Implementing Interfaces to Transport Services

Implementing Interfaces to Transport Services Karlstad University

Universitetsgatan 2 651 88 Karlstad Sweden anna.brunstrom@kau.se

Apple Inc.

One Apple Park Way Cupertino, California 95014 United States of America tpauly@apple.com

Netflix

121 Albright Way Los Gatos, CA 95032 United States of America ietf@tenghardt.net

SAP SE

Konrad-Zuse-Ring 10 14469 Potsdam Germany philipp@tiesel.net

University of Oslo

PO Box 1080 Blindern 0316 Oslo Norway michawe@ifi.uio.no

Transport TAPS Working Group Internet-Draft The Transport Services system enables applications to use transport protocols flexibly for network communication and defines a protocol-independent Transport Services Application Programming Interface (API) that is based on an asynchronous, event-driven interaction pattern. This document serves as a guide to implementing such a system.

Introduction The Transport Services architecture defines a system that allows applications to flexibly use transport networking protocols. The API that such a system exposes to applications is defined as the Transport Services API . This API is designed to be generic across multiple transport protocols and sets of protocol features. This document serves as a guide to implementing a system that provides a Transport Services API. This guide offers suggestions to developers, but it is not prescriptive: implementations are free to take any desired form as long as the API specification in is honored. It is the job of an implementation of a Transport Services system to turn the requests of an application into decisions on how to establish connections, and how to transfer data over those connections once established. The terminology used in this document is based on the Transport Services architecture .

Implementing Connection Objects The connection objects that are exposed to applications for Transport Services are:

the Preconnection, the bundle of properties that describes the application constraints on, and preferences for, the transport;
the Connection, the basic object that represents a flow of data as Messages in either direction between the Local and Remote Endpoints;
and the Listener, a passive waiting object that delivers new Connections.

Preconnection objects should be implemented as bundles of properties that an application can both read and write. A Preconnection object influences a Connection only at one point in time: when the Connection is created. Connection objects represent the interface between the application and the implementation to manage transport state, and conduct data transfer. During the process of establishment (), the Connection will not necessarily be immediately bound to a transport protocol instance, since multiple candidate Protocol Stacks might be raced. Once a Preconnection has been used to create an outbound Connection or a Listener, the implementation should ensure that the copy of the properties held by the Connection or Listener cannot be mutated by the application making changes to the original Preconnection object. This may involve the implementation performing a deep-copy, copying the object with all the objects that it references. Once the Connection is established, the Transport Services implementation maps actions and events to the details of the chosen Protocol Stack. For example, the same Connection object may ultimately represent a single transport protocol instance (e.g., a TCP connection, a TLS session over TCP, a UDP flow with fully-specified Local and Remote Endpoints, a DTLS session, a SCTP stream, a QUIC stream, or an HTTP/2 stream). The Connection Properties held by a Connection or Listener are independent of other Connections that are not part of the same Connection Group. Connection establishment is only a local operation for a connectionless protocols, which serves to simplify the local send/receive functions and to filter the traffic for the specified addresses and ports (for example using UDP or UDP-Lite transport without a connection handshake procedure). Once Initiate has been called, the Selection Properties and Endpoint information of the created Connection are immutable (i.e, an application is not able to later modify the properties of a Connection by manipulating the original Preconnection object). Listener objects are created with a Preconnection, at which point their configuration should be considered immutable by the implementation. The process of listening is described in .

Implementing Pre-Establishment The pre-establishment phase allows applications to specify properties for the Connections that they are about to make, or to query the API about potential Connections they could make. During pre-establishment the application specifies one or more Endpoints to be used for communication as well as protocol preferences and constraints via Selection Properties and, if desired, also Connection Properties. Section 4 of states that Connection Properties should preferably be configured during pre-establishment, because they can serve as input to decisions that are made by the implementation (e.g., the capacity profile can guide usage of a protocol offering scavenger-type congestion control). The implementation stores these properties as a part of the Preconnection object for use during connection establishment. For Selection Properties that are not provided by the application, the implementation uses the default values specified in the Transport Services API ().

Configuration-time errors The Transport Services system should have a list of supported protocols available, which each have transport features reflecting the capabilities of the protocol. Once an application specifies its Transport Properties, the Transport Services system matches the required and prohibited properties against the transport features of the available protocols (see Section 6.2 of for the definition of property preferences). In the following cases, failure should be detected during pre-establishment:

A request by an application for properties that cannot be satisfied by any of the available protocols. For example, if an application requires perMsgReliability, but no such feature is available in any protocol on the host running the Transport Services system this should result in an error.
A request by an application for properties that are in conflict with each other, such as specifying required and prohibited properties that cannot be satisfied by any protocol. For example, if an application prohibits reliability but then requires perMsgReliability, this mismatch should result in an error.

To avoid allocating resources that are not finally needed, it is important that configuration-time errors fail as early as possible.

Role of system policy The properties specified during pre-establishment have a close relationship to system policy. The implementation is responsible for combining and reconciling several different sources of preferences when establishing Connections. These include, but are not limited to:

Application preferences, i.e., preferences specified during the pre-establishment via Selection Properties.
Dynamic system policy, i.e., policy compiled from internally and externally acquired information about available network interfaces, supported transport protocols, and current/previous Connections. Examples of ways to externally retrieve policy-support information are through OS-specific statistics/measurement tools and tools that reside on middleboxes and routers.
Default implementation policy, i.e., predefined policy by OS or application.

In general, any protocol or path used for a Connection must conform to all three sources of constraints. A violation that occurs at any of the policy layers should cause a protocol or path to be considered ineligible for use. If such a violation prevents a Connection from being established, this should be communicated to the application, e.g. via the EstablishmentError event. For an example of application preferences leading to constraints, an application may prohibit the use of metered network interfaces for a given Connection to avoid user cost. Similarly, the system policy at a given time may prohibit the use of such a metered network interface from the application's process. Lastly, the implementation itself may default to disallowing certain network interfaces unless explicitly requested by the application. It is expected that the database of system policies and the method of looking up these policies will vary across various platforms. An implementation should attempt to look up the relevant policies for the system in a dynamic way to make sure it is reflecting an accurate version of the system policy, since the system's policy regarding the application's traffic may change over time due to user or administrative changes.

Implementing Connection Establishment The process of establishing a network connection begins when an application expresses intent to communicate with a Remote Endpoint by calling Initiate, at which point the Preconnection object contains all constraints or requirements the application has configured. The establishment process can be considered complete once there is at least one Protocol Stack that has completed any required setup to the point that it can transmit and receive the application's data. Connection establishment is divided into two top-level steps: Candidate Gathering (defined in Section 4.2.1 of ), to identify the paths, protocols, and endpoints to use (see ); and Candidate Racing (defined in Section 4.2.2 of ), in which the necessary protocol handshakes are conducted so that the Transport Services system can select which set to use (see ). Candidate Racing involves attempting multiple options for connection establishment, and choosing the first option to succeed as the Protocol Stack to use for the connection. These attempts are usually staggered, starting each next option after a delay, but they can also be performed in parallel or only after waiting for failures. For ease of illustration, this document structures the candidates for racing as a tree (see ). This is not meant to restrict implementations from structuring racing candidates differently. The most simple example of this process might involve identifying the single IP address to which the implementation wishes to connect, using the system's current default path (i.e., using the default interface), and starting a TCP handshake to establish a stream to the specified IP address. However, each step may also differ depending on the requirements of the connection: if the endpoint is defined as a hostname and port, then there may be multiple resolved addresses that are available; there may also be multiple paths available, (in this case using an interface other than the default system interface); and some protocols may not need any transport handshake to be considered "established" (such as UDP), while other connections may utilize layered protocol handshakes, such as TLS over TCP. Whenever an implementation has multiple options for connection establishment, it can view the set of all individual connection establishment options as a single, aggregate connection establishment. The aggregate set conceptually includes every valid combination of endpoints, paths, and protocols. As an example, consider an implementation that initiates a TCP connection to a hostname + port endpoint, and has two valid interfaces available (Wi-Fi and LTE). The hostname resolves to a single IPv4 address on the Wi-Fi network, and resolves to the same IPv4 address on the LTE network, as well as a single IPv6 address. The aggregate set of connection establishment options can be viewed as follows: [Endpoint: 192.0.2.1:80] [Interface: Wi-Fi] [Protocol: TCP] |-> [Endpoint: 192.0.2.1:80] [Interface: LTE] [Protocol: TCP] |-> [Endpoint: 2001:DB8::1.80] [Interface: LTE] [Protocol: TCP] ]]> Any one of these sub-entries on the aggregate connection attempt would satisfy the original application intent. The concern of this section is the algorithm defining which of these options to try, when, and in what order. During Candidate Gathering (), an implementation prunes and sorts branches according to the Selection Property preferences (Section 6.2 of . It first excludes all protocols and paths that match a Prohibit property or do not match all Require properties. Then it will sort branches according to Preferred properties, Avoided properties, and possibly other criteria.

Structuring Candidates as a Tree As noted above, the consideration of multiple candidates in a gathering and racing process can be conceptually structured as a tree; this terminological convention is used throughout this document. Each leaf node of the tree represents a single, coherent connection attempt, with an endpoint, a network path, and a set of protocols that can directly negotiate and send data on the network. Each node in the tree that is not a leaf represents a connection attempt that is either underspecified, or else includes multiple distinct options. For example, when connecting on an IP network, a connection attempt to a hostname and port is underspecified, because the connection attempt requires a resolved IP address as its Remote Endpoint. In this case, the node represented by the connection attempt to the hostname is a parent node, with child nodes for each IP address. Similarly, an implementation that is allowed to connect using multiple interfaces will have a parent node of the tree for the decision between the network paths, with a branch for each interface. The example aggregate connection attempt above can be drawn as a tree by grouping the addresses resolved on the same interface into branches: The rest of this section will use a notation scheme to represent this tree. The root node (or parent node) of the tree will be represented by a single integer, such as "1". ("1" is used assuming that this is the first connection made by the system; future connections created by the application would allocate numbers in an increasing manner.) Each child of that node will have an integer that identifies it, from 1 to the number of children. That child node will be uniquely identified by concatenating its integer to its parent's identifier with a dot in between, such as "1.1" and "1.2". Each node will be summarized by a tuple of three elements: endpoint, path (labeled here by interface), and protocol. In Protocol Stacks, the layers are separated by '/' and ordered with the protocol closest to the application first. The above example can now be written more succinctly as: When an implementation is asked to establish a single connection, only one of the leaf nodes in the candidate set is needed to transfer data. Thus, once a single leaf node becomes ready to use, then the connection establishment tree is considered ready. One way to implement this is by having every leaf node update the state of its parent node when it becomes ready, until the root node of the tree is ready, which then notifies the application that the Connection as a whole is ready to use. A connection establishment tree may consist of only a single node, such as a connection attempt to an IP address over a single interface with a single protocol. A root node may also only have one child (or leaf) node, such as a when a hostname resolves to only a single IP address.

