further STREAM frames are needed.
• The most recent set of acknowledgments are sent in ACK frames. An ACK frame SHOULD contain all unacknowledged acknowledgments, as described in Section 7.15.2.
• Cancellation of stream transmission, as carried in a RST_STREAM frame, is sent until acknowledged or until all stream data is acknowledged by the peer (that is, either the “Reset Recvd” or “Data Recvd” state is reached on the send stream). The content of a RST_STREAM frame MUST NOT change when it is sent again.
• Similarly, a request to cancel stream transmission, as encoded in a STOP_SENDING frame, is sent until the receive stream enters either a “Data Recvd” or “Reset Recvd” state, see Section 9.3.
• Connection close signals, including those that use CONNECTION_CLOSE and APPLICATION_CLOSE frames, are not sent again when packet loss is detected, but as described in Section 6.13.
• The current connection maximum data is sent in MAX_DATA frames. An updated value is sent in a MAX_DATA frame if the packet containing the most recently sent MAX_DATA frame is declared lost, or when the endpoint decides to update the limit. Care is necessary to avoid sending this frame too often as the limit can increase frequently and cause an unnecessarily large number of MAX_DATA frames to be sent.
• The current maximum stream data offset is sent in MAX_STREAM_DATA frames. Like MAX_DATA, an updated value is sent when the packet containing the most recent MAX_STREAM_DATA frame for a stream is lost or when the limit is updated, with care taken to prevent the frame from being sent too often. An endpoint SHOULD stop sending MAX_STREAM_DATA frames when the receive stream enters a “Size Known” state.
• The maximum stream ID for a stream of a given type is sent in MAX_STREAM_ID frames. Like MAX_DATA, an updated value is sent when a packet containing the most recent MAX_STREAM_ID for a stream type frame is declared lost or when the limit is updated, with care taken to prevent the frame from being sent too often.
• Blocked signals are carried in BLOCKED, STREAM_BLOCKED, and STREAM_ID_BLOCKED frames. BLOCKED streams have connection scope, STREAM_BLOCKED frames have stream scope, and STREAM_ID_BLOCKED frames are scoped to a specific stream type. New frames are sent if packets containing the most recent frame for a scope is lost, but only while the endpoint is blocked on the corresponding limit. These frames always include the limit that is causing blocking at the time that they are transmitted.
• A liveness or path validation check using PATH_CHALLENGE frames is sent periodically until a matching PATH_RESPONSE frame is received or until there is no remaining need for liveness or path validation checking. PATH_CHALLENGE frames include a different payload each time they are sent.
• Responses to path validation using PATH_RESPONSE frames are sent just once. A new PATH_CHALLENGE frame will be sent if another PATH_RESPONSE frame is needed.
• New connection IDs are sent in NEW_CONNECTION_ID frames and retransmitted if the packet containing them is lost.
• PADDING frames contain no information, so lost PADDING frames do not require repair.

Upon detecting losses, a sender MUST take appropriate congestion control action. The details of loss detection and congestion control are described in [QUIC-RECOVERY].¶

The QUIC packet size includes the QUIC header and integrity check, but not the UDP or IP header.¶

Clients MUST ensure that the first Initial packet it sends is sent in a UDP datagram that is at least 1200 octets. Padding the Initial packet or including a 0-RTT packet in the same datagram are ways to meet this requirement. Sending a UDP datagram of this size ensures that the network path supports a reasonable Maximum Transmission Unit (MTU), and helps reduce the amplitude of amplification attacks caused by server responses toward an unverified client address.¶

The datagram containing the first Initial packet from a client MAY exceed 1200 octets if the client believes that the Path Maximum Transmission Unit (PMTU) supports the size that it chooses.¶

A server MAY send a CONNECTION_CLOSE frame with error code PROTOCOL_VIOLATION in response to the first Initial packet it receives from a client if the UDP datagram is smaller than 1200 octets. It MUST NOT send any other frame type in response, or otherwise behave as if any part of the offending packet was processed as valid.¶

The Path Maximum Transmission Unit (PMTU) is the maximum size of the entire IP header, UDP header, and UDP payload. The UDP payload includes the QUIC packet header, protected payload, and any authentication fields.¶

All QUIC packets SHOULD be sized to fit within the estimated PMTU to avoid IP fragmentation or packet drops. To optimize bandwidth efficiency, endpoints SHOULD use Packetization Layer PMTU Discovery ([PLPMTUD]). Endpoints MAY use PMTU Discovery ([PMTUDv4], [PMTUDv6]) for detecting the PMTU, setting the PMTU appropriately, and storing the result of previous PMTU determinations.¶

In the absence of these mechanisms, QUIC endpoints SHOULD NOT send IP packets larger than 1280 octets. Assuming the minimum IP header size, this results in a QUIC packet size of 1232 octets for IPv6 and 1252 octets for IPv4. Some QUIC implementations MAY wish to be more conservative in computing allowed QUIC packet size given unknown tunneling overheads or IP header options.¶

QUIC endpoints that implement any kind of PMTU discovery SHOULD maintain an estimate for each combination of local and remote IP addresses. Each pairing of local and remote addresses could have a different maximum MTU in the path.¶

QUIC depends on the network path supporting a MTU of at least 1280 octets. This is the IPv6 minimum MTU and therefore also supported by most modern IPv4 networks. An endpoint MUST NOT reduce its MTU below this number, even if it receives signals that indicate a smaller limit might exist.¶

