TRANSPORTATION

LAYER
KEVIN LEONARDO BARBOSA MORENO
NETWORKING FUNDAMENTALS
JAVIER DAZA PIRAGAUTA
 Introduction…

 The transport layer is the fourth layer in the open system interconnection (OSI)
model, and is responsible for end-to-end communication over a network.
 Layer in charge of transporting the data (inside the packet) from the source
machine to the destination machine, independent of the type of physical network
you are using.
 Transport Layer…

 It provides logical communication between application processes running on

different hosts within a layered architecture of protocols and other network
components.
 The transport layer is also responsible for the management of error correction,
providing quality and reliability to the end user.
 This layer enables the host to send and receive error corrected data, packets or
messages over a network and is the network component that allows multiplexing.
 Transport Service Primitives…

 To allow users to access the transport service, the transport layer must provide
some operations to the application programs, that is, an interface of the transport
service. Each transport service has its own interface.
 The primitives of a simple transport would be:
 Listen: It is blocked until some process tries the contact.
 Connect: Actively try to establish a connection.
 Send: Send information.
 Receive: Locks until a Data TPDU arrives.
 Disconnect: This side wants to release the connection.
 Elements of Transport Protocole…

 The transport service is implemented using a transport protocol between two

transport entities. In some respects, transport protocols resemble network
protocols. Both are responsible for error control, sequencing, and flow control.
Addressing…
 When a process wants to establish a connection to a remote application process, it
must specify which one it will connect to. (Who do I send the message to?) The
method normally used is to define transport addresses in which the processes can
listen for connection requests.
 On the Internet, these endpoints are called ports, but we will use the generic term
of TSAP (Access Point for Transportation Service). Analog network layer
endpoints are called NSAP (Network Service Access Point). IP addresses are
examples of NSAPs.
 Establishing a connection ...
Establishing a connection seems easy, but it's actually surprisingly difficult. At first glance, it would seem that it is
enough to send a TPDU (Transport Protocol Data Unit) with the connection request and wait for the other to accept
the connection. The problem comes when the network can lose, store, or duplicate packets. The main problem is the
existence of delayed duplicates. This can be solved in several ways (none is very satisfactory).
Releasing a connection ...
Releasing a connection is easier than establishing it. However, there are more pitfalls than one might imagine. There
are two termination styles for a connection: asymmetric release and symmetric release. Asymmetric release is the
way the telephone mechanism works: when a party hangs up, the connection is interrupted. Symmetric release treats
the connection as two separate unidirectional connections, and requires each to be released separately.
 Flow control and buffering ...
One of the key aspects is flow control. We need a scheme to prevent a fast emitter from overflowing a slow receiver.
The main difference is that a router typically has relatively few lines, and a host can have numerous connections.
This difference makes it impractical to use the implementation that is done at the link layer.
Multiplexing ...
Multiplexing multiple conversations across connections, virtual circuits, or physical links plays an important role at
different layers of the network architecture. The need for multiplexing may arise in the transport layer for various
reasons. For example, if only one network address is available on a host, all transport connections from that machine
will have to use it. When a TPDU arrives, some mechanism is needed to know which process to assign it to.
 Recovery from falls ...
If hosts and routers are subject to crashes, recovery is critical. If the transport entity is entirely within the hosts,
network and router crash recovery is straightforward. If the network layer provides datagram service, transport
entities expect loss of some TPDUs all the time, and they know how to handle it. If the network layer provides
connection-oriented service, then the loss of a virtual circuit is handled by establishing a new one and polling the
remote transport entity to know which TPDUs it has received and which it has not.
 Internet Transport Protocols…

 User Datagram Protocol (UDP) and TCP are the basic transport-level protocols
for making connections between Internet hosts. Both TCP and UDP allow
programs to send messages and receive messages from applications on other
hosts. When an application sends a request to the transport layer to send a
message, UDP and TCP break the information down into packets, add a packet
header that includes the destination address, and send the information to the
network layer for further processing. Both TCP and UDP use protocol ports on the
host to identify the specific destination of the message.
 User datagram protocol
Sometimes an application on one network needs to send messages to a specific application or process
on another network. UDP provides a datagram means of communication between applications on
Internet hosts.
Reliable Datagram Sockets through InfiniBand and RoCE
Reliable Datagram Sockets (RDS) is a connectionless, record-oriented protocol that provides an order-
free, duplication-free service through InfiniBand and RDMA over Converged Ethernet (RoCE). RDS
exposes the Socket API User Datagram Protocol (UDP) subset.
The Transmission Control Protocol
TCP provides reliable data transmission between Internet hosts.
TCP Header Field Definitions Next, we will
Briefly describe each of the Transmission Control Protocol (TCP) fields.
 Transport Protocols…

 The Internet provides a way to obtain packets (convenient data drives for
computers and routers) from any host computer to one or more host computers.
However, network protocols do not guarantee the delivery of a package. In fact, a
package may be lost, it may arrive after others send it later, or it may become
distorted. There might even be a package that just didn't ship!

