]> Google

martin.h.duke@gmail.com

Web and Internet Transport Transport and Services Working Group udp ecn Explicit Congestion Notification (ECN) applies to all transport protocols in principle. However, it had limited deployment for UDP until QUIC became widely adopted. As a result, documentation of UDP socket APIs for ECN on various platforms is sparse. This document records the results of experimenting with these APIs in order to get ECN working on UDP for Chromium on Apple, Linux, and Windows platforms. This document provides a snapshot of ECN state of affairs as supported by these platforms at the time of writing. Readers should refer to the documentations of the various platforms for an up-to-date information on the matter. About This Document The latest revision of this draft can be found at . Status information for this document may be found at . Discussion of this document takes place on the Transport and Services Working Group Working Group mailing list (), which is archived at . Subscribe at . Source for this draft and an issue tracker can be found at .

Introduction defines a two-bit field in the IP header for Explicit Congestion Notification (ECN), which provides network feedback to endpoint congestion controllers. This has historically mostly been relevant to TCP (), where any incoming ECN codepoints are internally consumed by the kernel, and therefore imply no application interface except enabling and disabling the capability. The Stream Control Transport Protocol (SCTP) () has long supported ECN in its design. SCTP is sometimes carried over DTLS and UDP (). In principle, user-space implementers might have leveraged UDP ECN APIs to deliver ECN codepoints between SCTP and the UDP socket. At the time of publication, the IETF is not aware of any such efforts. defines ECN over RTP over UDP. The IETF is aware of a research implementation, but cannot confirm any commercial deployments. QUIC runs over UDP and has seen wider deployment than SCTP. The Low Latency, Low Loss, Scalable Throughput (L4S) architecture () and QUIC have combined to increase interest in ECN over UDP. The Chromium Projects () provide a widely-deployed protocol library that includes QUIC. An effort to provide ECN support for QUIC on the many platforms on which Chromium is deployed revealed that many ECN-related UDP socket interfaces are poorly documented. This informational document provides a record of that experience, to encourage further support for ECN in other QUIC implementations, and indeed any consumer of ECN codepoints that operates over UDP. It is not a standards-track document and does not bind platforms to any API, or suggest any such API. Many socket APIs continue to reference the "ToS (Type of Service) byte", including the IP_TOS label, even though obsoleted that in 1998. That 8-bit field now contains a 6-bit Differentiated Services Code Point (DSCP) and the 2-bit ECN field. This document focuses on the APIs for the C and C++ languages. Other languages are likely to have different syntax and capabilities. The "ecn-examples" code repository () is extremely compact code that can verify the information in this document. This document addresses access to the ECN field in the IP header via socket APIs. It does not address UDP transport-layer options , which are a separate extension mechanism operating at a different layer.

Conventions and Definitions This document is not a general tutorial on UDP socket programming, and assumes familiarity with basic socket concepts like binding, socket options, and common system error codes. Throughout this document, "Apple" refers to both macOS and iOS. "Linux" guidance also applies to Android.

Receiving ECN codepoints Network devices can change the ECN codepoint in the IP header. Since this feedback is required at the packet sender, the packet receiver needs to extract this codepoint from the UDP socket in order to report to the sender. There are two components to this: setting the socket to report incoming ECN marks, and retrieving the ECN codepoint for each incoming packet. Note that Apple platforms additionally provide a framework for network connections that allows receiving ECN flags when using UDP without traditional socket option semantics. When sending or receiving UDP datagrams, IP protocol metadata carries ECN information in both directions. See .

Setting the socket to report incoming ECN codepoints

Linux, Apple, and FreeBSD To receive ECN codepoints, applications set a socket option to true using a setsockopt() call. On all platforms, IPv4 sockets require the IPPROTO_IP-level socket option with name IP_RECVTOS to be set. On all platforms, IPv6 sockets require the IPPROTO_IPV6-level socket option with name IPV6_RECVTCLASS to be set. If the IPv6 socket is not IPv6 only, on Linux hosts it is required to also set the IPPROTO_IP-level socket option IP_RECVTOS to receive ECN codepoints for UDP/IPv4 packets. This step is not necessary for Apple and FreeBSD. At the time of writing, an example implementation can be found at .