If a QUIC endpoint determines that the PMTU between any pair of local and remote IP addresses has fallen below 1280 octets, it MUST immediately cease sending QUIC packets on the affected path. This could result in termination of the connection if an alternative path cannot be found.¶

Streams are identified by an unsigned 62-bit integer, referred to as the Stream ID. The least significant two bits of the Stream ID are used to identify the type of stream (unidirectional or bidirectional) and the initiator of the stream.¶

The least significant bit (0x1) of the Stream ID identifies the initiator of the stream. Clients initiate even-numbered streams (those with the least significant bit set to 0); servers initiate odd-numbered streams (with the bit set to 1). Separation of the stream identifiers ensures that client and server are able to open streams without the latency imposed by negotiating for an identifier.¶

If an endpoint receives a frame for a stream that it expects to initiate (i.e., odd-numbered for the client or even-numbered for the server), but which it has not yet opened, it MUST close the connection with error code STREAM_STATE_ERROR.¶

The second least significant bit (0x2) of the Stream ID differentiates between unidirectional streams and bidirectional streams. Unidirectional streams always have this bit set to 1 and bidirectional streams have this bit set to 0.¶

The two type bits from a Stream ID therefore identify streams as summarized in Table 5.¶

The first bi-directional stream opened by the client is stream 0.¶

A QUIC endpoint MUST NOT reuse a Stream ID. Streams of each type are created in numeric order. Streams that are used out of order result in opening all lower-numbered streams of the same type in the same direction.¶

Stream IDs are encoded as a variable-length integer (see Section 7.1).¶

This section describes the two types of QUIC stream in terms of the states of their send or receive components. Two state machines are described: one for streams on which an endpoint transmits data (Section 9.2.1); another for streams from which an endpoint receives data (Section 9.2.2).¶

Unidirectional streams use the applicable state machine directly. Bidirectional streams use both state machines. For the most part, the use of these state machines is the same whether the stream is unidirectional or bidirectional. The conditions for opening a stream are slightly more complex for a bidirectional stream because the opening of either send or receive sides causes the stream to open in both directions.¶

An endpoint can open streams up to its maximum stream limit in any order, however endpoints SHOULD open the send side of streams for each type in order.¶

Note:

These states are largely informative. This document uses stream states to describe rules for when and how different types of frames can be sent and the reactions that are expected when different types of frames are received. Though these state machines are intended to be useful in implementing QUIC, these states aren’t intended to constrain implementations. An implementation can define a different state machine as long as its behavior is consistent with an implementation that implements these states.

If an endpoint is no longer interested in the data it is receiving on a stream, it MAY send a STOP_SENDING frame identifying that stream to prompt closure of the stream in the opposite direction. This typically indicates that the receiving application is no longer reading data it receives from the stream, but is not a guarantee that incoming data will be ignored.¶

STREAM frames received after sending STOP_SENDING are still counted toward the connection and stream flow-control windows, even though these frames will be discarded upon receipt. This avoids potential ambiguity about which STREAM frames count toward flow control.¶

A STOP_SENDING frame requests that the receiving endpoint send a RST_STREAM frame. An endpoint that receives a STOP_SENDING frame MUST send a RST_STREAM frame for that stream, and can use an error code of STOPPING. If the STOP_SENDING frame is received on a send stream that is already in the “Data Sent” state, a RST_STREAM frame MAY still be sent in order to cancel retransmission of previously-sent STREAM frames.¶

STOP_SENDING SHOULD only be sent for a receive stream that has not been reset. STOP_SENDING is most useful for streams in the “Recv” or “Size Known” states.¶

An endpoint is expected to send another STOP_SENDING frame if a packet containing a previous STOP_SENDING is lost. However, once either all stream data or a RST_STREAM frame has been received for the stream - that is, the stream is in any state other than “Recv” or “Size Known” - sending a STOP_SENDING frame is unnecessary.¶

An endpoint limits the number of concurrently active incoming streams by adjusting the maximum stream ID. An initial value is set in the transport parameters (see Section 6.6.1) and is subsequently increased by MAX_STREAM_ID frames (see Section 7.8).¶

The maximum stream ID is specific to each endpoint and applies only to the peer that receives the setting. That is, clients specify the maximum stream ID the server can initiate, and servers specify the maximum stream ID the client can initiate. Each endpoint may respond on streams initiated by the other peer, regardless of whether it is permitted to initiated new streams.¶

Endpoints MUST NOT exceed the limit set by their peer. An endpoint that receives a STREAM frame with an ID greater than the limit it has sent MUST treat this as a stream error of type STREAM_ID_ERROR (Section 11), unless this is a result of a change in the initial limits (see Section 6.6.2).¶

A receiver cannot renege on an advertisement; that is, once a receiver advertises a stream ID via a MAX_STREAM_ID frame, advertising a smaller maximum ID has no effect. A sender MUST ignore any MAX_STREAM_ID frame that does not increase the maximum stream ID.¶

Once a stream is created, endpoints may use the stream to send and receive data. Each endpoint may send a series of STREAM frames encapsulating data on a stream until the stream is terminated in that direction. Streams are an ordered byte-stream abstraction, and they have no other structure within them. STREAM frame boundaries are not expected to be preserved in retransmissions from the sender or during delivery to the application at the receiver.¶