Branch Types There are three types of branching from a parent node into one or more child nodes. Any parent node of the tree must only use one type of branching.

Derived Endpoints If a connection originally targets a single endpoint, there may be multiple endpoints of different types that can be derived from the original. This creates an ordered list of the derived endpoints according to application preference, system policy and expected performance. DNS hostname-to-address resolution is the most common method of endpoint derivation. When trying to connect to a hostname endpoint on a traditional IP network, the implementation should send all applicable DNS queries. Commonly, this will include both A (IPv4) and AAAA (IPv6) records if both address families are supported on the local interface. This can also include SRV records , SVCB and HTTPS records , or other future record types. The algorithm for ordering and racing these addresses should follow the recommendations in Happy Eyeballs . DNS-Based Service Discovery can also provide an endpoint derivation step. When trying to connect to a named service, the client may discover one or more hostname and port pairs on the local network using multicast DNS . These hostnames should each be treated as a branch that can be attempted independently from other hostnames. Each of these hostnames might resolve to one or more addresses, which would create multiple layers of branching. Applications can influence which derived endpoints are allowed and preferred via Selection Properties set on the Preconnection. For example, setting a preference for useTemporaryLocalAddress would prefer the use of IPv6 over IPv4, and requiring useTemporaryLocalAddress would eliminate IPv4 options, since IPv4 does not support temporary addresses.

Network Paths If a client has multiple network paths available to it, e.g., a mobile client with interfaces for both Wi-Fi and Cellular connectivity, it can attempt a connection over any of the paths. This represents a branch point in the connection establishment. Similar to a derived endpoint, the paths should be ranked based on preference, system policy, and performance. Attempts should be started on one path (e.g., a specific interface), and then successively on other paths (or interfaces) after delays based on the expected path round-trip-time or other available metrics. The same approach applies to any situation in which the client is aware of multiple links or views of the network. A single interface may be shared by multiple network paths, each with a coherent set of addresses, routes, DNS server, and more. A path may also represent a virtual interface service such as a Virtual Private Network (VPN). The list of available paths should be constrained by any requirements the application sets, as well as by the system policy.

Protocol Options Differences in possible protocol compositions and options can also provide a branching point in connection establishment. This allows clients to be resilient to situations in which a certain protocol is not functioning on a server or network. This approach is commonly used for connections with optional proxy server configurations. A single connection might have several options available: an HTTP-based proxy, a SOCKS-based proxy, or no proxy. As above, these options should be ranked based on preference, system policy, and performance and attempted in succession. This approach also allows a client to attempt different sets of application and transport protocols that, when available, could provide preferable features. For example, the protocol options could involve QUIC over UDP on one branch, and HTTP/2 over TLS over TCP on the other: Another example is racing SCTP with TCP: Implementations that support racing protocols and protocol options should maintain a history of which protocols and protocol options were successfully established, on a per-network and per-endpoint basis (see ). This information can influence future racing decisions to prioritize or prune branches.

Branching Order-of-Operations Branch types ought to occur in a specific order relative to one another to avoid creating leaf nodes with invalid or incompatible settings. In the example above, it would be invalid to branch for derived endpoints (the DNS results for www.example.com) before branching between interface paths, since there are situations when the results will be different across networks due to private names or different supported IP versions. Implementations need to be careful to branch in a consistent order that results in usable leaf nodes whenever there are multiple branch types that could be used from a single node. This document recommends the following order of operations for branching:

Network Paths
Protocol Options
Derived Endpoints

where a lower number indicates higher precedence and therefore higher placement in the tree. Branching between paths is the first in the list because results across multiple interfaces are likely not related to one another: endpoint resolution may return different results, especially when using locally resolved host and service names, and which protocols are supported and preferred may differ across interfaces. Thus, if multiple paths are attempted, the overall connection establishment process can be seen as a race between the available paths or interfaces. Protocol options are next checked in order. Whether or not a set of protocols, or protocol-specific options, can successfully connect is generally not dependent on which specific IP address is used. Furthermore, the Protocol Stacks being attempted may influence or altogether change the endpoints being used. Adding a proxy to a connection's branch will change the endpoint to the proxy's IP address or hostname. Choosing an alternate protocol may also modify the ports that should be selected. Branching for derived endpoints is the final step, and may have multiple layers of derivation or resolution, such as DNS service resolution and DNS hostname resolution. For example, if the application has indicated both a preference for WiFi over LTE and for a feature only available in SCTP, branches will be first sorted accord to path selection, with WiFi attempted first. Then, branches with SCTP will be attempted first within their subtree according to the properties influencing protocol selection. However, if the implementation has current cache information that SCTP is not available on the path over WiFi, there would be no SCTP node in the WiFi subtree. Here, the path over WiFi will be attempted first, and, if connection establishment succeeds, TCP will be used. Thus, the Selection Property preferring WiFi takes precedence over the Property that led to a preference for SCTP.

Sorting Branches Implementations should sort the branches of the tree of connection options in order of their preference rank, from most preferred to least preferred as specified by Selection Properties . Leaf nodes on branches with higher rankings represent connection attempts that will be raced first. In addition to the properties provided by the application, an implementation may include additional criteria such as cached performance estimates, see , or system policy, see , in the ranking. Two examples of how Selection and Connection Properties may be used to sort branches are provided below:

"Interface Instance or Type" (property name interface): If the application specifies an interface type to be preferred or avoided, implementations should accordingly rank the paths. If the application specifies an interface type to be required or prohibited, an implementation is expected to exclude the non-conforming paths.
"Capacity Profile" (property name connCapacityProfile): An implementation can use the capacity profile to prefer paths that match an application's expected traffic profile. This match will use cached performance estimates, see . Some examples of path preferences based on capacity profiles include:
- Low Latency/Interactive: Prefer paths with the lowest expected Round Trip Time, based on observed Round Trip Time estimates;
- Low Latency/Non-Interactive: Prefer paths with a low expected Round Trip Time, but can tolerate delay variation;
- Constant-Rate Streaming: Prefer paths that are expected to satisfy the requested stream send or receive bitrate, based on the observed maximum throughput;
- Capacity-Seeking: Prefer adapting to paths to determine the highest available capacity, based on the observed maximum throughput.

As another example, branch sorting can also be influenced by bounds on the send or receive rate (Selection Properties minSendRate / minRecvRate / maxSendRate / maxRecvRate): if the application indicates a bound on the expected send or receive bitrate, an implementation may prefer a path that can likely provide the desired bandwidth, based on cached maximum throughput, see . The application may know the send or receive bitrate from metadata in adaptive HTTP streaming, such as MPEG-DASH. Implementations process the Properties (Section 6.2 of ) in the following order: Prohibit, Require, Prefer, Avoid. If Selection Properties contain any prohibited properties, the implementation should first purge branches containing nodes with these properties. For required properties, it should only keep branches that satisfy these requirements. Finally, it should order the branches according to the preferred properties, and finally use any avoided properties as a tiebreaker. When ordering branches, an implementation can give more weight to properties that the application has explicitly set, than to the properties that are default. The available protocols and paths on a specific system and in a specific context can change; therefore, the result of sorting and the outcome of racing may vary, even when using the same Selection and Connection Properties. However, an implementation ought to provide a consistent outcome to applications, e.g., by preferring protocols and paths that are already used by existing Connections that specified similar Properties.

Candidate Gathering The step of gathering candidates involves identifying which paths, protocols, and endpoints may be used for a given Connection. This list is determined by the requirements, prohibitions, and preferences of the application as specified in the Selection Properties.

Gathering Endpoint Candidates Both Local and Remote Endpoint Candidates must be discovered during connection establishment. To support Interactive Connectivity Establishment (ICE) , or similar protocols that involve out-of-band indirect signalling to exchange candidates with the Remote Endpoint, it is important to query the set of candidate Local Endpoints, and provide the Protocol Stack with a set of candidate Remote Endpoints, before the Local Endpoint attempts to establish connections.

Local Endpoint candidates The set of possible Local Endpoints is gathered. In a simple case, this merely enumerates the local interfaces and protocols, and allocates ephemeral source ports. For example, a system that has WiFi and Ethernet and supports IPv4 and IPv6 might gather four candidate Local Endpoints (IPv4 on Ethernet, IPv6 on Ethernet, IPv4 on WiFi, and IPv6 on WiFi) that can form the source for a transient. If NAT traversal is required, the process of gathering Local Endpoints becomes broadly equivalent to the ICE Candidate Gathering phase (see Section 5.1.1 of ). The endpoint determines its server reflexive Local Endpoints (i.e., the translated address of a Local Endpoint, on the other side of a NAT, e.g via a STUN sever ) and relayed Local Endpoints (e.g., via a TURN server or other relay), for each interface and network protocol. These are added to the set of candidate Local Endpoints for this connection. Gathering Local Endpoints is primarily a local operation, although it might involve exchanges with a STUN server to derive server reflexive Local Endpoints, or with a TURN server or other relay to derive relayed Local Endpoints. However, it does not involve communication with the Remote Endpoint.

Remote Endpoint Candidates The Remote Endpoint is typically a name that needs to be resolved into a set of possible addresses that can be used for communication. Resolving the Remote Endpoint is the process of recursively performing such name lookups, until fully resolved, to return the set of candidates for the Remote Endpoint of this Connection. How this resolution is done will depend on the type of the Remote Endpoint, and can also be specific to each Local Endpoint. A common case is when the Remote Endpoint is a DNS name, in which case it is resolved to give a set of IPv4 and IPv6 addresses representing that name. Some types of Remote Endpoint might require more complex resolution. Resolving the Remote Endpoint for a peer-to-peer connection might involve communication with a rendezvous server, which in turn contacts the peer to gain consent to communicate and retrieve its set of candidate Local Endpoints, which are returned and form the candidate remote addresses for contacting that peer. Resolving the Remote Endpoint is not a local operation. It will involve a directory service, and can require communication with the Remote Endpoint to rendezvous and exchange peer addresses. This can expose some or all of the candidate Local Endpoints to the Remote Endpoint.

Candidate Racing The primary goal of the Candidate Racing process is to successfully negotiate a Protocol Stack to an endpoint over an interface to connect a single leaf node of the tree with as little delay and as few unnecessary connections attempts as possible. Optimizing these two factors improves the user experience, while minimizing network load. This section covers the dynamic aspect of connection establishment. The tree described above is a useful conceptual and architectural model. However, an implementation is unable to know all of the nodes that will be used until steps like name resolution have occurred, and many of the possible branches ultimately might not be attempted. There are three different approaches to racing the attempts for different nodes of the connection establishment tree:

Simultaneous
Staggered
Failover

Each approach is appropriate in different use-cases and branch types. However, to avoid consuming unnecessary network resources, implementations should not use simultaneous racing as a default approach. The timing algorithms for racing should remain independent across branches of the tree. Any timer or racing logic is isolated to a given parent node, and is not ordered precisely with regards to children of other nodes.

