| Internet-Draft | Stream Namespaces for QUIC | February 2024 |
| Vasiliev | Expires 24 August 2024 | [Page] |
QUIC Stream Namespaces provide an extension to the QUIC protocol that enables multiplexing multiple logical groups of streams within the same connection, while providing flow control isolation.¶
This note is to be removed before publishing as an RFC.¶
Status information for this document may be found at https://datatracker.ietf.org/doc/draft-vvv-quic-namespaces/.¶
Discussion of this document takes place on the quic Working Group mailing list (mailto:quic@ietf.org), which is archived at https://example.com/WG. Subscribe at https://www.ietf.org/mailman/listinfo/quic/.¶
Source for this draft and an issue tracker can be found at https://github.com/https://github.com/vasilvv/draft-vvv-quic-namespaces.¶
This Internet-Draft is submitted in full conformance with the provisions of BCP 78 and BCP 79.¶
Internet-Drafts are working documents of the Internet Engineering Task Force (IETF). Note that other groups may also distribute working documents as Internet-Drafts. The list of current Internet-Drafts is at https://datatracker.ietf.org/drafts/current/.¶
Internet-Drafts are draft documents valid for a maximum of six months and may be updated, replaced, or obsoleted by other documents at any time. It is inappropriate to use Internet-Drafts as reference material or to cite them other than as "work in progress."¶
This Internet-Draft will expire on 24 August 2024.¶
Copyright (c) 2024 IETF Trust and the persons identified as the document authors. All rights reserved.¶
This document is subject to BCP 78 and the IETF Trust's Legal Provisions Relating to IETF Documents (https://trustee.ietf.org/license-info) in effect on the date of publication of this document. Please review these documents carefully, as they describe your rights and restrictions with respect to this document. Code Components extracted from this document must include Revised BSD License text as described in Section 4.e of the Trust Legal Provisions and are provided without warranty as described in the Revised BSD License.¶
QUIC [RFC9000] provides an ordered bytestream abstraction called streams. Streams are subject to various flow control mechanisms that allow a network endpoint to control how much resources a peer is allowed to consume. Some of the flow control mechanisms are scoped to a single stream; others are global to the entire connection. The connection-level flow control mechanisms are a good fit in cases when all of the streams originate from the same entity; however, in cases when multiple logical entities share the same connection, a single global limit may lead to one entity starving another. This document provides a mechanism by which a single QUIC connection can have multiple namespaces, each with its own resource limits for streams.¶
The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in BCP 14 [RFC2119] [RFC8174] when, and only when, they appear in all capitals, as shown here.¶
A QUIC namespace is a 62-bit unique ID number. In the initial state, every namespace ID is assumed to exist, but have a MAX_STREAMS number associated with it set to 0 for all types of streams, and a MAX_DATA value of 0 in both directions. A peer opens a namespace by sending a combination of MAX_DATA and MAX_STREAMS frames for that namespace. The recepient may response with either its own MAX_DATA and MAX_STREAMS, confirming the response, or it may close the namespace. Frames that do not have a namespace ID associated with them are said to be a part of the default namespace.¶
Note that there is no way to set a namespace-specific initial_max_stream_data parameters; those remain connection-global.¶
An NS frame (frame type=0x29c5) is a frame that alters the meaning of the frame that comes immediately after it. If the subsequent frame has a stream ID in it, that ID refers to the stream with the corresponding ID in the specified namespace. If the subsequent frame alters connection-global flow control limits, those limits are altered for the namespace in question, instead of the default namespace.¶
NS Frame {
Type (i) = 0x29c5,
Namespace ID (i),
}
The following frames are allowed to follow the NS frame: STREAM, RESET_STREAM, STOP_SENDING, MAX_DATA, MAX_STREAM_DATA, MAX_STREAMS, DATA_BLOCKED, STREAM_DATA_BLOCKED, STREAMS_BLOCKED. Extensions that define their own frames can define their own semantics of interacting with namespaces. If a frame that is not listed above and does not have extension semantics defined for it is prefixed with an NS frame, the recepient MUST close the connection with a PROTOCOL_VIOLATION error code. Same applies to an NS frame that is not followed by anything.¶
Note that this intentionally does not define NS prefix for the DATAGRAM frames [RFC9221], as datagrams already have pre-defined mechanisms for multiplexing (such as [RFC9297]) that may conflict with QUIC stream namespaces, and there is no technical advantage of using an NS frame with datagrams over doing multiplexing within the datagram payload.¶
A CLOSE_NAMESPACE frame indicates to the peer that the sender will not process any further data received for a given namespace. The sender can discard all of the state related to the namespace after sending this frame.¶
CLOSE_NAMESPACE Frame {
Type (i) = 0x29c6,
Namespace ID (i),
}
TODO Security¶
TODO: discuss the issue where the peer has to remember flow control limits for arbitrary unexpected namespaces.¶
TODO: add a transport parameter to negotiate this feature.¶
TODO acknowledge.¶