When new data is to be sent on a stream, a sender MUST set the encapsulating STREAM frame’s offset field to the stream offset of the first byte of this new data. The first octet of data on a stream has an offset of 0. An endpoint is expected to send every stream octet. The largest offset delivered on a stream MUST be less than 2^62.¶

QUIC makes no specific allowances for partial reliability or delivery of stream data out of order. Endpoints MUST be able to deliver stream data to an application as an ordered byte-stream. Delivering an ordered byte-stream requires that an endpoint buffer any data that is received out of order, up to the advertised flow control limit.¶

An endpoint could receive the same octets multiple times; octets that have already been received can be discarded. The value for a given octet MUST NOT change if it is sent multiple times; an endpoint MAY treat receipt of a changed octet as a connection error of type PROTOCOL_VIOLATION.¶

An endpoint MUST NOT send data on any stream without ensuring that it is within the data limits set by its peer.¶

Flow control is described in detail in Section 10, and congestion control is described in the companion document [QUIC-RECOVERY].¶

Stream multiplexing has a significant effect on application performance if resources allocated to streams are correctly prioritized. Experience with other multiplexed protocols, such as HTTP/2 [HTTP2], shows that effective prioritization strategies have a significant positive impact on performance.¶

QUIC does not provide frames for exchanging prioritization information. Instead it relies on receiving priority information from the application that uses QUIC. Protocols that use QUIC are able to define any prioritization scheme that suits their application semantics. A protocol might define explicit messages for signaling priority, such as those defined in HTTP/2; it could define rules that allow an endpoint to determine priority based on context; or it could leave the determination to the application.¶

A QUIC implementation SHOULD provide ways in which an application can indicate the relative priority of streams. When deciding which streams to dedicate resources to, QUIC SHOULD use the information provided by the application. Failure to account for priority of streams can result in suboptimal performance.¶

Stream priority is most relevant when deciding which stream data will be transmitted. Often, there will be limits on what can be transmitted as a result of connection flow control or the current congestion controller state.¶

Giving preference to the transmission of its own management frames ensures that the protocol functions efficiently. That is, prioritizing frames other than STREAM frames ensures that loss recovery, congestion control, and flow control operate effectively.¶

CRYPTO frames SHOULD be prioritized over other streams prior to the completion of the cryptographic handshake. This includes the retransmission of the second flight of client handshake messages, that is, the TLS Finished and any client authentication messages.¶

STREAM data in frames determined to be lost SHOULD be retransmitted before sending new data, unless application priorities indicate otherwise. Retransmitting lost stream data can fill in gaps, which allows the peer to consume already received data and free up flow control window.¶

There are some edge cases which must be considered when dealing with stream and connection level flow control. Given enough time, both endpoints must agree on flow control state. If one end believes it can send more than the other end is willing to receive, the connection will be torn down when too much data arrives.¶

Conversely if a sender believes it is blocked, while endpoint B expects more data can be received, then the connection can be in a deadlock, with the sender waiting for a MAX_DATA or MAX_STREAM_DATA frame which will never come.¶

On receipt of a RST_STREAM frame, an endpoint will tear down state for the matching stream and ignore further data arriving on that stream. This could result in the endpoints getting out of sync, since the RST_STREAM frame may have arrived out of order and there may be further bytes in flight. The data sender would have counted the data against its connection level flow control budget, but a receiver that has not received these bytes would not know to include them as well. The receiver must learn the number of bytes that were sent on the stream to make the same adjustment in its connection flow controller.¶

To avoid this de-synchronization, a RST_STREAM sender MUST include the final byte offset sent on the stream in the RST_STREAM frame. On receiving a RST_STREAM frame, a receiver definitively knows how many bytes were sent on that stream before the RST_STREAM frame, and the receiver MUST use the final offset to account for all bytes sent on the stream in its connection level flow controller.¶

As with flow control, this document leaves when and how many streams to make available to a peer via MAX_STREAM_ID to implementations, but offers a few considerations. MAX_STREAM_ID frames constitute minimal overhead, while withholding MAX_STREAM_ID frames can prevent the peer from using the available parallelism.¶

Implementations will likely want to increase the maximum stream ID as peer-initiated streams close. A receiver MAY also advance the maximum stream ID based on current activity, system conditions, and other environmental factors.¶

The final offset is the count of the number of octets that are transmitted on a stream. For a stream that is reset, the final offset is carried explicitly in a RST_STREAM frame. Otherwise, the final offset is the offset of the end of the data carried in a STREAM frame marked with a FIN flag, or 0 in the case of incoming unidirectional streams.¶

An endpoint will know the final offset for a stream when the receive stream enters the “Size Known” or “Reset Recvd” state.¶

An endpoint MUST NOT send data on a stream at or beyond the final offset.¶

Once a final offset for a stream is known, it cannot change. If a RST_STREAM or STREAM frame causes the final offset to change for a stream, an endpoint SHOULD respond with a FINAL_OFFSET_ERROR error (see Section 11). A receiver SHOULD treat receipt of data at or beyond the final offset as a FINAL_OFFSET_ERROR error, even after a stream is closed. Generating these errors is not mandatory, but only because requiring that an endpoint generate these errors also means that the endpoint needs to maintain the final offset state for closed streams, which could mean a significant state commitment.¶

Data sent in CRYPTO frames is not flow controlled in the same way as STREAM frames. QUIC relies on the cryptographic protocol implementation to avoid excessive buffering of data, see [QUIC-TLS]. The implementation SHOULD provide an interface to QUIC to tell it about its buffering limits so that there is not excessive buffering at multiple layers.¶