Simultaneous Simultaneous racing is when multiple alternate branches are started without waiting for any one branch to make progress before starting the next alternative. This means the attempts are effectively simultaneous. Simultaneous racing should be avoided by implementations, since it consumes extra network resources and establishes state that might not be used.

Staggered Staggered racing can be used whenever a single node of the tree has multiple child nodes. Based on the order determined when building the tree, the first child node will be initiated immediately, followed by the next child node after some delay. Once that second child node is initiated, the third child node (if present) will begin after another delay, and so on until all child nodes have been initiated, or one of the child nodes successfully completes its negotiation. Staggered racing attempts can proceed in parallel. Implementations should not terminate an earlier child connection attempt upon starting a secondary child. If a child node fails to establish connectivity (as in ) before the delay time has expired for the next child, the next child should be started immediately. Staggered racing between IP addresses for a generic Connection should follow the Happy Eyeballs algorithm described in . provides guidance for racing when performing Interactive Connectivity Establishment (ICE). Generally, the delay before starting a given child node ought to be based on the length of time the previously started child node is expected to take before it succeeds or makes progress in connection establishment. Algorithms like Happy Eyeballs choose a delay based on how long the transport connection handshake is expected to take. When performing staggered races in multiple branch types (such as racing between network interfaces, and then racing between IP addresses), a longer delay may be chosen for some branch types. For example, when racing between network interfaces, the delay should also take into account the amount of time it takes to prepare the network interface (such as radio association) and name resolution over that interface, in addition to the delay that would be added for a single transport connection handshake. Since the staggered delay can be chosen based on dynamic information, such as predicted Round Trip Time, implementations should define upper and lower bounds for delay times. These bounds are implementation-specific, and may differ based on which branch type is being used.

Failover If an implementation or application has a strong preference for one branch over another, the branching node may choose to wait until one child has failed before starting the next. Failure of a leaf node is determined by its protocol negotiation failing or timing out; failure of a parent branching node is determined by all of its children failing. An example in which failover is recommended is a race between a preferred Protocol Stack that uses a proxy and an alternate Protocol Stack that bypasses the proxy. Failover is useful in case the proxy is down or misconfigured, but any more aggressive type of racing may end up unnecessarily avoiding a proxy that was preferred by policy.

Completing Establishment The process of connection establishment completes when one leaf node of the tree has successfully completed negotiation with the Remote Endpoint, or else all nodes of the tree have failed to connect. The first leaf node to complete its connection is then used by the application to send and receive data. This is signalled to the application using the Ready event in the API (). Successes and failures of a given attempt should be reported up to parent nodes (towards the root of the tree). For example, in the following case, if 1.1.1 fails to connect, it reports the failure to 1.1. Since 1.1 has no other child nodes, it also has failed and reports that failure to 1. Because 1.2 has not yet failed, 1 is not considered to have failed. Since 1.2 has not yet started, it is started and the process continues. Similarly, if 1.1.1 successfully connects, then it marks 1.1 as connected, which propagates to the root node 1. At this point, the Connection as a whole is considered to be successfully connected and ready to process application data. If a leaf node has successfully completed its connection, all other attempts should be made ineligible for use by the application for the original request. New connection attempts that involve transmitting data on the network ought not to be started after another leaf node has already successfully completed, because the Connection as a whole has now been established. An implementation could choose to let certain handshakes and negotiations complete to gather metrics that influence future connections. Keeping additional connections is generally not recommended, because those attempts were slower to connect and may exhibit less desirable properties.

Determining Successful Establishment On a per-protocol basis, implementations may select different criteria by which a leaf node is considered to be successfully connected. If the only protocol being used is a transport protocol with a clear handshake, like TCP, then the obvious choice is to declare that node "connected" when the three-way handshake has been completed. If the only protocol being used is an connectionless protocol, like UDP, the implementation may consider the node fully "connected" the moment it determines a route is present, before sending any packets on the network, see further . When the Initiate action is called without any Messages being sent at the same time, depending on the protocols involved, it is not guaranteed that the Remote Endpoint will be notified of this, and hence a passive endpoint's application may not receive a ConnectionReceived event until it receives the first Message on the new Connection. For Protocol Stacks with multiple handshakes, the decision becomes more nuanced. If the Protocol Stack involves both TLS and TCP, an implementation could determine that a leaf node is connected after the TCP handshake is complete, or it can wait for the TLS handshake to complete as well. The benefit of declaring completion when the TCP handshake finishes, and thus stopping the race for other branches of the tree, is reduced burden on the network and Remote Endpoints from further connection attempts that are likely to be abandoned. On the other hand, by waiting until the TLS handshake is complete, an implementation avoids the scenario in which a TCP handshake completes quickly, but TLS negotiation is either very slow or fails altogether in particular network conditions or to a particular endpoint. To avoid the issue of TLS possibly failing, the implementation should not generate a Ready event for the Connection until the TLS handshake is complete. If all of the leaf nodes fail to connect during racing, i.e. none of the configurations that satisfy all requirements given in the Transport Properties actually work over the available paths, then the Transport Services system should report an EstablishmentError to the application. An EstablishmentError event should also be generated in case the Transport Services system finds no usable candidates to race.

Establishing multiplexed connections Multiplexing several Connections over a single underlying transport connection requires that the Connections to be multiplexed belong to the same Connection Group (as is indicated by the application using the Clone action). When the underlying transport connection supports multi-streaming, the Transport Services System can map each Connection in the Connection Group to a different stream of this connection. For such streams, there is often no explicit connection establishment procedure for the new stream prior to sending data on it (e.g., with SCTP). In this case, the same considerations apply to determining stream establishment as apply to establishing a UDP connection, as discussed in . This means that there might not be any "establishment" message (like a TCP SYN).

Handling connectionless protocols While protocols that use an explicit handshake to validate a connection to a peer can be used for racing multiple establishment attempts in parallel, connectionless protocols such as raw UDP do not offer a way to validate the presence of a peer or the usability of a Connection without application feedback. An implementation should consider such a Protocol Stack to be established as soon as the Transport Services system has selected a path on which to send data. However, this can cause a problem if a specific peer is not reachable over the network using the connectionless protocol, or data cannot be exchanged with the peer for any other reason. To handle the lack of an explicit handshake in the underlying protocol, an application can use a Message Framer () on top of a connectionless protocol to only mark a specific connection attempt as ready when some data has been received, or after some application-level handshake has been performed by the Message Framer.

Implementing Listeners When an implementation is asked to Listen, it registers with the system to wait for incoming traffic to the Local Endpoint. If no Local Endpoint is specified, the implementation should use an ephemeral port. If the Selection Properties do not require a single network interface or path, but allow the use of multiple paths, the Listener object should register for incoming traffic on all of the network interfaces or paths that conform to the Properties. The set of available paths can change over time, so the implementation should monitor network path changes, and change the registration of the Listener across all usable paths as appropriate. When using multiple paths, the Listener is generally expected to use the same port for listening on each. If the Selection Properties allow multiple protocols to be used for listening, and the implementation supports it, the Listener object should support receiving inbound connections for each eligible protocol on each eligible path.

Implementing Listeners for Connected Protocols Connected protocols such as TCP and TLS-over-TCP have a strong mapping between the Local and Remote Endpoints (four-tuple) and their protocol connection state. These map into Connection objects. Whenever a new inbound handshake is being started, the Listener should generate a new Connection object and pass it to the application.

Implementing Listeners for Connectionless Protocols Connectionless protocols such as UDP and UDP-lite generally do not provide the same mechanisms that connected protocols do to offer Connection objects. Implementations should wait for incoming packets for connectionless protocols on a listening port and should perform four-tuple matching of packets to existing Connection objects if possible. If a matching Connection object does not exist, an incoming packet from a connectionless protocol should cause a new Connection object to be created.

Implementing Listeners for Multiplexed Protocols Protocols that provide multiplexing of streams can listen for entirely new connections as well as for new sub-connections (streams of an already existing connection). A new stream arrival on an existing connection is presented to the application as a new Connection. This new Connection is grouped with all other Connections that are multiplexed via the same protocol.

Implementing Sending and Receiving Data The most basic mapping for sending a Message is an abstraction of datagrams, in which the transport protocol naturally deals in discrete packets (such as UDP). Each Message here corresponds to a single datagram. For protocols that expose byte-streams (such as TCP), the only delineation provided by the protocol is the end of the stream in a given direction. Each Message in this case corresponds to the entire stream of bytes in a direction. These Messages may be quite long, in which case they can be sent in multiple parts. Protocols that provide framing (such as length-value protocols, or protocols that use delimiters like HTTP/1.1) may support Message sizes that do not fit within a single datagram. Each Message for framing protocols corresponds to a single frame, which may be sent either as a complete Message in the underlying protocol, or in multiple parts. Messages themselves generally consist of bytes passed in the messageData parameter intended to be processed at an application layer. However, Message objects presented through the API can carry associated Message Properties passed through the messageContext parameter. When these are Protocol Specific Properties, they can include metadata that exists separately from a byte encoding. For example, these Properties can include name-value pairs of information, like HTTP header fields. In such cases, Messages might be "empty", insofar as they contain zero bytes in the messageData parameter, but can still include data in the messageContext that is interpreted by the Protocol Stack.

Sending Messages The effect of the application sending a Message is determined by the top-level protocol in the established Protocol Stack. That is, if the top-level protocol provides an abstraction of framed Messages over a connection, the receiving application will be able to obtain multiple Messages on that connection, even if the framing protocol is built on a byte-stream protocol like TCP.