 To counter this, host computers incorporate transport protocols, which use the
Internet to transport application information, but also send a variety of other
information to provide verification and correction or recovery of such errors.
 The User Datagram Protocol or UDP…
UDP is a `` send and forget '' protocol. It provides enough control information at the beginning of each package to
know which application is running and to check if the package was distorted in the path. UDP is used by applications
that don't normally require a response, and they really don't care if the other end received the message. A typical
example of this could be a server that advertises time on the network, unsolicited.

The reliable data protocol, or RDP…

RDP is a generic name for a collection of protocols; the most relevant here is the one used by Prospero (see below).
RDP-type protocols are similar to TCP, but with reduced complexity at the start and end of a conversation, and with
good support for data fragment exchange sequences, often known as Remote Procedure Calls, or sometimes calls
transactions incorrectly.

Transmission control protocol or TCP…

TCP is the protocol module that provides reliability and security. TCP is designed to cope with the full range of
network failures and elegantly accommodates the resources available on the network. It even tries to be fair to all
users.
 Network Performance…

 Network performance refers to the quality of service measures of a network as

seen by the customer.

 There are many different ways to measure the performance of a network, as each
network is different in nature and design. Performance can also be modeled and
simulated instead of measured; An example of this is using state transition
diagrams to model queue performance or to use a network simulator.
 Performance Measures…
 Bandwidth commonly measured in bits / second is the maximum rate of information transfer.
 Throughput is the actual rate of information transfer.
 Latency the delay between the sender and the receiver that decodes it, this is mainly a function of the travel time
of the signals and the processing time at any node that passes through the information
 Jitter variation in packet delay at information receiver
 Error rate the number of corrupted bits expressed as a percentage or fraction of the total sent
Bandwidth…
 Available channel bandwidth and achievable signal-to-noise ratio determine the maximum performance possible.
In general, it is not possible to send more data than dictated by the Shannon-Hartley theorem.
Performance…
 Throughput is the number of messages successfully delivered per unit of time. Performance is controlled by
available bandwidth, as well as available signal-to-noise ratio and hardware limitations. The performance for the
purposes of this article will be understood as measured from the arrival of the first bit of data to the receiver, to
decouple the concept of performance from the concept of latency. For discussions of this type, the terms
'performance' and 'bandwidth' are often used interchangeably.
 Latency…
The speed of light imposes a minimum propagation time on all electromagnetic signals. Latency cannot be reduced
below
where s is the distance and c m is the speed of light in between. This means approximately 1 millisecond additional
round trip time (RTT) for 100 km / 62 miles of distance between hosts.
Jitter…
Jitter is the unwanted deviation from the actual periodicity of an assumed periodic signal in electronics and
telecommunications, often relative to a reference clock source. The fluctuation can be observed in characteristics
such as the frequency of successive pulses, the amplitude of the signal or the phase of the periodic signals. Jitter is an
important, and generally unwanted, factor in the design of almost all communication links (eg USB, PCI-e, SATA,
OC-48). In clock recovery applications it is called sync jitter. [one]
Error Rate…
In digital transmission, the number of bit errors is the number of bits received from a data stream through a
communication channel that have been altered due to noise, interference, distortion, or bit synchronization errors.
Interaction of factors…
All of the above factors, along with user requirements and perceptions, play a role in determining the perceived
"strength" or utility of a network connection. The relationship between performance, latency, and user experience is
best understood in the context of a shared network medium and as a scheduling issue.
 Conclusion…

 The transport layer takes application messages and transmits those message
segments into Layer 3, the networking layer. Once the receiving side has the
segments, they are reassembled into messages and passed on to Layer 7, the
application layer.
 Web Graphy…

 https://www.techopedia.com/definition/9760/transport-layer
 https://sites.google.com/site/sabyrodriguezgamez/unidad2/2-2-capa-de-transporte
 https://www.ecured.cu/Capa_de_transporte
 https://www.cl.cam.ac.uk/~jac22/books/www/book/node21.html
 https://www.ibm.com/support/knowledgecenter/en/ssw_aix_72/network/protocols
_transp_level.html
 https://en.wikipedia.org/wiki/Network_performance