Errors that result in the connection being unusable, such as an obvious violation of protocol semantics or corruption of state that affects an entire connection, MUST be signaled using a CONNECTION_CLOSE or APPLICATION_CLOSE frame (Section 7.4, Section 7.5). An endpoint MAY close the connection in this manner even if the error only affects a single stream.¶

Application protocols can signal application-specific protocol errors using the APPLICATION_CLOSE frame. Errors that are specific to the transport, including all those described in this document, are carried in a CONNECTION_CLOSE frame. Other than the type of error code they carry, these frames are identical in format and semantics.¶

A CONNECTION_CLOSE or APPLICATION_CLOSE frame could be sent in a packet that is lost. An endpoint SHOULD be prepared to retransmit a packet containing either frame type if it receives more packets on a terminated connection. Limiting the number of retransmissions and the time over which this final packet is sent limits the effort expended on terminated connections.¶

An endpoint that chooses not to retransmit packets containing CONNECTION_CLOSE or APPLICATION_CLOSE risks a peer missing the first such packet. The only mechanism available to an endpoint that continues to receive data for a terminated connection is to use the stateless reset process (Section 6.13.4).¶

An endpoint that receives an invalid CONNECTION_CLOSE or APPLICATION_CLOSE frame MUST NOT signal the existence of the error to its peer.¶

If an application-level error affects a single stream, but otherwise leaves the connection in a recoverable state, the endpoint can send a RST_STREAM frame (Section 7.3) with an appropriate error code to terminate just the affected stream.¶

Other than STOPPING (Section 9.3), RST_STREAM MUST be instigated by the application and MUST carry an application error code. Resetting a stream without knowledge of the application protocol could cause the protocol to enter an unrecoverable state. Application protocols might require certain streams to be reliably delivered in order to guarantee consistent state between endpoints.¶

QUIC error codes are 16-bit unsigned integers.¶

This section lists the defined QUIC transport error codes that may be used in a CONNECTION_CLOSE frame. These errors apply to the entire connection.¶

NO_ERROR (0x0):

An endpoint uses this with CONNECTION_CLOSE to signal that the connection is being closed abruptly in the absence of any error.

INTERNAL_ERROR (0x1):

The endpoint encountered an internal error and cannot continue with the connection.

SERVER_BUSY (0x2):

The server is currently busy and does not accept any new connections.

FLOW_CONTROL_ERROR (0x3):

An endpoint received more data than it permitted in its advertised data limits (see Section 10).

STREAM_ID_ERROR (0x4):

An endpoint received a frame for a stream identifier that exceeded its advertised maximum stream ID.

STREAM_STATE_ERROR (0x5):

An endpoint received a frame for a stream that was not in a state that permitted that frame (see Section 9.2).

FINAL_OFFSET_ERROR (0x6):

An endpoint received a STREAM frame containing data that exceeded the previously established final offset. Or an endpoint received a RST_STREAM frame containing a final offset that was lower than the maximum offset of data that was already received. Or an endpoint received a RST_STREAM frame containing a different final offset to the one already established.

FRAME_ENCODING_ERROR (0x7):

An endpoint received a frame that was badly formatted. For instance, an empty STREAM frame that omitted the FIN flag, or an ACK frame that has more acknowledgment ranges than the remainder of the packet could carry.

TRANSPORT_PARAMETER_ERROR (0x8):

An endpoint received transport parameters that were badly formatted, included an invalid value, was absent even though it is mandatory, was present though it is forbidden, or is otherwise in error.

VERSION_NEGOTIATION_ERROR (0x9):

An endpoint received transport parameters that contained version negotiation parameters that disagreed with the version negotiation that it performed. This error code indicates a potential version downgrade attack.

PROTOCOL_VIOLATION (0xA):

An endpoint detected an error with protocol compliance that was not covered by more specific error codes.

INVALID_MIGRATION (0xC):

A peer has migrated to a different network when the endpoint had disabled migration.

CRYPTO_ERROR (0x1XX):

The cryptographic handshake failed. A range of 256 values is reserved for carrying error codes specific to the cryptographic handshake that is used. Codes for errors occurring when TLS is used for the crypto handshake are described in Section 4.8 of [QUIC-TLS].

See Section 13.3 for details of registering new error codes.¶

Application protocol error codes are 16-bit unsigned integers, but the management of application error codes are left to application protocols. Application protocol error codes are used for the RST_STREAM (Section 7.3) and APPLICATION_CLOSE (Section 7.5) frames.¶

There is no restriction on the use of the 16-bit error code space for application protocols. However, QUIC reserves the error code with a value of 0 to mean STOPPING. The application error code of STOPPING (0) is used by the transport to cancel a stream in response to receipt of a STOP_SENDING frame.¶

As an encrypted and authenticated transport QUIC provides a range of protections against denial of service. Once the cryptographic handshake is complete, QUIC endpoints discard most packets that are not authenticated, greatly limiting the ability of an attacker to interfere with existing connections.¶

Once a connection is established QUIC endpoints might accept some unauthenticated ICMP packets (see Section 8.4.1), but the use of these packets is extremely limited. The only other type of packet that an endpoint might accept is a stateless reset (Section 6.13.4) which relies on the token being kept secret until it is used.¶