Message Properties The API allows various properties to be associated with each Message, which should be implemented as discussed below.

msgLifetime: this should be implemented by removing the Message from the queue of pending Messages after the Lifetime has expired. A queue of pending Messages within the Transport Services implementation that have yet to be handed to the Protocol Stack can always support this property, but once a Message has been sent into the send buffer of a protocol, only certain protocols may support removing it from their send buffer. For example, a Transport Services implementation cannot remove bytes from a TCP send buffer, while it can remove data from a SCTP send buffer using the partial reliability extension . When there is no standing queue of Messages within the system, and the Protocol Stack does not support the removal of a Message from the stack's send buffer, this property may be ignored.
msgPriority: this represents the ability to prioritize a Message over other Messages. This can be implemented by the Transport Services system by re-ordering Messages that have yet to be handed to the Protocol Stack, or by giving relative priority hints to protocols that support priorities per Message. For example, an implementation of HTTP/2 could choose to send Messages of different Priority on streams of different priority.
msgOrdered: when this is false, this disables the requirement of in-order-delivery for protocols that support configurable ordering. When the Protocol Stack does not support configurable ordering, this property may be ignored.
safelyReplayable: when this is true, this means that the Message can be used by a transport mechanism that might deliver it multiple times -- e.g., as a result of racing multiple transports or as part of TCP Fast Open. Also, protocols that do not protect against duplicated Messages, such as UDP (when used directly, without a protocol layered atop), can only be used with Messages that are Safely Replayable. When a Transport Services system is permitted to replay Messages, replay protection could be provided by the application.
final: when this is true, this means that the sender will not send any further Messages. The Connection need not be closed (in case the Protocol Stack supports half-close operation, like TCP). Any Messages sent after a Message marked final will result in a SendError.
msgChecksumLen: when this is set to any value other than Full Coverage, it sets the minimum protection in protocols that allow limiting the checksum length (e.g. UDP-Lite). If the Protocol Stack does not support checksum length limitation, this property may be ignored.
msgReliable: When true, the property specifies that the Message must be reliably transmitted. When false, and if unreliable transmission is supported by the underlying protocol, then the Message should be unreliably transmitted. If the underlying protocol does not support unreliable transmission, the Message should be reliably transmitted.
msgCapacityProfile: When true, this expresses a wish to override the Generic Connection Property connCapacityProfile for this Message. Depending on the value, this can, for example, be implemented by changing the DSCP value of the associated packet (note that the guidelines in Section 6 of apply; e.g., the DSCP value should not be changed for different packets within a reliable transport protocol session or DCCP connection).
noFragmentation: Setting this avoids network-layer fragmentation. Messages exceeding the transport's current estimate of its maximum packet size (the singularTransmissionMsgMaxLen Connection Property) can result in transport segmentation when permitted, or generate an error. When used with transports running over IP version 4, the Don't Fragment bit should be set to avoid on-path IP fragmentation ().
noSegmentation: When set, this property limits the Message size to the transport's current estimate of its maximum packet size (the singularTransmissionMsgMaxLen Connection Property). Messages larger than this size generate an error. Setting this avoids transport-layer segmentation and network-layer fragmentation. When used with transports running over IP version 4, the Don't Fragment bit should be set to avoid on-path IP fragmentation ().

Send Completion The application should be notified (using a Sent, Expired or SendError event) whenever a Message or partial Message has been consumed by the Protocol Stack, or has failed to send. The time at which a Message is considered to have been consumed by the Protocol Stack may vary depending on the protocol. For example, for a basic datagram protocol like UDP, this may correspond to the time when the packet is sent into the interface driver. For a protocol that buffers data in queues, like TCP, this may correspond to when the data has entered the send buffer. The time at which a Message failed to send is when the Transport Services implementation (including the Protocol Stack) has experienced a failure related to sending; this can depend on protocol-specific timeouts.

Batching Sends Since sending a Message may involve a context switch between the application and the Transport Services system, sending patterns that involve multiple small Messages can incur high overhead if each needs to be enqueued separately. To avoid this, the application can indicate a batch of Send actions through the API. When this is used, the implementation can defer the processing of Messages until the batch is complete.

Receiving Messages Similar to sending, receiving a Message is determined by the top-level protocol in the established Protocol Stack. The main difference with receiving is that the size and boundaries of the Message are not known beforehand. The application can communicate in its Receive action the parameters for the Message, which can help the Transport Services implementation know how much data to deliver and when. For example, if the application only wants to receive a complete Message, the implementation should wait until an entire Message (datagram, stream, or frame) is read before delivering any Message content to the application. This requires the implementation to understand where Messages end, either via a supplied Message Framer or because the top-level protocol in the established Protocol Stack preserves message boundaries. The application can also control the flow of received data by specifying the minimum and maximum number of bytes of Message content it wants to receive at one time. If a Connection finishes before a requested Receive action can be satisfied, the Transport Services system should deliver any partial Message content outstanding, or if none is available, an indication that there will be no more received Messages.

Handling of data for fast-open protocols Several protocols allow sending higher-level protocol or application data during their protocol establishment, such as TCP Fast Open and TLS 1.3 . This approach is referred to as sending Zero-RTT (0-RTT) data. This is a desirable feature, but poses challenges to an implementation that uses racing during connection establishment. The application can express its preference for sending messagess as 0-RTT data by using the zeroRttMsg Selection Property on the Preconnection. Then, the application can provide the message to send as 0-RTT data via the InitiateWithSend action. In order to be sent as 0-RTT data, the message needs to be marked with the safelyReplayable send paramteter. In general, 0-RTT data may be replayed (for example, if a TCP SYN contains data, and the SYN is retransmitted, the data will be retransmitted as well but may be considered as a new connection instead of a retransmission). When racing connections, different leaf nodes have the opportunity to send the same data independently. If data is truly safely replayable, this is permissible. Once the application has provided its 0-RTT data, a Transport Services implementation should keep a copy of this data and provide it to each new leaf node that is started and for which a protocol instance supporting 0-RTT is being used. Note that the amount of data that can actually be sent as 0-RTT data varies by protocol, so any given Protocol Stack might only consume part of the saved data prior to becoming established. The implementation needs to keep track of how much data a particular Protocol Stack has consumed, and ensure that any pending 0-RTT-eligible data from the application is handled before subsequent Messages. It is also possible for Protocol Stacks within a particular leaf node to use a 0-RTT handshakes in a lower-level protocol without any safely replayable application data if a higher-level protocol in the stack has idempotent handshake data to send. For example, TCP Fast Open could use a Client Hello from TLS as its 0-RTT data, without any data being provided by the application. 0-RTT handshakes often rely on previous state, such as TCP Fast Open cookies, previously established TLS tickets, or out-of-band distributed pre-shared keys (PSKs). Implementations should be aware of security concerns around using these tokens across multiple addresses or paths when racing. In the case of TLS, any given ticket or PSK should only be used on one leaf node, since servers will likely reject duplicate tickets in order to prevent replays (see section-8.1 ). If implementations have multiple tickets available from a previous connection, each leaf node attempt can use a different ticket. In effect, each leaf node will send the same early application data, yet encoded (encrypted) differently on the wire.

Implementing Message Framers Message Framers are functions that define simple transformations between application Message data and raw transport protocol data. Generally, a Message Framer implements a simple application protocol that can either be provided by the Transport Services implementation or by the application. It is optional for Transport Services system implementations to provide Message Framers: the specification does not prescribe any particular Message Framers to be implemented. A Framer can encapsulate or encode outbound Messages, decapsulate or decode inbound data into Messages, and implement parts of protocols that do not directly map to application Messages (such as protocol handshakes or preludes before Message exchange). While many protocols can be represented as Message Framers, for the purposes of the Transport Services API, these are ways for applications or application frameworks to define their own Message parsing to be included within a Connection's Protocol Stack. As an example, TLS is a protocol that is natively supported by the Transport Services API, even though it could also serve the purpose of framing data over TCP. Most Message Framers fall into one of two categories:

Header-prefixed record formats, such as a basic Type-Length-Value (TLV) structure
Delimiter-separated formats, such as HTTP/1.1

Common Message Framers can be provided by a Transport Services implementation, but an implementation ought to allow custom Message Framers to be defined by the application or some other piece of software. This section describes one possible API for defining Message Framers, as an example.

Defining Message Framers A Message Framer is primarily defined by the code that handles events for a framer implementation, specifically how it handles inbound and outbound data parsing. The function that implements custom framing logic will be referred to as the "framer implementation", which may be provided by a Transport Services implementation or the application itself. The Message Framer refers to the object or function within the main Connection implementation that delivers events to the custom framer implementation whenever data is ready to be parsed or framed. The API examples in this section use the notation conventions for the Transport Services API defined in . The Transport Services implementation needs to ensure that all of the events and actions taken on a Message Framer are synchronized to ensure consistent behavior. For example, some of the actions defined below (such as PrependFramer and StartPassthrough) modify how data flows in a protocol stack, and require synchronization with sending and parsing data in the Message Framer. When a Connection establishment attempt begins, an event can be delivered to notify the framer implementation that a new Connection is being created. Similarly, a stop event can be delivered when a Connection is being torn down. The framer implementation can use the Connection object to look up specific properties of the Connection or the network being used that may influence how to frame Messages. Start MessageFramer -> Stop ]]> When a Message Framer generates a Start event, the framer implementation has the opportunity to start writing some data prior to the Connection delivering its Ready event. This allows the implementation to communicate control data to the Remote Endpoint that can be used to parse Messages. Once the framer implementation has completed its setup or handshake, it can indicate to the application that it is ready for handling data with this call. Similarly, when a Message Framer generates a Stop event, the framer implementation has the opportunity to write some final data or clear up its local state before the Closed event is delivered to the Application. The framer implementation can indicate that it has finished with this call. At any time if the implementation encounters a fatal error, it can also cause the Connection to fail and provide an error. Should the framer implementation deem the candidate selected during racing unsuitable, it can signal this to the Transport Services API by failing the Connection prior to marking it as ready. If there are no other candidates available, the Connection will fail. Otherwise, the Connection will select a different candidate and the Message Framer will generate a new Start event. Before an implementation marks a Message Framer as ready, it can also dynamically add a protocol or framer above it in the stack. This allows protocols that need to add TLS conditionally, like STARTTLS , to modify the Protocol Stack based on a handshake result. A Message Framer might also choose to go into a passthrough mode once an initial exchange or handshake has been completed, such as the STARTTLS case mentioned above. This can also be useful for proxy protocols like SOCKS or HTTP CONNECT . In such cases, a Message Framer implementation can intercept sending and receiving of Messages at first, but then indicate that no more processing is needed.

Sender-side Message Framing Message Framers generate an event whenever a Connection sends a new Message. The parameters to the event align with the Send action in the API (). NewSentMessage ]]> Upon receiving this event, a framer implementation is responsible for performing any necessary transformations and sending the resulting data back to the Message Framer, which will in turn send it to the next protocol. To improve performance, implementations should ensure that there is a way to pass the original data through without copying. To provide an example, a simple protocol that adds the length of the Message data as a header would receive the NewSentMessage event, create a data representation of the length of the Message data, and then send a block of data that is the concatenation of the length header and the original Message data.

Receiver-side Message Framing In order to parse a received flow of data into Messages, the Message Framer notifies the framer implementation whenever new data is available to parse. The parameters to the events and calls for receiving data with a framer align with the Receive action in the API (). HandleReceivedData ]]> Upon receiving this event, the framer implementation can inspect the inbound data. The data is parsed from a particular cursor representing the unprocessed data. The application requests a specific amount of data it needs to have available in order to parse. If the data is not available, the parse fails. (messageData, messageContext, endOfMessage) ]]> The framer implementation can directly advance the receive cursor once it has parsed data to effectively discard data (for example, discard a header once the content has been parsed). To deliver a Message to the application, the framer implementation can either directly deliver data that it has allocated, or deliver a range of data directly from the underlying transport and simultaneously advance the receive cursor. Note that MessageFramer.DeliverAndAdvanceReceiveCursor allows the framer implementation to earmark bytes as part of a Message even before they are received by the transport. This allows the delivery of very large Messages without requiring the implementation to directly inspect all of the bytes. To provide an example, a simple protocol that parses the length of the Message data as a header value would receive the HandleReceivedData event, and call Parse with a minimum and maximum set to the length of the header field. Once the parse succeeded, it would call AdvanceReceiveCursor with the length of the header field, and then call DeliverAndAdvanceReceiveCursor with the length of the body that was parsed from the header, marking the new Message as complete.

