In computer networking
, the Transport Layer
is a group of methods and protocols within a layered architecture of network components, within which it is responsible for encapsulating application data blocks into datagrams
(packets) suitable for transfer to the network infrastructure for transmission to the destination host, or managing the reverse transaction by abstracting network datagrams and delivering them to an application. Such protocol are also referred to as "transport protocols".
Transport layers are contained in both the TCP/IP model (RFC 1122), which is the foundation of the Internet, and the Open Systems Interconnection (OSI) model of general networking. The definitions of the Transport Layer are slightly different in these two models. This article primarily refers to the TCP/IP model. See also the _Transport_layer.
The first transport protocol was the Transmission Control Protocol (TCP). It lent its name to the title of the entire Internet Protocol Suite, TCP/IP. It is used for connection-oriented transmissions, whereas the connectionless User Datagram Protocol (UDP) is used for simpler messaging transmissions. TCP is the more complex protocol, due to its stateful design incorporating reliable transmission. Other prominent protocols in this group are the Datagram Congestion Control Protocol (DCCP) and the Stream Control Transmission Protocol (SCTP).
Transport Layer functions
The Transport Layer is responsible for delivering data to the appropriate application process on the host computers. This involves statistical multiplexing
of data from different application processes, i.e. forming data packets, and adding source and destination port numbers in the header of each Transport Layer data packet. Together with the source and destination IP address, the port numbers constitutes a network socket
, i.e. an identification address of the process-to-process communication. In the OSI model, this function is supported by the Session Layer
Some Transport Layer protocols, for example TCP, but not UDP, support virtual circuits, i.e. provide connection oriented communication over an underlying packet oriented datagram network. A byte-stream is delivered while hiding the packet mode communication for the application processes. This involves connection establishment, dividing of the data stream into packets called segments, segment numbering and reordering of out-of order data.
Finally, some Transport Layer protocols, for example TCP, but not UDP, provide end-to-end reliable communication, i.e. error recovery by means of error detecting code and automatic repeat request (ARQ) protocol. The ARQ protocol also provides flow control, which may be combined with congestion avoidance.
UDP is a very simple protocol, and does not provide virtual circuits, nor reliable communication, delegating these functions to the application program. UDP packets are called datagrams, rather than segments.
TCP is used for many protocols, including  web browsing and email transfer. UDP may be used for multicasting and broadcasting, since retransmissions are not possible to a large amount of hosts. UDP typically gives higher throughput and shorter latency, and is therefore often used for real-time multimedia communication where packet loss occasionally can be accepted, for example IP-TV and IP-telephony, and for online computer games.
In many non-IP-based networks, for example X.25, Frame Relay and ATM, the connection oriented communication is implemented at network layer or data link layer rather than the Transport Layer. In X.25, in telephone network modems and in wireless communication systems, reliable node-to-node communication is implemented at lower protocol layers.
The OSI/X.25 protocol suite defines five classes of the OSI transport protocol, ranging from class 0 (which is also known as TP0 and provides the least error recovery) to class 4 (which is also known as TP4 and is designed for less reliable networks, similar to the Internet).
Transport Layer services
There is a long list of services that can be optionally provided by the Transport Layer. None of them are compulsory, because not all applications require all available services.
- Connection-oriented: This is normally easier to deal with than connection-less models, so where the Network layer only provides a connection-less service, often a connection-oriented service is built on top of that in the Transport Layer.
- Same Order Delivery: The Network layer doesn't generally guarantee that packets of data will arrive in the same order that they were sent, but often this is a desirable feature, so the Transport Layer provides it. The simplest way of doing this is to give each packet a number, and allow the receiver to reorder the packets.
- Reliable data: Packets may be lost in routers, switches, bridges and hosts due to network congestion, when the packet queues are filled and the network nodes have to delete packets. Packets may be lost or corrupted in Ethernet due to interference and noise, since Ethernet does not retransmit corrupted packets. Packets may be delivered in the wrong order by an underlying network. Some Transport Layer protocols, for example TCP, can fix this. By means of an error detection code, for example a checksum, the transport protocol may check that the data is not corrupted, and verify that by sending an ACK message to the sender. Automatic repeat request schemes may be used to retransmit lost or corrupted data. By introducing segment numbering in the Transport Layer packet headers, the packets can be sorted in order. Of course, error free is impossible, but it is possible to substantially reduce the numbers of undetected errors.
- Flow control: The amount of memory on a computer is limited, and without flow control a larger computer might flood a computer with so much information that it can't hold it all before dealing with it. Nowadays, this is not a big issue, as memory is cheap while bandwidth is comparatively expensive, but in earlier times it was more important. Flow control allows the receiver to say "Whoa!" before it is overwhelmed. Sometimes this is already provided by the network, but where it is not, the Transport Layer may add it on.
- Congestion avoidance: Network congestion occurs when a queue buffer of a network node is full and starts to drop packets. Automatic repeat request may keep the network in a congested state. This situation can be avoided by adding congestion avoidance to the flow control, including slow-start. This keeps the bandwidth consumption at a low level in the beginning of the transmission, or after packet retransmission.
- Byte orientation: Rather than dealing with things on a packet-by-packet basis, the Transport Layer may add the ability to view communication just as a stream of bytes. This is nicer to deal with than random packet sizes, however, it rarely matches the communication model which will normally be a sequence of messages of user defined sizes.
- Ports: (Part of the Transport Layer in the TCP/IP model, but of the Session Layer in the OSI model) Ports are essentially ways to address multiple entities in the same location. For example, the first line of a postal address is a kind of port, and distinguishes between different occupants of the same house. Computer applications will each listen for information on their own ports, which is why you can use more than one network-based application at the same time.
Comparison of principal transport protocols
- AEP, AppleTalk Echo Protocol
- ATP, AppleTalk Transaction Protocol
- CUDP, Cyclic UDP
- DCCP, Datagram Congestion Control Protocol
- FCIP, Fiber Channel over TCP/IP
- FCP, Fiber Channel Protocol
- IL, IL Protocol
- iSCSI, Internet Small Computer System Interface
- LCP (in PPP), Link Control Protocol
- NBF, NetBIOS Frames protocol
- NBP, Name Binding Protocol
- RTMP, Routing Table Maintenance Protocol
- SCSI, Small Computer System Interface
- SCTP, Stream Control Transmission Protocol
- SPX, Sequenced Packet Exchange
- SST, Structured Stream Transport
- TCP, Transmission Control Protocol
- UDP, User Datagram Protocol
- UDP Lite