During the creation of a connection, QUIC only provides protection against attack from off the network path. All QUIC packets contain proof that the recipient saw a preceding packet from its peer.¶

The first mechanism used is the source and destination connection IDs, which are required to match those set by a peer. Except for an Initial and stateless reset packets, an endpoint only accepts packets that include a destination connection that matches a connection ID the endpoint previously chose. This is the only protection offered for Version Negotiation packets.¶

The destination connection ID in an Initial packet is selected by a client to be unpredictable, which serves an additional purpose. The packets that carry the cryptographic handshake are protected with a key that is derived from this connection ID and salt specific to the QUIC version. This allows endpoints to use the same process for authenticating packets that they receive as they use after the cryptographic handshake completes. Packets that cannot be authenticated are discarded. Protecting packets in this fashion provides a strong assurance that the sender of the packet saw the Initial packet and understood it.¶

These protections are not intended to be effective against an attacker that is able to receive QUIC packets prior to the connection being established. Such an attacker can potentially send packets that will be accepted by QUIC endpoints. This version of QUIC attempts to detect this sort of attack, but it expects that endpoints will fail to establish a connection rather than recovering. For the most part, the cryptographic handshake protocol [QUIC-TLS] is responsible for detecting tampering during the handshake, though additional validation is required for version negotiation (see Section 6.6.4).¶

Endpoints are permitted to use other methods to detect and attempt to recover from interference with the handshake. Invalid packets may be identified and discarded using other methods, but no specific method is mandated in this document.¶

An attacker might be able to receive an address validation token (Section 6.9) from the server and then release the IP address it used to acquire that token. The attacker may, in the future, spoof this same address (which now presumably addresses a different endpoint), and initiate a 0-RTT connection with a server on the victim’s behalf. The attacker can then spoof ACK frames to the server which cause the server to send excessive amounts of data toward the new owner of the IP address.¶

There are two possible mitigations to this attack. The simplest one is that a server can unilaterally create a gap in packet-number space. In the non-attack scenario, the client will send an ACK frame with the larger value for largest acknowledged. In the attack scenario, the attacker could acknowledge a packet in the gap. If the server sees an acknowledgment for a packet that was never sent, the connection can be aborted.¶

The second mitigation is that the server can require that acknowledgments for sent packets match the encryption level of the sent packet. This mitigation is useful if the connection has an ephemeral forward-secure key that is generated and used for every new connection. If a packet sent is protected with a forward-secure key, then any acknowledgments that are received for them MUST also be forward-secure protected. Since the attacker will not have the forward secure key, the attacker will not be able to generate forward-secure protected packets with ACK frames.¶

An endpoint that acknowledges packets it has not received might cause a congestion controller to permit sending at rates beyond what the network supports. An endpoint MAY skip packet numbers when sending packets to detect this behavior. An endpoint can then immediately close the connection with a connection error of type PROTOCOL_VIOLATION (see Section 6.13.3).¶

The attacks commonly known as Slowloris [SLOWLORIS] try to keep many connections to the target endpoint open and hold them open as long as possible. These attacks can be executed against a QUIC endpoint by generating the minimum amount of activity necessary to avoid being closed for inactivity. This might involve sending small amounts of data, gradually opening flow control windows in order to control the sender rate, or manufacturing ACK frames that simulate a high loss rate.¶

QUIC deployments SHOULD provide mitigations for the Slowloris attacks, such as increasing the maximum number of clients the server will allow, limiting the number of connections a single IP address is allowed to make, imposing restrictions on the minimum transfer speed a connection is allowed to have, and restricting the length of time an endpoint is allowed to stay connected.¶

An adversarial endpoint might intentionally fragment the data on stream buffers in order to cause disproportionate memory commitment. An adversarial endpoint could open a stream and send some STREAM frames containing arbitrary fragments of the stream content.¶

The attack is mitigated if flow control windows correspond to available memory. However, some receivers will over-commit memory and advertise flow control offsets in the aggregate that exceed actual available memory. The over-commitment strategy can lead to better performance when endpoints are well behaved, but renders endpoints vulnerable to the stream fragmentation attack.¶

QUIC deployments SHOULD provide mitigations against the stream fragmentation attack. Mitigations could consist of avoiding over-committing memory, delaying reassembly of STREAM frames, implementing heuristics based on the age and duration of reassembly holes, or some combination.¶

An adversarial endpoint can open lots of streams, exhausting state on an endpoint. The adversarial endpoint could repeat the process on a large number of connections, in a manner similar to SYN flooding attacks in TCP.¶

Normally, clients will open streams sequentially, as explained in Section 9.1. However, when several streams are initiated at short intervals, transmission error may cause STREAM DATA frames opening streams to be received out of sequence. A receiver is obligated to open intervening streams if a higher-numbered stream ID is received. Thus, on a new connection, opening stream 2000001 opens 1 million streams, as required by the specification.¶

The number of active streams is limited by the concurrent stream limit transport parameter, as explained in Section 9.4. If chosen judiciously, this limit mitigates the effect of the stream commitment attack. However, setting the limit too low could affect performance when applications expect to open large number of streams.¶

An on-path attacker could manipulate the value of ECN codepoints in the IP header to influence the sender’s rate. [RFC3168] discusses manipulations and their effects in more detail.¶