Implementing Connection Management Once a Connection is established, the Transport Services API allows applications to interact with the Connection by modifying or inspecting Connection Properties. A Connection can also generate error events in the form of SoftError events. The set of Connection Properties that are supported for setting and getting on a Connection are described in . For any properties that are generic, and thus could apply to all protocols being used by a Connection, the Transport Services implementation should store the properties in storage common to all protocols, and notify all protocol instances in the Protocol Stack whenever the properties have been modified by the application. and offer guidance on how to do this for TCP, MPTCP, SCTP, UDP and UDP-Lite; see {#specific-protocol-considerations} for a description of a back-tracking method to find the relevant protocol primitives using these documents. For Protocol-specific Properties, such as the User Timeout that applies to TCP, the Transport Services implementation only needs to update the relevant protocol instance. If an error is encountered in setting a property (for example, if the application tries to set a TCP-specific property on a Connection that is not using TCP), the action must fail gracefully. The application must be informed of the error, but the Connection itself must not be terminated. When protocol instances in the Protocol Stack report generic or protocol-specific errors, the API will deliver them to the application as SoftError events. These allow the application to be informed of ICMP errors, and other similar events.

Pooled Connection For applications that do not need in-order delivery of Messages, the Transport Services implementation may distribute Messages of a single Connection across several underlying transport connections or multiple streams of multi-streaming connections between endpoints, as long as all of these satisfy the Selection Properties. The Transport Services implementation will then hide this connection management and only expose a single Connection object, which we here call a "Pooled Connection". This is in contrast to Connection Groups, which explicitly expose combined treatment of Connections, giving the application control over multiplexing, for example. Pooled Connections can be useful when the application using the Transport Services system implements a protocol such as HTTP, which employs request/response pairs and does not require in-order delivery of responses. This enables implementations of Transport Services systems to realize transparent connection coalescing, connection migration, and to perform per-message endpoint and path selection by choosing among multiple underlying connections.

Handling Path Changes When a path change occurs, e.g., when the IP address of an interface changes or a new interface becomes available, the Transport Services implementation is responsible for notifying the Protocol Instance of the change. The path change may interrupt connectivity on a path for an active Connection or provide an opportunity for a transport that supports multipath or migration to adapt to the new paths. Note that, in the model of the Transport Services API, migration is considered a part of multipath connectivity; it is just a limiting policy on multipath usage. If the multipath Selection Property is set to Disabled, migration is disallowed. For protocols that do not support multipath or migration, the Protocol Instances should be informed of the path change, but should not be forcibly disconnected if the previously used path becomes unavailable. There are many common usage scenarios that can lead to a path becoming temporarily unavailable, and then recovering before the transport protocol reaches a timeout error. These are particularly common using mobile devices. Examples include: an Ethernet cable becoming unplugged and then plugged back in; a device losing a Wi-Fi signal while a user is in an elevator, and reattaching when the user leaves the elevator; and a user losing the radio signal while riding a train through a tunnel. If the device is able to rejoin a network with the same IP address, a stateful transport connection can generally resume. Thus, while it is useful for a Protocol Instance to be aware of a temporary loss of connectivity, the Transport Services implementation should not aggressively close Connections in these scenarios. If the Protocol Stack includes a transport protocol that supports multipath connectivity, the Transport Services implementation should also inform the Protocol Instance about potentially new paths that become permissible based on the multipath Selection Property and the multipathPolicy Connection Property choices made by the application. A protocol can then establish new subflows over new paths while an active path is still available or, if migration is supported, also after a break has been detected, and should attempt to tear down subflows over paths that are no longer used. The Connection Property multipathPolicy of the Transport Services API allows an application to indicate when and how different paths should be used. However, detailed handling of these policies is implementation-specific. For example, if the multipath Selection Property is set to active, the decision about when to create a new path or to announce a new path or set of paths to the Remote Endpoint, e.g., in the form of additional IP addresses, is implementation-specific. If the Protocol Stack includes a transport protocol that does not support multipath, but does support migrating between paths, the update to the set of available paths can trigger the connection to be migrated. In the case of a Pooled Connection , the Transport Services implementation may add connections over new paths to the pool if permissible based on the multipath policy and Selection Properties. In the case that a previously used path becomes unavailable, the Transport Services system may disconnect all connections that require this path, but should not disconnect the pooled Connection object exposed to the application. The strategy to do so is implementation-specific, but should be consistent with the behavior of multipath transports.

Implementing Connection Termination For Close (which leads to a Closed event) and Abort (which leads to a ConnectionError event), the application might find it useful to be informed when a peer closes or aborts a Connection. Whether this is possible depends on the underlying protocol, and no guarantees can be given. When an underlying transport connection supports multi-streaming (such as SCTP), the Transport Services system can use a stream reset procedure to cause a Finish event upon a Close action from the peer .

Cached State Beyond a single Connection's lifetime, it is useful for an implementation to keep state and history. This cached state can help improve future Connection establishment due to re-using results and credentials, and favoring paths and protocols that performed well in the past. Cached state may be associated with different endpoints for the same Connection, depending on the protocol generating the cached content. For example, session tickets for TLS are associated with specific endpoints, and thus should be cached based on a connection's hostname endpoint (if applicable). However, performance characteristics of a path are more likely tied to the IP address and subnet being used.

Protocol state caches Some protocols will have long-term state to be cached in association with endpoints. This state often has some time after which it is expired, so the implementation should allow each protocol to specify an expiration for cached content. Examples of cached protocol state include:

The DNS protocol can cache resolved addresses (such as those retrieved from A and AAAA queries), associated with a Time To Live (TTL) to be used for future hostname resolutions without requiring asking the DNS resolver again.
TLS caches session state and tickets based on a hostname, which can be used for resuming sessions with a server.
TCP can cache cookies for use in TCP Fast Open.

Cached protocol state is primarily used during Connection establishment for a single Protocol Stack, but may be used to influence an implementation's preference between several candidate Protocol Stacks. For example, if two IP address endpoints are otherwise equally preferred, an implementation may choose to attempt a connection to an address for which it has a TCP Fast Open cookie. Applications can use the Transport Services API to request that a Connection Group maintain a separate cache for protocol state. Connections in the group will not use cached state from Connections outside the group, and Connections outside the group will not use state cached from Connections inside the group. This may be necessary, for example, if application-layer identifiers rotate and clients wish to avoid linkability via trackable TLS tickets or TFO cookies.

Performance caches In addition to protocol state, Protocol Instances should provide data into a performance-oriented cache to help guide future protocol and path selection. Some performance information can be gathered generically across several protocols to allow predictive comparisons between protocols on given paths:

Observed Round Trip Time
Connection establishment latency
Connection establishment success rate

These items can be cached on a per-address and per-subnet granularity, and averaged between different values. The information should be cached on a per-network basis, since it is expected that different network attachments will have different performance characteristics. Besides Protocol Instances, other system entities may also provide data into performance-oriented caches. This could for instance be signal strength information reported by radio modems like Wi-Fi and mobile broadband or information about the battery-level of the device. Furthermore, the system may cache the observed maximum throughput on a path as an estimate of the available bandwidth. An implementation should use this information, when possible, to influence preference between candidate paths, endpoints, and protocol options. Eligible options that historically had significantly better performance than others should be selected first when gathering candidates (see ) to ensure better performance for the application. The reasonable lifetime for cached performance values will vary depending on the nature of the value. Certain information, like the connection establishment success rate to a Remote Endpoint using a given Protocol Stack, can be stored for a long period of time (hours or longer), since it is expected that the capabilities of the Remote Endpoint are not changing very quickly. On the other hand, the Round Trip Time observed by TCP over a particular network path may vary over a relatively short time interval. For such values, the implementation should remove them from the cache more quickly, or treat older values with less confidence/weight. provides guidance about sharing of TCP Control Block information between connections on initialization.

Specific Transport Protocol Considerations Each protocol that is supported by a Transport Services implementation should have a well-defined API mapping. API mappings for a protocol are important for Connections in which a given protocol is the "top" of the Protocol Stack. For example, the mapping of the Send function for TCP applies to Connections in which the application directly sends over TCP. Each protocol has a notion of Connectedness. Possible values for Connectedness are:

Connectionless. Connectionless protocols do not establish explicit state between endpoints, and do not perform a handshake during Connection establishment.
Connected. Connected (also called "connection-oriented") protocols establish state between endpoints, and perform a handshake during connection establishment. The handshake may be 0-RTT to send data or resume a session, but bidirectional traffic is required to confirm connectedness.
Multiplexing Connected. Multiplexing Connected protocols share properties with Connected protocols, but also explictly support opening multiple application-level flows. This means that they can support cloning new Connection objects without a new explicit handshake.

Protocols also have a notion of Data Unit. Possible values for Data Unit are:

Byte-stream. Byte-stream protocols do not define any message boundaries of their own apart from the end of a stream in each direction.
Datagram. Datagram protocols define message boundaries at the same level of transmission, such that only complete (not partial) messages are supported.
Message. Message protocols support message boundaries that can be sent and received either as complete or partial messages. Maximum message lengths can be defined, and messages can be partially reliable.

Below, terms in capitals with a dot (e.g., "CONNECT.SCTP") refer to the primitives with the same name in section 4 of . For further implementation details, the description of these primitives in points to section 3 of and section 3 of , which refers back to the relevant specifications for each protocol. This back-tracking method applies to all elements of (see appendix D of ): they are listed in appendix A of with an implementation hint in the same style, pointing back to section 4 of . This document presents the protocol mappings defined in . Other protocol mappings can be provided as separate documents, following the mapping template in .