Windows Windows documentation recommends using the function WSASetRecvIPEcn() to enable ECN codepoint reporting regardless of the IP version. This function dates to Windows 10 Build 20348, according to . However, this can also be accomplished by calling setsockopt() and using options of level IPPROTO_IP and name IP_RECVECN for IPv4, and IPPROTO_IPV6 and IPV6_RECVECN for IPv6. These options are documented at . For IPv6 sockets that are not IPv6 only, WSASetRecvIPEcn() will not enable ECN reporting for IPv4. This requires a separate setsockopt() call using the IP_RECVECN option. If a socket is bound to a IPv4-mapped IPv6 address (i.e. it is of the format ::ffff:<IPv4 address>, see ) calls to WSASetRecvIpEcn() return error EINVAL. These sockets should instead use an explicit setsockopt() call to set IP_RECVECN. At the time of writing, an example implementation can be found at .

Retrieving ECN codepoints on incoming packets All platforms described in this document require the use of a recvmsg() call to read data from the socket to retrieve the ECN codepoint, because that information is provided as ancillary data. Those platforms all return zero or more "cmsg"s that contain requested information about the arriving packet. Examples of the technique described below can be found at and .

Linux If a UDP/IPv4 message is received, Linux will include a cmsg of level IPPROTO_IP and type IP_TOS. The cmsg data contains an unsigned char. This applies to IPv4 sockets and IPv6 socket, which are not IPv6 only. If a UDP/IPv6 message is received, Linux will include a cmsg of level IPPROTO_IPV6 and type IPV6_TCLASS. The cmsg data contains an int. This applies to IPv6 sockets. The cmsg data contains the entire IP header byte, which includes the DSCP and the ECN codepoint. The ECN codepoint constitutes the two least-significant bits of this byte. The same applies to the Linux-specific recvmmsg() call.

Apple and FreeBSD If a UDP/IPv4 message is received on an IPv4 socket, the ancillary data will contain a cmsg of level IPPROTO_IP and type IP_RECVTOS. The cmsg data contains an unsigned char. If a UDP/IPv6 or UDP/IPv4 message is received on an IPv6 socket, the ancillary data will contain a cmsg of level IPPROTO_IPV6 and type IPV6_TCLASS. The cmsg data contains an int. The cmsg data contains the entire IP header byte, which includes the DSCP and the ECN codepoint. The ECN codepoint constitutes the two least-significant bits of this byte.

Windows If the incoming packet is UDP/IPv4, the socket will include a cmsg of level IPPROTO_IP and type IP_ECN. The cmsg data contains an int. If the incoming packet is UDP/IPv6, the socket will include a cmsg of level IPPROTO_IPV6 and type IPV6_ECN. The cmsg data contains an int. The cmsg data solely consists of the ECN codepoint, and requires no further bitwise operations.

Sending ECN codepoints Existing ECN specifications (, ) envision a particular connection consistently sending the same ECN codepoint. It might transition that marking after successfully completing a handshake, recognizing the path or the peer do not support ECN, or transitioning to a new path. Therefore, using a socket option to configure a consistent marking is generally more resource-efficient. However, some server designs receive all incoming packets on a single socket. As the many connections that constitute this packet stream may have different support for ECN, it is suitable to provide the ECN codepoint on a per-packet basis. Note that Apple platforms additionally provide a framework for network connections that allows sending ECN flags when using UDP without traditional socket option semantics. When sending or receiving UDP datagrams, IP protocol metadata carries ECN information in both directions. See .