An on-the-side attacker can duplicate and send packets with modified ECN codepoints to affect the sender’s rate. If duplicate packets are discarded by a receiver, an off-path attacker will need to race the duplicate packet against the original to be successful in this attack. Therefore, QUIC receivers ignore ECN codepoints set in duplicate packets (see Section 6.8).¶

Stateless resets create a possible denial of service attack analogous to a TCP reset injection. This attack is possible if an attacker is able to cause a stateless reset token to be generated for a connection with a selected connection ID. An attacker that can cause this token to be generated can reset an active connection with the same connection ID.¶

If a packet can be routed to different instances that share a static key, for example by changing an IP address or port, then an attacker can cause the server to send a stateless reset. To defend against this style of denial service, endpoints that share a static key for stateless reset (see Section 6.13.4.2) MUST be arranged so that packets with a given connection ID always arrive at an instance that has connection state, unless that connection is no longer active.¶

In the case of a cluster that uses dynamic load balancing, it’s possible that a change in load balancer configuration could happen while an active instance retains connection state; even if an instance retains connection state, the change in routing and resulting stateless reset will result in the connection being terminated. If there is no chance in the packet being routed to the correct instance, it is better to send a stateless reset than wait for connections to time out. However, this is acceptable only if the routing cannot be influenced by an attacker.¶

IANA [SHALL add/has added] a registry for “QUIC Transport Parameters” under a “QUIC Protocol” heading.¶

The “QUIC Transport Parameters” registry governs a 16-bit space. This space is split into two spaces that are governed by different policies. Values with the first byte in the range 0x00 to 0xfe (in hexadecimal) are assigned via the Specification Required policy [RFC8126]. Values with the first byte 0xff are reserved for Private Use [RFC8126].¶

Value:

The numeric value of the assignment (registrations will be between 0x0000 and 0xfeff).

Parameter Name:

A short mnemonic for the parameter.

Specification:

A reference to a publicly available specification for the value.

The nominated expert(s) verify that a specification exists and is readily accessible. Expert(s) are encouraged to be biased towards approving registrations unless they are abusive, frivolous, or actively harmful (not merely aesthetically displeasing, or architecturally dubious).¶

The initial contents of this registry are shown in Table 7.¶

IANA [SHALL add/has added] a registry for “QUIC Frame Types” under a “QUIC Protocol” heading.¶

The “QUIC Frame Types” registry governs a 62-bit space. This space is split into three spaces that are governed by different policies. Values between 0x00 and 0x3f (in hexadecimal) are assigned via the Standards Action or IESG Review policies [RFC8126]. Values from 0x40 to 0x3fff operate on the Specification Required policy [RFC8126]. All other values are assigned to Private Use [RFC8126].¶

Value:

The numeric value of the assignment (registrations will be between 0x00 and 0x3fff). A range of values MAY be assigned.

Frame Name:

A short mnemonic for the frame type.

Specification:

A reference to a publicly available specification for the value.

The nominated expert(s) verify that a specification exists and is readily accessible. Specifications for new registrations need to describe the means by which an endpoint might determine that it can send the identified type of frame. An accompanying transport parameter registration (see Section 13.1) is expected for most registrations. The specification needs to describe the format and assigned semantics of any fields in the frame.¶

Expert(s) are encouraged to be biased towards approving registrations unless they are abusive, frivolous, or actively harmful (not merely aesthetically displeasing, or architecturally dubious).¶

The initial contents of this registry are tabulated in Table 3.¶

IANA [SHALL add/has added] a registry for “QUIC Transport Error Codes” under a “QUIC Protocol” heading.¶

The “QUIC Transport Error Codes” registry governs a 16-bit space. This space is split into two spaces that are governed by different policies. Values with the first byte in the range 0x00 to 0xfe (in hexadecimal) are assigned via the Specification Required policy [RFC8126]. Values with the first byte 0xff are reserved for Private Use [RFC8126].¶

Value:

The numeric value of the assignment (registrations will be between 0x0000 and 0xfeff).

Code:

A short mnemonic for the parameter.

Description:

A brief description of the error code semantics, which MAY be a summary if a specification reference is provided.

Specification:

A reference to a publicly available specification for the value.

The initial contents of this registry are shown in Table 8. Values from 0xFF00 to 0xFFFF are reserved for Private Use [RFC8126].¶

• Streams open when higher-numbered streams of the same type open (#1342, #1549)
• Split initial stream flow control limit into 3 transport parameters (#1016, #1542)
• All flow control transport parameters are optional (#1610)
• Removed UNSOLICITED_PATH_RESPONSE error code (#1265, #1539)
• Permit stateless reset in response to any packet (#1348, #1553)
• Recommended defense against stateless reset spoofing (#1386, #1554)
• Prevent infinite stateless reset exchanges (#1443, #1627)
• Forbid processing of the same packet number twice (#1405, #1624)
• Added a packet number decoding example (#1493)
• More precisely define idle timeout (#1429, #1614, #1652)
• Corrected format of Retry packet and prevented looping (#1492, #1451, #1448, #1498)
• Permit 0-RTT after receiving Version Negotiation or Retry (#1507, #1514, #1621)
• Permit Retry in response to 0-RTT (#1547, #1552)
• Looser verification of ECN counters to account for ACK loss (#1555, #1481, #1565)
• Remove frame type field from APPLICATION_CLOSE (#1508, #1528)