TCP Connectedness: Connected Data Unit: Byte-stream

Connection Object:: TCP connections between two hosts map directly to Connection objects.
Initiate:: CONNECT.TCP. Calling Initiate on a TCP Connection causes it to reserve a local port, and send a SYN to the Remote Endpoint.
InitiateWithSend:: CONNECT.TCP with parameter user message. Early safely replayable data is sent on a TCP Connection in the SYN, as TCP Fast Open data.
Ready:: A TCP Connection is ready once the three-way handshake is complete.
EstablishmentError:: Failure of CONNECT.TCP. TCP can throw various errors during connection setup. Specifically, it is important to handle a RST being sent by the peer during the handshake.
ConnectionError:: Once established, TCP throws errors whenever the connection is disconnected, such as due to receiving a RST from the peer.
Listen:: LISTEN.TCP. Calling Listen for TCP binds a local port and prepares it to receive inbound SYN packets from peers.
ConnectionReceived:: TCP Listeners will deliver new connections once they have replied to an inbound SYN with a SYN-ACK.
Clone:: Calling Clone on a TCP Connection creates a new Connection with equivalent parameters. These Connections, and Connections generated via later calls to Clone on an Established Connection, form a Connection Group. To realize entanglement for these Connections, with the exception of connPriority, changing a Connection Property on one of them must affect the Connection Properties of the others too. No guarantees of honoring the Connection Property connPriority are given, and thus it is safe for an implementation of a Transport Services system to ignore this property. When it is reasonable to assume that Connections traverse the same path (e.g., when they share the same encapsulation), support for it can also experimentally be implemented using a congestion control coupling mechanism (see for example or ).
Send:: SEND.TCP. TCP does not on its own preserve message boundaries. Calling Send on a TCP connection lays out the bytes on the TCP send stream without any other delineation. Any Message marked as Final will cause TCP to send a FIN once the Message has been completely written, by calling CLOSE.TCP immediately upon successful termination of SEND.TCP. Note that transmitting a Message marked as Final should not cause the Closed event to be delivered to the application, as it will still be possible to receive data until the peer closes or aborts the TCP connection.
Receive:: With RECEIVE.TCP, TCP delivers a stream of bytes without any Message delineation. All data delivered in the Received or ReceivedPartial event will be part of a single stream-wide Message that is marked Final (unless a Message Framer is used). EndOfMessage will be delivered when the TCP Connection has received a FIN (CLOSE-EVENT.TCP) from the peer. Note that reception of a FIN should not cause the Closed event to be delivered to the application, as it will still be possible for the application to send data.
Close:: Calling Close on a TCP Connection indicates that the Connection should be gracefully closed (CLOSE.TCP) by sending a FIN to the peer. It will then still be possible to receive data until the peer closes or aborts the TCP connection. The Closed event will be issued upon reception of a FIN.
Abort:: Calling Abort on a TCP Connection indicates that the Connection should be immediately closed by sending a RST to the peer (ABORT.TCP).
CloseGroup:: Calling CloseGroup on a TCP Connection (CLOSE.TCP) is identical to calling Close on this Connection and on all Connections in the same ConnectionGroup.
AbortGroup:: Calling AbortGroup on a TCP Connection (ABORT.TCP) is identical to calling Abort on this Connection and on all Connections in the same ConnectionGroup.

MPTCP Connectedness: Connected Data Unit: Byte-stream The Transport Services API mappings for MPTCP are identical to TCP. MPTCP adds support for multipath properties, such as multipath and multipathPolicy, and actions for managing paths, such as AddRemote and RemoveRemote.

UDP Connectedness: Connectionless Data Unit: Datagram

Connection Object:: UDP Connections represent a pair of specific IP addresses and ports on two hosts.
Initiate:: CONNECT.UDP. Calling Initiate on a UDP Connection causes it to reserve a local port, but does not generate any traffic.
InitiateWithSend:: Early data on a UDP Connection does not have any special meaning. The data is sent whenever the Connection is Ready.
Ready:: A UDP Connection is ready once the system has reserved a local port and has a path to send to the Remote Endpoint.
EstablishmentError:: UDP Connections can only generate errors on initiation due to port conflicts on the local system.
ConnectionError:: UDP Connections can only generate Connection errors in response to Abort calls. (Once in use, UDP Connections can also generate SoftError events (ERROR.UDP(-Lite)) upon receiving ICMP notifications indicating failures in the network.)
Listen:: LISTEN.UDP. Calling Listen for UDP binds a local port and prepares it to receive inbound UDP datagrams from peers.
ConnectionReceived:: UDP Listeners will deliver new connections once they have received traffic from a new Remote Endpoint.
Clone:: Calling Clone on a UDP Connection creates a new Connection with equivalent parameters. The two Connections are otherwise independent.
Send:: SEND.UDP(-Lite). Calling Send on a UDP connection sends the data as the payload of a complete UDP datagram. Marking Messages as Final does not change anything in the datagram's contents. Upon sending a UDP datagram, some relevant fields and flags in the IP header can be controlled: DSCP (SET_DSCP.UDP(-Lite)), DF in IPv4 (SET_DF.UDP(-Lite)) and ECN flag (SET_ECN.UDP(-Lite)).
Receive:: RECEIVE.UDP(-Lite). UDP only delivers complete Messages to Received, each of which represents a single datagram received in a UDP packet. Upon receiving a UDP datagram, the ECN flag from the IP header can be obtained (GET_ECN.UDP(-Lite)).
Close:: Calling Close on a UDP Connection (ABORT.UDP(-Lite)) releases the local port reservation. The Connection then issues a Closed event.
Abort:: Calling Abort on a UDP Connection (ABORT.UDP(-Lite)) is identical to calling Close, except that the Connection will send a ConnectionError event rather than a Closed event.
CloseGroup:: Calling CloseGroup on a UDP Connection (ABORT.UDP(-Lite)) is identical to calling Close on this Connection and on all Connections in the same ConnectionGroup.
AbortGroup:: Calling AbortGroup on a UDP Connection (ABORT.UDP(-Lite)) is identical to calling Close on this Connection and on all Connections in the same ConnectionGroup.

UDP-Lite Connectedness: Connectionless Data Unit: Datagram The Transport Services API mappings for UDP-Lite are identical to UDP. In addition, UDP-Lite supports the msgChecksumLen and recvChecksumLen Properties that allow an application to specify the minimum number of bytes in a Message that need to be covered by a checksum.

UDP Multicast Receive Connectedness: Connectionless Data Unit: Datagram

Connection Object:: Established UDP Multicast Receive connections represent a pair of specific IP addresses and ports. The direction Selection Property must be set to unidirectional receive, and the Local Endpoint must be configured with a group IP address and a port.
Initiate:: Calling Initiate on a UDP Multicast Receive Connection causes an immediate EstablishmentError. This is an unsupported operation.
InitiateWithSend:: Calling InitiateWithSend on a UDP Multicast Receive Connection causes an immediate EstablishmentError. This is an unsupported operation.
Ready:: A UDP Multicast Receive Connection is ready once the system has received traffic for the appropriate group and port.
EstablishmentError:: UDP Multicast Receive Connections generate an EstablishmentError indicating that joining a multicast group failed if Initiate is called.
ConnectionError:: The only ConnectionError generated by a UDP Multicast Receive Connection is in response to an Abort call.
Listen:: LISTEN.UDP. Calling Listen for UDP Multicast Receive binds a local port, prepares it to receive inbound UDP datagrams from peers, and issues a multicast host join. If a Remote Endpoint with an address is supplied, the join is Source-specific Multicast, and the path selection is based on the route to the Remote Endpoint. If a Remote Endpoint is not supplied, the join is Any-source Multicast, and the path selection is based on the outbound route to the group supplied in the Local Endpoint.

There are cases where it is required to open multiple connections for the same address(es). For example, one Connection might be opened for a multicast group to for a multicast control bus, and another application later opens a separate Connection to the same group to send signals to and/or receive signals from the common bus. In such cases, the Transport Services system needs to explicitly enable re-use of the same set of addresses (equivalent to setting SO_REUSEADDR in the socket API).

ConnectionReceived:: UDP Multicast Receive Listeners will deliver new Connections once they have received traffic from a new Remote Endpoint.
Clone:: Calling Clone on a UDP Multicast Receive Connection creates a new Connection with equivalent parameters. The two Connections are otherwise independent.
Send:: SEND.UDP(-Lite). Calling Send on a UDP Multicast Receive connection causes an immediate SendError. This is an unsupported operation.
Receive:: RECEIVE.UDP(-Lite). The Receive operation in a UDP Multicast Receive connection only delivers complete Messages to Received, each of which represents a single datagram received in a UDP packet. Upon receiving a UDP datagram, the ECN flag from the IP header can be obtained (GET_ECN.UDP(-Lite)).
Close:: Calling Close on a UDP Multicast Receive Connection (ABORT.UDP(-Lite)) releases the local port reservation and leaves the group. The Connection then issues a Closed event.
Abort:: Calling Abort on a UDP Multicast Receive Connection (ABORT.UDP(-Lite)) is identical to calling Close, except that the Connection will send a ConnectionError event rather than a Closed event.
CloseGroup:: Calling CloseGroup on a UDP Multicast Receive Connection (ABORT.UDP(-Lite)) is identical to calling Close on this Connection and on all Connections in the same ConnectionGroup.
AbortGroup:: Calling AbortGroup on a UDP Multicast Receive Connection (ABORT.UDP(-Lite)) is identical to calling Close on this Connection and on all Connections in the same ConnectionGroup.

SCTP Connectedness: Connected Data Unit: Message

Connection Object:: Connection objects can be mapped to an SCTP association or a stream in an SCTP association. Mapping Connection objects to SCTP streams is called "stream mapping" and has additional requirements as follows. The following explanation assumes a client-server communication model.

Stream mapping requires an association to already be in place between the client and the server, and it requires the server to understand that a new incoming stream should be represented as a new Connection object by the Transport Services system. A new SCTP stream is created by sending an SCTP message with a new stream id. Thus, to implement stream mapping, the Transport Services API must provide a newly created Connection object to the application upon the reception of such a message. The necessary semantics to implement a Transport Services system Close and Abort primitives are provided by the stream reconfiguration (reset) procedure described in . This also allows to re-use a stream id after resetting ("closing") the stream. To implement this functionality, SCTP stream reconfiguration must be supported by both the client and the server side. To avoid head-of-line blocking, stream mapping should only be implemented when both sides support message interleaving . This allows a sender to schedule transmissions between multiple streams without risking that transmission of a large message on one stream might block transmissions on other streams for a long time. To avoid conflicts between stream ids, the following procedure is recommended: the first Connection, for which the SCTP association has been created, must always use stream id zero. All additional Connections are assigned to unused stream ids in growing order. To avoid a conflict when both endpoints map new Connections simultaneously, the peer which initiated association must use even stream ids whereas the remote side must map its Connections to odd stream ids. Both sides maintain a status map of the assigned stream ids. Generally, new streams should consume the lowest available (even or odd, depending on the side) stream id; this rule is relevant when lower ids become available because Connection objects associated with the streams are closed. SCTP stream mapping as described here has been implemented in a research prototype; a desription of this implementation is given in .