On a per-socket basis

Apple, FreeBSD, and Linux For sending UDP/IPv4 packets on an IPv4 socket, Apple, FreeBSD, and Linux platforms allow the outgoing ECN codepoint to be configured by using the IPPROTO_IP-level socket option with name IP_TOS. The value has the type int. For sending UDP/IPv6 packets on an IPv6 socket, Apple, FreeBSD, and Linux platforms allow the outgoing ECN codepoint to be configured by using the IPPROTO_IPV6-level socket option with name IPV6_TCLASS. The value has the type int. For sending UDP/IPv4 packets on an IPv6 socket, Linux platforms allow the outgoing ECN codepoint to be configured by using the IPPROTO_IP-level socket option with name IP_TOS. For sending UDP/IPv4 packets on an IPv6 socket, Apple and FreeBSD platforms allow the outgoing ECN codepoint to be configured by using the IPPROTO_IPV6-level socket option with name IPV6_TCLASS. Except for Apple platforms, this setsockopt() call also sets the Differentiated Services Code Point (DSCP) that make up the rest of the header byte. Applications making this call ought to preserve any existing DSCP setting, which might require an additional getsockopt() call, to avoid overriding commands set by other code in the stack. If there are multiple threads making changes to this byte, further safeguards will be necessary. An example of the technique described above can be found at .

Windows At the time of this writing, Windows does not provide a way to configure marking on a per-socket basis.

On a per-packet basis Packets can be individually marked with ECN codepoints using the ancillary data that accompanies a sendmsg() call.

Apple, FreeBSD, and Linux For sending UDP/IPv4 packets on an IPv4 socket, Apple, FreeBSD, and Linux use a cmsg with level IPPROTO_IP and type IP_TOS. On Apple and Linux the type of data is int and for FreeBSD it is unsigned char. For sending UDP/IPv6 packets on an IPv6 socket, Apple, FreeBSD, and Linux use a cmsg with level IPPROTO_IPV6 and type IPV6_TCLASS. The type of the data is int. For sending UDP/IPv4 packets on an IPv6 socket, Linux requires a cmsg with level IPPROTO_IP and type IP_TOS. Apple and FreeBSD accept a cmsg with level IPPROTO_IPV6 and type IPV6_TCLASS. The same applies to the Linux-specific sendmmsg() call.

Windows Windows uses a cmsg with level IPPROTO_IP and type IP_ECN for IPv4 packets. Windows uses a cmsg with level IPPROTO_IPV6 and type IPV6_ECN for IPv6 packets. An example of the technique described above can be found at .

Security Considerations The security implications of ECN are documented in and . This document is a guide to enabling these capabilities, which incurs no additional security considerations. Note that implementing ECN capabilities on some platforms, but not others, can help peers identify the operating system in use by a host, which can have privacy implications. This document aims to mitigate that possibility.

IANA Considerations This document has no IANA actions.