• Changes to integration of the TLS handshake (#829, #1018, #1094, #1165, #1190, #1233, #1242, #1252, #1450, #1458)
  • The cryptographic handshake uses CRYPTO frames, not stream 0
  • QUIC packet protection is used in place of TLS record protection
  • Separate QUIC packet number spaces are used for the handshake
  • Changed Retry to be independent of the cryptographic handshake
  • Added NEW_TOKEN frame and Token fields to Initial packet
  • Limit the use of HelloRetryRequest to address TLS needs (like key shares)
• Enable server to transition connections to a preferred address (#560, #1251, #1373)
• Added ECN feedback mechanisms and handling; new ACK_ECN frame (#804, #805, #1372)
• Changed rules and recommendations for use of new connection IDs (#1258, #1264, #1276, #1280, #1419, #1452, #1453, #1465)
• Added a transport parameter to disable intentional connection migration (#1271, #1447)
• Packets from different connection ID can’t be coalesced (#1287, #1423)
• Fixed sampling method for packet number encryption; the length field in long headers includes the packet number field in addition to the packet payload (#1387, #1389)
• Stateless Reset is now symmetric and subject to size constraints (#466, #1346)
• Added frame type extension mechanism (#58, #1473)

• Enable server to transition connections to a preferred address (#560, #1251)
• Packet numbers are encrypted (#1174, #1043, #1048, #1034, #850, #990, #734, #1317, #1267, #1079)
• Packet numbers use a variable-length encoding (#989, #1334)
• STREAM frames can now be empty (#1350)

• Swap payload length and packed number fields in long header (#1294)
• Clarified that CONNECTION_CLOSE is allowed in Handshake packet (#1274)
• Spin bit reserved (#1283)
• Coalescing multiple QUIC packets in a UDP datagram (#1262, #1285)
• A more complete connection migration (#1249)
• Refine opportunistic ACK defense text (#305, #1030, #1185)
• A Stateless Reset Token isn’t mandatory (#818, #1191)
• Removed implicit stream opening (#896, #1193)
• An empty STREAM frame can be used to open a stream without sending data (#901, #1194)
• Define stream counts in transport parameters rather than a maximum stream ID (#1023, #1065)
• STOP_SENDING is now prohibited before streams are used (#1050)
• Recommend including ACK in Retry packets and allow PADDING (#1067, #882)
• Endpoints now become closing after an idle timeout (#1178, #1179)
• Remove implication that Version Negotiation is sent when a packet of the wrong version is received (#1197)

• Added PATH_CHALLENGE and PATH_RESPONSE frames to replace PING with Data and PONG frame. Changed ACK frame type from 0x0e to 0x0d. (#1091, #725, #1086)
• A server can now only send 3 packets without validating the client address (#38, #1090)
• Delivery order of stream data is no longer strongly specified (#252, #1070)
• Rework of packet handling and version negotiation (#1038)
• Stream 0 is now exempt from flow control until the handshake completes (#1074, #725, #825, #1082)
• Improved retransmission rules for all frame types: information is retransmitted, not packets or frames (#463, #765, #1095, #1053)
• Added an error code for server busy signals (#1137)
• Endpoints now set the connection ID that their peer uses. Connection IDs are variable length. Removed the omit_connection_id transport parameter and the corresponding short header flag. (#1089, #1052, #1146, #821, #745, #821, #1166, #1151)

• Clarified requirements for BLOCKED usage (#65, #924)
• BLOCKED frame now includes reason for blocking (#452, #924, #927, #928)
• GAP limitation in ACK Frame (#613)
• Improved PMTUD description (#614, #1036)
• Clarified stream state machine (#634, #662, #743, #894)
• Reserved versions don’t need to be generated deterministically (#831, #931)
• You don’t always need the draining period (#871)
• Stateless reset clarified as version-specific (#930, #986)
• initial_max_stream_id_x transport parameters are optional (#970, #971)
• Ack Delay assumes a default value during the handshake (#1007, #1009)
• Removed transport parameters from NewSessionTicket (#1015)

• The long header now has version before packet number (#926, #939)
• Rename and consolidate packet types (#846, #822, #847)
• Packet types are assigned new codepoints and the Connection ID Flag is inverted (#426, #956)
• Removed type for Version Negotiation and use Version 0 (#963, #968)
• Streams are split into unidirectional and bidirectional (#643, #656, #720, #872, #175, #885)
  • Stream limits now have separate uni- and bi-directional transport parameters (#909, #958)
  • Stream limit transport parameters are now optional and default to 0 (#970, #971)
• The stream state machine has been split into read and write (#634, #894)
• Employ variable-length integer encodings throughout (#595)
• Improvements to connection close
  • Added distinct closing and draining states (#899, #871)
  • Draining period can terminate early (#869, #870)
  • Clarifications about stateless reset (#889, #890)
• Address validation for connection migration (#161, #732, #878)
• Clearly defined retransmission rules for BLOCKED (#452, #65, #924)
• negotiated_version is sent in server transport parameters (#710, #959)
• Increased the range over which packet numbers are randomized (#864, #850, #964)

• Replaced FNV-1a with AES-GCM for all “Cleartext” packets (#554)
• Split error code space between application and transport (#485)
• Stateless reset token moved to end (#820)
• 1-RTT-protected long header types removed (#848)
• No acknowledgments during draining period (#852)
• Remove “application close” as a separate close type (#854)
• Remove timestamps from the ACK frame (#841)
• Require transport parameters to only appear once (#792)