Initiate:: If this is the only Connection object that is assigned to the SCTP Association or stream mapping is not used, CONNECT.SCTP is called. Else, unless the Selection Property activeReadBeforeSend is Preferred or Required, a new stream is used: if there are enough streams available, Initiate is a local operation that assigns a new stream id to the Connection object. The number of streams is negotiated as a parameter of the prior CONNECT.SCTP call, and it represents a trade-off between local resource usage and the number of Connection objects that can be mapped without requiring a reconfiguration signal. When running out of streams, ADD_STREAM.SCTP must be called.
InitiateWithSend:: If this is the only Connection object that is assigned to the SCTP association or stream mapping is not used, CONNECT.SCTP is called with the "user message" parameter. Else, a new stream is used (see Initiate for how to handle running out of streams), and this just sends the first message on a new stream.
Ready:: Initiate or InitiateWithSend returns without an error, i.e. SCTP's four-way handshake has completed. If an association with the peer already exists, stream mapping is used and enough streams are available, a Connection object instantly becomes Ready after calling Initiate or InitiateWithSend.
EstablishmentError:: Failure of CONNECT.SCTP.
ConnectionError:: TIMEOUT.SCTP or ABORT-EVENT.SCTP.
Listen:: LISTEN.SCTP. If an association with the peer already exists and stream mapping is used, Listen just expects to receive a new message with a new stream id (chosen in accordance with the stream id assignment procedure described above).
ConnectionReceived:: LISTEN.SCTP returns without an error (a result of successful CONNECT.SCTP from the peer), or, in case of stream mapping, the first message has arrived on a new stream (in this case, Receive is also invoked).
Clone:: Calling Clone on an SCTP association creates a new Connection object and assigns it a new stream id in accordance with the stream id assignment procedure described above. If there are not enough streams available, ADD_STREAM.SCTP must be called.
Send:: SEND.SCTP. Message Properties such as msgLifetime and msgOrdered map to parameters of this primitive.
Receive:: RECEIVE.SCTP. The "partial flag" of RECEIVE.SCTP invokes a ReceivedPartial event.

Close: If this is the only Connection object that is assigned to the SCTP association, CLOSE.SCTP is called, and the Closed event will be delivered to the application upon the ensuing CLOSE-EVENT.SCTP. Else, the Connection object is one out of several Connection objects that are assigned to the same SCTP assocation, and RESET_STREAM.SCTP must be called, which informs the peer that the stream will no longer be used for mapping and can be used by future Initiate, InitiateWithSend or Listen calls. At the peer, the event RESET_STREAM-EVENT.SCTP will fire, which the peer must answer by issuing RESET_STREAM.SCTP too. The resulting local RESET_STREAM-EVENT.SCTP informs the Transport Services system that the stream id can now be re-used by the next Initiate, InitiateWithSend or Listen calls, and invokes a Closed event towards the application. Abort: If this is the only Connection object that is assigned to the SCTP association, ABORT.SCTP is called. Else, the Connection object is one out of several Connection objects that are assigned to the same SCTP assocation, and shutdown proceeds as described under Close. CloseGroup: Calling CloseGroup calls CLOSE.SCTP, closing all Connections in the SCTP association. AbortGroup: Calling AbortGroup calls ABORT.SCTP, immediately closing all Connections in the SCTP association. In addition to the API mappings described above, when there are multiple Connection objects assigned to the same SCTP association, SCTP can support Connection properties such as connPriority and connScheduler where CONFIGURE_STREAM_SCHEDULER.SCTP can be called to adjust the priorities of streams in the SCTP association.

IANA Considerations RFC-EDITOR: Please remove this section before publication. This document has no actions for IANA.

Security Considerations outlines general security consideration and requirements for any system that implements the Transport Services archtecture. provides further discussion on security and privacy implications of the Transport Services API. This document provides additional guidance on implementation specifics for the Transport Services API and as such the security considerations in both of these documents apply. The next two subsections discuss further considerations that are specific to mechanisms specified in this document.

Considerations for Candidate Gathering The Security Considerations of the Transport Services Architecture forbids gathering and racing with Protocol Stacks that do not have equivalent security properties. Therefore, implementations need to avoid downgrade attacks that allow network interference to cause the implementation to select less secure, or entirely insecure, combinations of paths and protocols.

Considerations for Candidate Racing See for security considerations around racing with 0-RTT data. An attacker that knows a particular device is racing several options during connection establishment may be able to block packets for the first connection attempt, thus inducing the device to fall back to a secondary attempt. This is a problem if the secondary attempts have worse security properties that enable further attacks. Implementations should ensure that all options have equivalent security properties to avoid incentivizing attacks. Since results from the network can determine how a connection attempt tree is built, such as when DNS returns a list of resolved endpoints, it is possible for the network to cause an implementation to consume significant on-device resources. Implementations should limit the maximum amount of state allowed for any given node, including the number of child nodes, especially when the state is based on results from the network.

Acknowledgements This work has received funding from the European Union's Horizon 2020 research and innovation programme under grant agreement No. 644334 (NEAT) and No. 815178 (5GENESIS). This work has been supported by Leibniz Prize project funds of DFG - German Research Foundation: Gottfried Wilhelm Leibniz-Preis 2011 (FKZ FE 570/4-1). This work has been supported by the UK Engineering and Physical Sciences Research Council under grant EP/R04144X/1. This work has been supported by the Research Council of Norway under its "Toppforsk" programme through the "OCARINA" project. Thanks to Colin Perkins, Tom Jones, Karl-Johan Grinnemo, Gorry Fairhurst, for their contributions to the design of this specification. Thanks also to Stuart Cheshire, Josh Graessley, David Schinazi, and Eric Kinnear for their implementation and design efforts, including Happy Eyeballs, that heavily influenced this work.