References Normative References This memo specifies the incorporation of ECN (Explicit Congestion Notification) to TCP and IP, including ECN's use of two bits in the IP header. [STANDARDS-TRACK] Informative References n.d. n.d. n.d. n.d. n.d. n.d. n.d. This document specifies the Transmission Control Protocol (TCP). TCP is an important transport-layer protocol in the Internet protocol stack, and it has continuously evolved over decades of use and growth of the Internet. Over this time, a number of changes have been made to TCP as it was specified in RFC 793, though these have only been documented in a piecemeal fashion. This document collects and brings those changes together with the protocol specification from RFC 793. This document obsoletes RFC 793, as well as RFCs 879, 2873, 6093, 6429, 6528, and 6691 that updated parts of RFC 793. It updates RFCs 1011 and 1122, and it should be considered as a replacement for the portions of those documents dealing with TCP requirements. It also updates RFC 5961 by adding a small clarification in reset handling while in the SYN-RECEIVED state. The TCP header control bits from RFC 793 have also been updated based on RFC 3168. This document describes the Stream Control Transmission Protocol (SCTP) and obsoletes RFC 4960. It incorporates the specification of the chunk flags registry from RFC 6096 and the specification of the I bit of DATA chunks from RFC 7053. Therefore, RFCs 6096 and 7053 are also obsoleted by this document. In addition, RFCs 4460 and 8540, which describe errata for SCTP, are obsoleted by this document. SCTP was originally designed to transport Public Switched Telephone Network (PSTN) signaling messages over IP networks. It is also suited to be used for other applications, for example, WebRTC. SCTP is a reliable transport protocol operating on top of a connectionless packet network, such as IP. It offers the following services to its users: The design of SCTP includes appropriate congestion avoidance behavior and resistance to flooding and masquerade attacks. The Stream Control Transmission Protocol (SCTP) is a transport protocol originally defined to run on top of the network protocols IPv4 or IPv6. This document specifies how SCTP can be used on top of the Datagram Transport Layer Security (DTLS) protocol. Using the encapsulation method described in this document, SCTP is unaware of the protocols being used below DTLS; hence, explicit IP addresses cannot be used in the SCTP control chunks. As a consequence, the SCTP associations carried over DTLS can only be single-homed. This memo specifies how Explicit Congestion Notification (ECN) can be used with the Real-time Transport Protocol (RTP) running over UDP, using the RTP Control Protocol (RTCP) as a feedback mechanism. It defines a new RTCP Extended Report (XR) block for periodic ECN feedback, a new RTCP transport feedback message for timely reporting of congestion events, and a Session Traversal Utilities for NAT (STUN) extension used in the optional initialisation method using Interactive Connectivity Establishment (ICE). Signalling and procedures for negotiation of capabilities and initialisation methods are also defined. [STANDARDS-TRACK] This document defines the core of the QUIC transport protocol. QUIC provides applications with flow-controlled streams for structured communication, low-latency connection establishment, and network path migration. QUIC includes security measures that ensure confidentiality, integrity, and availability in a range of deployment circumstances. Accompanying documents describe the integration of TLS for key negotiation, loss detection, and an exemplary congestion control algorithm. This document describes the L4S architecture, which enables Internet applications to achieve low queuing latency, low congestion loss, and scalable throughput control. L4S is based on the insight that the root cause of queuing delay is in the capacity-seeking congestion controllers of senders, not in the queue itself. With the L4S architecture, all Internet applications could (but do not have to) transition away from congestion control algorithms that cause substantial queuing delay and instead adopt a new class of congestion controls that can seek capacity with very little queuing. These are aided by a modified form of Explicit Congestion Notification (ECN) from the network. With this new architecture, applications can have both low latency and high throughput. The architecture primarily concerns incremental deployment. It defines mechanisms that allow the new class of L4S congestion controls to coexist with 'Classic' congestion controls in a shared network. The aim is for L4S latency and throughput to be usually much better (and rarely worse) while typically not impacting Classic performance. This document defines the IP header field, called the DS (for differentiated services) field. [STANDARDS-TRACK] Transport protocols are extended through the use of transport header options. This document updates RFC 768 (UDP) by indicating the location, syntax, and semantics for UDP transport layer options within the surplus area after the end of the UDP user data but before the end of the IP datagram. This specification defines the addressing architecture of the IP Version 6 (IPv6) protocol. The document includes the IPv6 addressing model, text representations of IPv6 addresses, definition of IPv6 unicast addresses, anycast addresses, and multicast addresses, and an IPv6 node's required addresses. This document obsoletes RFC 3513, "IP Version 6 Addressing Architecture". [STANDARDS-TRACK]

Acknowledgments The author would like to thank Ryan Hamilton, who provided constant advice through this effort. Randall Meyer from Apple and Nick Grifka from Microsoft provided useful hints about the behavior of their respective operating systems. However, the author takes full responsibility for any errors above. Michael Tuexen wrote the "ecn-examples" code that was very helpful in verifying the conclusions in this document. He also made multiple editorial contributions. Neal Cardwell, Gorry Fairhurst, Max Franke, Rodney Grimes, Will Hawkins, Guillaume Hétier, Max Inden, Jonathan Lennox, Colin Perkins, Marten Seemann, and Greg White made improvements to this draft.