• Stateless token is server-only (#726)
• Refactor section on connection termination (#733, #748, #328, #177)
• Limit size of Version Negotiation packet (#585)
• Clarify when and what to ack (#736)
• Renamed STREAM_ID_NEEDED to STREAM_ID_BLOCKED
• Clarify Keep-alive requirements (#729)

• Introduce STOP_SENDING frame, RST_STREAM only resets in one direction (#165)
• Removed GOAWAY; application protocols are responsible for graceful shutdown (#696)
• Reduced the number of error codes (#96, #177, #184, #211)
• Version validation fields can’t move or change (#121)
• Removed versions from the transport parameters in a NewSessionTicket message (#547)
• Clarify the meaning of “bytes in flight” (#550)
• Public reset is now stateless reset and not visible to the path (#215)
• Reordered bits and fields in STREAM frame (#620)
• Clarifications to the stream state machine (#572, #571)
• Increased the maximum length of the Largest Acknowledged field in ACK frames to 64 bits (#629)
• truncate_connection_id is renamed to omit_connection_id (#659)
• CONNECTION_CLOSE terminates the connection like TCP RST (#330, #328)
• Update labels used in HKDF-Expand-Label to match TLS 1.3 (#642)

• Change STREAM and RST_STREAM layout
• Add MAX_STREAM_ID settings

• The size of the initial packet payload has a fixed minimum (#267, #472)
• Define when Version Negotiation packets are ignored (#284, #294, #241, #143, #474)
• The 64-bit FNV-1a algorithm is used for integrity protection of unprotected packets (#167, #480, #481, #517)
• Rework initial packet types to change how the connection ID is chosen (#482, #442, #493)
• No timestamps are forbidden in unprotected packets (#542, #429)
• Cryptographic handshake is now on stream 0 (#456)
• Remove congestion control exemption for cryptographic handshake (#248, #476)
• Version 1 of QUIC uses TLS; a new version is needed to use a different handshake protocol (#516)
• STREAM frames have a reduced number of offset lengths (#543, #430)
• Split some frames into separate connection- and stream- level frames (#443)
  • WINDOW_UPDATE split into MAX_DATA and MAX_STREAM_DATA (#450)
  • BLOCKED split to match WINDOW_UPDATE split (#454)
  • Define STREAM_ID_NEEDED frame (#455)
• A NEW_CONNECTION_ID frame supports connection migration without linkability (#232, #491, #496)
• Transport parameters for 0-RTT are retained from a previous connection (#405, #513, #512)
  • A client in 0-RTT no longer required to reset excess streams (#425, #479)
• Expanded security considerations (#440, #444, #445, #448)

• Defined short and long packet headers (#40, #148, #361)
• Defined a versioning scheme and stable fields (#51, #361)
• Define reserved version values for “greasing” negotiation (#112, #278)
• The initial packet number is randomized (#35, #283)
• Narrow the packet number encoding range requirement (#67, #286, #299, #323, #356)
• Defined client address validation (#52, #118, #120, #275)
• Define transport parameters as a TLS extension (#49, #122)
• SCUP and COPT parameters are no longer valid (#116, #117)
• Transport parameters for 0-RTT are either remembered from before, or assume default values (#126)
• The server chooses connection IDs in its final flight (#119, #349, #361)
• The server echoes the Connection ID and packet number fields when sending a Version Negotiation packet (#133, #295, #244)
• Defined a minimum packet size for the initial handshake packet from the client (#69, #136, #139, #164)
• Path MTU Discovery (#64, #106)
• The initial handshake packet from the client needs to fit in a single packet (#338)
• Forbid acknowledgment of packets containing only ACK and PADDING (#291)
• Require that frames are processed when packets are acknowledged (#381, #341)
• Removed the STOP_WAITING frame (#66)
• Don’t require retransmission of old timestamps for lost ACK frames (#308)
• Clarified that frames are not retransmitted, but the information in them can be (#157, #298)
• Error handling definitions (#335)
• Split error codes into four sections (#74)
• Forbid the use of Public Reset where CONNECTION_CLOSE is possible (#289)
• Define packet protection rules (#336)
• Require that stream be entirely delivered or reset, including acknowledgment of all STREAM frames or the RST_STREAM, before it closes (#381)
• Remove stream reservation from state machine (#174, #280)
• Only stream 1 does not contribute to connection-level flow control (#204)
• Stream 1 counts towards the maximum concurrent stream limit (#201, #282)
• Remove connection-level flow control exclusion for some streams (except 1) (#246)
• RST_STREAM affects connection-level flow control (#162, #163)
• Flow control accounting uses the maximum data offset on each stream, rather than bytes received (#378)
• Moved length-determining fields to the start of STREAM and ACK (#168, #277)
• Added the ability to pad between frames (#158, #276)
• Remove error code and reason phrase from GOAWAY (#352, #355)
• GOAWAY includes a final stream number for both directions (#347)
• Error codes for RST_STREAM and CONNECTION_CLOSE are now at a consistent offset (#249)
• Defined priority as the responsibility of the application protocol (#104, #303)

• Replaced DIVERSIFICATION_NONCE flag with KEY_PHASE flag
• Defined versioning
• Reworked description of packet and frame layout
• Error code space is divided into regions for each component
• Use big endian for all numeric values

• Adopted as base for draft-ietf-quic-tls
• Updated authors/editors list
• Added IANA Considerations section
• Moved Contributors and Acknowledgments to appendices