References Normative References An Architecture for Transport Services Apple Inc. Google Switzerland GmbH Karlstad University University of Aberdeen University of Glasgow This document describes an architecture for exposing transport protocol features to applications for network communication, a Transport Services system. The Transport Services Application Programming Interface (API) is based on an asynchronous, event-driven interaction pattern. This API uses messages for representing data transfer to applications, and describes how implementations can use multiple IP addresses, multiple protocols, and multiple paths, and provide multiple application streams. This document further defines common terminology and concepts to be used in definitions of a Transport Service API and a Transport Services implementation. An Abstract Application Layer Interface to Transport Services Google Switzerland GmbH University of Oslo Netflix University of Aberdeen Ericsson University of Glasgow SAP SE Apple Inc. This document describes an abstract application programming interface, API, to the transport layer that enables the selection of transport protocols and network paths dynamically at runtime. This API enables faster deployment of new protocols and protocol features without requiring changes to the applications. The specified API follows the Transport Services architecture by providing asynchronous, atomic transmission of messages. It is intended to replace the BSD sockets API as the common interface to the transport layer, in an environment where endpoints could select from multiple interfaces and potential transport protocols. Happy Eyeballs Version 2: Better Connectivity Using Concurrency Many communication protocols operating over the modern Internet use hostnames. These often resolve to multiple IP addresses, each of which may have different performance and connectivity characteristics. Since specific addresses or address families (IPv4 or IPv6) may be blocked, broken, or sub-optimal on a network, clients that attempt multiple connections in parallel have a chance of establishing a connection more quickly. This document specifies requirements for algorithms that reduce this user-visible delay and provides an example algorithm, referred to as "Happy Eyeballs". This document obsoletes the original algorithm description in RFC 6555. Hypertext Transfer Protocol Version 2 (HTTP/2) This specification describes an optimized expression of the semantics of the Hypertext Transfer Protocol (HTTP), referred to as HTTP version 2 (HTTP/2). HTTP/2 enables a more efficient use of network resources and a reduced perception of latency by introducing header field compression and allowing multiple concurrent exchanges on the same connection. It also introduces unsolicited push of representations from servers to clients. This specification is an alternative to, but does not obsolete, the HTTP/1.1 message syntax. HTTP's existing semantics remain unchanged. Guidelines for Multihomed and IPv4/IPv6 Dual-Stack Interactive Connectivity Establishment (ICE) This document provides guidelines on how to make Interactive Connectivity Establishment (ICE) conclude faster in multihomed and IPv4/IPv6 dual-stack scenarios where broken paths exist. The provided guidelines are backward compatible with the original ICE specification (see RFC 5245). On the Usage of Transport Features Provided by IETF Transport Protocols This document describes how the transport protocols Transmission Control Protocol (TCP), MultiPath TCP (MPTCP), Stream Control Transmission Protocol (SCTP), User Datagram Protocol (UDP), and Lightweight User Datagram Protocol (UDP-Lite) expose services to applications and how an application can configure and use the features that make up these services. It also discusses the service provided by the Low Extra Delay Background Transport (LEDBAT) congestion control mechanism. The description results in a set of transport abstractions that can be exported in a transport services (TAPS) API. Transport Features of the User Datagram Protocol (UDP) and Lightweight UDP (UDP-Lite) This is an informational document that describes the transport protocol interface primitives provided by the User Datagram Protocol (UDP) and the Lightweight User Datagram Protocol (UDP-Lite) transport protocols. It identifies the datagram services exposed to applications and how an application can configure and use the features offered by the Internet datagram transport service. RFC 8303 documents the usage of transport features provided by IETF transport protocols, describing the way UDP, UDP-Lite, and other transport protocols expose their services to applications and how an application can configure and use the features that make up these services. This document provides input to and context for that document, as well as offers a road map to documentation that may help users of the UDP and UDP-Lite protocols. TCP Fast Open This document describes an experimental TCP mechanism called TCP Fast Open (TFO). TFO allows data to be carried in the SYN and SYN-ACK packets and consumed by the receiving end during the initial connection handshake, and saves up to one full round-trip time (RTT) compared to the standard TCP, which requires a three-way handshake (3WHS) to complete before data can be exchanged. However, TFO deviates from the standard TCP semantics, since the data in the SYN could be replayed to an application in some rare circumstances. Applications should not use TFO unless they can tolerate this issue, as detailed in the Applicability section. The Transport Layer Security (TLS) Protocol Version 1.3 This document specifies version 1.3 of the Transport Layer Security (TLS) protocol. TLS allows client/server applications to communicate over the Internet in a way that is designed to prevent eavesdropping, tampering, and message forgery. This document updates RFCs 5705 and 6066, and obsoletes RFCs 5077, 5246, and 6961. This document also specifies new requirements for TLS 1.2 implementations. A Minimal Set of Transport Services for End Systems This document recommends a minimal set of Transport Services offered by end systems and gives guidance on choosing among the available mechanisms and protocols. It is based on the set of transport features in RFC 8303. Informative References Transparent Flow Mapping for NEAT ctrlTCP: Reducing Latency through Coupled, Heterogeneous Multi-Flow TCP Congestion Control n.d. UDP Usage Guidelines The User Datagram Protocol (UDP) provides a minimal message-passing transport that has no inherent congestion control mechanisms. This document provides guidelines on the use of UDP for the designers of applications, tunnels, and other protocols that use UDP. Congestion control guidelines are a primary focus, but the document also provides guidance on other topics, including message sizes, reliability, checksums, middlebox traversal, the use of Explicit Congestion Notification (ECN), Differentiated Services Code Points (DSCPs), and ports. Because congestion control is critical to the stable operation of the Internet, applications and other protocols that choose to use UDP as an Internet transport must employ mechanisms to prevent congestion collapse and to establish some degree of fairness with concurrent traffic. They may also need to implement additional mechanisms, depending on how they use UDP. Some guidance is also applicable to the design of other protocols (e.g., protocols layered directly on IP or via IP-based tunnels), especially when these protocols do not themselves provide congestion control. This document obsoletes RFC 5405 and adds guidelines for multicast UDP usage. A DNS RR for specifying the location of services (DNS SRV) This document describes a DNS RR which specifies the location of the server(s) for a specific protocol and domain. [STANDARDS-TRACK] Service binding and parameter specification via the DNS (DNS SVCB and HTTPS RRs) Google Akamai Technologies Akamai Technologies This document specifies the "SVCB" and "HTTPS" DNS resource record (RR) types to facilitate the lookup of information needed to make connections to network services, such as for HTTP origins. SVCB records allow a service to be provided from multiple alternative endpoints, each with associated parameters (such as transport protocol configuration), and are extensible to support future uses (such as keys for encrypting the TLS ClientHello). They also enable aliasing of apex domains, which is not possible with CNAME. The HTTPS RR is a variation of SVCB for use with HTTP [HTTP]. By providing more information to the client before it attempts to establish a connection, these records offer potential benefits to both performance and privacy. TO BE REMOVED: This document is being collaborated on in Github at: https://github.com/MikeBishop/dns-alt-svc (https://github.com/MikeBishop/dns-alt-svc). The most recent working version of the document, open issues, etc. should all be available there. The authors (gratefully) accept pull requests. DNS-Based Service Discovery This document specifies how DNS resource records are named and structured to facilitate service discovery. Given a type of service that a client is looking for, and a domain in which the client is looking for that service, this mechanism allows clients to discover a list of named instances of that desired service, using standard DNS queries. This mechanism is referred to as DNS-based Service Discovery, or DNS-SD. Multicast DNS As networked devices become smaller, more portable, and more ubiquitous, the ability to operate with less configured infrastructure is increasingly important. In particular, the ability to look up DNS resource record data types (including, but not limited to, host names) in the absence of a conventional managed DNS server is useful. Multicast DNS (mDNS) provides the ability to perform DNS-like operations on the local link in the absence of any conventional Unicast DNS server. In addition, Multicast DNS designates a portion of the DNS namespace to be free for local use, without the need to pay any annual fee, and without the need to set up delegations or otherwise configure a conventional DNS server to answer for those names. The primary benefits of Multicast DNS names are that (i) they require little or no administration or configuration to set them up, (ii) they work when no infrastructure is present, and (iii) they work during infrastructure failures. QUIC: A UDP-Based Multiplexed and Secure Transport 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. Interactive Connectivity Establishment (ICE): A Protocol for Network Address Translator (NAT) Traversal This document describes a protocol for Network Address Translator (NAT) traversal for UDP-based communication. This protocol is called Interactive Connectivity Establishment (ICE). ICE makes use of the Session Traversal Utilities for NAT (STUN) protocol and its extension, Traversal Using Relay NAT (TURN). This document obsoletes RFC 5245. Session Traversal Utilities for NAT (STUN) Session Traversal Utilities for NAT (STUN) is a protocol that serves as a tool for other protocols in dealing with Network Address Translator (NAT) traversal. It can be used by an endpoint to determine the IP address and port allocated to it by a NAT. It can also be used to check connectivity between two endpoints, and as a keep-alive protocol to maintain NAT bindings. STUN works with many existing NATs, and does not require any special behavior from them. STUN is not a NAT traversal solution by itself. Rather, it is a tool to be used in the context of a NAT traversal solution. This is an important change from the previous version of this specification (RFC 3489), which presented STUN as a complete solution. This document obsoletes RFC 3489. [STANDARDS-TRACK] Traversal Using Relays around NAT (TURN): Relay Extensions to Session Traversal Utilities for NAT (STUN) If a host is located behind a NAT, then in certain situations it can be impossible for that host to communicate directly with other hosts (peers). In these situations, it is necessary for the host to use the services of an intermediate node that acts as a communication relay. This specification defines a protocol, called TURN (Traversal Using Relays around NAT), that allows the host to control the operation of the relay and to exchange packets with its peers using the relay. TURN differs from some other relay control protocols in that it allows a client to communicate with multiple peers using a single relay address. [STANDARDS-TRACK] Differentiated Services (Diffserv) and Real-Time Communication This memo describes the interaction between Differentiated Services (Diffserv) network quality-of-service (QoS) functionality and real- time network communication, including communication based on the Real-time Transport Protocol (RTP). Diffserv is based on network nodes applying different forwarding treatments to packets whose IP headers are marked with different Diffserv Codepoints (DSCPs). WebRTC applications, as well as some conferencing applications, have begun using the Session Description Protocol (SDP) bundle negotiation mechanism to send multiple traffic streams with different QoS requirements using the same network 5-tuple. The results of using multiple DSCPs to obtain different QoS treatments within a single network 5-tuple have transport protocol interactions, particularly with congestion control functionality (e.g., reordering). In addition, DSCP markings may be changed or removed between the traffic source and destination. This memo covers the implications of these Diffserv aspects for real-time network communication, including WebRTC. SMTP Service Extension for Secure SMTP over Transport Layer Security This document describes an extension to the SMTP (Simple Mail Transfer Protocol) service that allows an SMTP server and client to use TLS (Transport Layer Security) to provide private, authenticated communication over the Internet. This gives SMTP agents the ability to protect some or all of their communications from eavesdroppers and attackers. [STANDARDS-TRACK] SOCKS Protocol Version 5 This memo describes a protocol that is an evolution of the previous version of the protocol, version 4 [1]. This new protocol stems from active discussions and prototype implementations. [STANDARDS-TRACK] Hypertext Transfer Protocol (HTTP/1.1): Message Syntax and Routing The Hypertext Transfer Protocol (HTTP) is a stateless application-level protocol for distributed, collaborative, hypertext information systems. This document provides an overview of HTTP architecture and its associated terminology, defines the "http" and "https" Uniform Resource Identifier (URI) schemes, defines the HTTP/1.1 message syntax and parsing requirements, and describes related security concerns for implementations. TCP Control Block Interdependence This memo provides guidance to TCP implementers that is intended to help improve connection convergence to steady-state operation without affecting interoperability. It updates and replaces RFC 2140's description of sharing TCP state, as typically represented in TCP Control Blocks, among similar concurrent or consecutive connections. The Congestion Manager This document describes the Congestion Manager (CM), an end-system module that enables an ensemble of multiple concurrent streams from a sender destined to the same receiver and sharing the same congestion properties to perform proper congestion avoidance and control, and allows applications to easily adapt to network congestion. [STANDARDS-TRACK] Stream Control Transmission Protocol (SCTP) Stream Reconfiguration Many applications that use the Stream Control Transmission Protocol (SCTP) want the ability to "reset" a stream. The intention of resetting a stream is to set the numbering sequence of the stream back to 'zero' with a corresponding notification to the application layer that the reset has been performed. Applications requiring this feature want it so that they can "reuse" streams for different purposes but still utilize the stream sequence number so that the application can track the message flows. Thus, without this feature, a new use of an old stream would result in message numbers greater than expected, unless there is a protocol mechanism to "reset the streams back to zero". This document also includes methods for resetting the transmission sequence numbers, adding additional streams, and resetting all stream sequence numbers. [STANDARDS-TRACK] Stream Schedulers and User Message Interleaving for the Stream Control Transmission Protocol The Stream Control Transmission Protocol (SCTP) is a message-oriented transport protocol supporting arbitrarily large user messages. This document adds a new chunk to SCTP for carrying payload data. This allows a sender to interleave different user messages that would otherwise result in head-of-line blocking at the sender. The interleaving of user messages is required for WebRTC data channels. Whenever an SCTP sender is allowed to send user data, it may choose from multiple outgoing SCTP streams. Multiple ways for performing this selection, called stream schedulers, are defined in this document. A stream scheduler can choose to either implement, or not implement, user message interleaving.

API Mapping Template Any protocol mapping for the Transport Services API should follow a common template. Connectedness: (Connectionless/Connected/Multiplexing Connected) Data Unit: (Byte-stream/Datagram/Message) Connection Object: Initiate: InitiateWithSend: Ready: EstablishmentError: ConnectionError: Listen: ConnectionReceived: Clone: Send: Receive: Close: Abort: CloseGroup: AbortGroup:

Reasons for errors The Transport Services API allows for the several generic error types to specify a more detailed reason about why an error occurred. This appendix lists some of the possible reasons.

InvalidConfiguration: The transport properties and endpoints provided by the application are either contradictory or incomplete. Examples include the lack of a Remote Endpoint on an active open or using a multicast group address while not requesting a unidirectional receive.
NoCandidates: The configuration is valid, but none of the available transport protocols can satisfy the transport properties provided by the application.
ResolutionFailed: The remote or local specifier provided by the application can not be resolved.
EstablishmentFailed: The Transport Services system was unable to establish a transport-layer connection to the Remote Endpoint specified by the application.
PolicyProhibited: The system policy prevents the Transport Services system from performing the action requested by the application.
NotCloneable: The Protocol Stack is not capable of being cloned.
MessageTooLarge: The Message is too big for the Transport Services system to handle.
ProtocolFailed: The underlying Protocol Stack failed.
InvalidMessageProperties: The Message Properties either contradict the Transport Properties or they can not be satisfied by the Transport Services system.
DeframingFailed: The data that was received by the underlying Protocol Stack could not be processed by the Message Framer.
ConnectionAborted: The connection was aborted by the peer.
Timeout: Delivery of a Message was not possible after a timeout.

Existing Implementations This appendix gives an overview of existing implementations, at the time of writing, of Transport Services systems that are (to some degree) in line with this document.

Apple's Network.framework:
- Network.framework is a transport-level API built for C, Objective-C, and Swift. It a connect-by-name API that supports transport security protocols. It provides userspace implementations of TCP, UDP, TLS, DTLS, proxy protocols, and allows extension via custom framers.
- Documentation: https://developer.apple.com/documentation/network
NEAT and NEATPy:
- NEAT is the output of the European H2020 research project "NEAT"; it is a user-space library for protocol-independent communication on top of TCP, UDP and SCTP, with many more features, such as a policy manager.
- Code: https://github.com/NEAT-project/neat
- NEAT project: https://www.neat-project.org
- NEATPy is a Python shim over NEAT which updates the NEAT API to be in line with version 6 of the Transport Services API draft.
- Code: https://github.com/theagilepadawan/NEATPy
PyTAPS:
- A TAPS implementation based on Python asyncio, offering protocol-independent communication to applications on top of TCP, UDP and TLS, with support for multicast.
- Code: https://github.com/fg-inet/python-asyncio-taps