Multimedia Communication & Transport Layer Protocols

0 | P a g e

RESEARCH PAPER

MULTIMEDIA COMMUNICATION AND TRANSPORT LAYER

PROTOCOLS

MWAMI MUGALA

1 | P a g e

INTRODUCTION

Computer networking has become the core centre of computing across many

platforms of devices; mobile phones, game consoles, tablets and computers in

general. The need to communicate using audio and video has been widely used and

is growing. This is often referred to as multimedia networking.

For communication to be complete, multimedia applications or any applications require

the services that are provided by the transport layer. The transport layer cannot be by

passed. This layer is an interface between the Application layer and the Network layer

of the TCP/IP protocol stack. The transport layer has protocols that play a huge role

in offering the necessary services for process to process delivery to be successful.

The protocols are responsible for the delivery of packets from the source to destination

application.

JUSTIFICATION

Multimedia communication refers to the sending and receiving of audio and video

traffic on a network. It can involve real-time interactive or conversational, stored media,

and streaming live audio/audio.

Multimedia data traffic can be sensitive to delay, jitter and is at some level tolerant to

packet loss. This traffic poses new requirements to the internet architecture and its

protocols. Network layer has the Internet Protocol (IP) which is a best effort protocol.

It only makes every effort to deliver the packets without any guarantees.

To achieve the desired quality of service required by multimedia traffic, application

developers need to carefully select which transport layer protocol to deploy for their

application.

The transport layer has three major protocols; Transmission Control Protocol (TCP),

User Datagram Protocol (UDP) and Stream Control Transmission Protocol.

2 | P a g e

MULTIMEDIA APPLICATIONS AND THE OSI LAYERS

All networking begins at the application layer which is the first layer of the TCP/IP. The

application layer is the starting point of all communication sessions and defines the

type of messages to be exchanged between the sender and receiver of a particular

application process. It also allows human interaction with the actual application to be

used for communication.

After the application layer has played it role, the packets are passed down to the

presentation layer. At this stage, offers encoding and decoding of data. This ensures

that the correct data format is encoded and sent to the destination so that the right

decoding system is used to read the data. In the case of Multimedia applications;

various video and audio formats exist. It is essential that the correct encoding scheme

is used. Other functions the presentation layer offers is data compressing and

encryption

The session layer comes after the presentation layer and it allows for the control of

dialogs or sessions.

THE RELATIONSHIP BETWEEN THE TRANSPORT LAYER AND MULTIMEDIA

APPLICATIONS

The transport layer resides between the application and network layers of the TCP

protocol stack and serves as an intermediary.

All applications running on a network must pass the data packets to the transport layer

to provide process to process communication. Among other things, the transport layer

is responsible for identifying the sending and receiving applications; this ensures that

the correct application gets the data that was intended for it. For this, the transport

layer uses port number addressing to uniquely identify the application or process.

3 | P a g e

However, the transport layer has protocols which each multimedia application should

deploy. These protocol offer different services, with some being conducive for

multimedia networking and some having a poor performance on the quality of service.

After the transport layer has played its role, the packets are handed over to the

Network layer. The network layer has the Internet Protocol which offers a best effort

delivery service.

QUALITY OF SERVICE REQUIREMENTS FOR MULTIMEDIA TRAFFIC

Multimedia applications differ from the regular elastic applications like e-mail, Web and

file sharing. The challenges in multimedia network arise due to the distinct

characteristics of audio and video data. Most multimedia applications are highly

sensitive to end-to-end delay and jitter but can tolerate occasional loss of data. The

traffic flow of this data needs to be continuous. Any delays or breaks in the flow can

have a negative impact on the user experience.

Delay and jitter are also very important in multimedia networking. Multimedia

applications have a maximum time required from the time the packets are transmitted

to the time they reach the destination; this is the delay time. Delay should be very

minimal and is measured in milliseconds (ms). Jitter on the other hand, is the measure

of smoothness of the audio/video. It is the variance in packet delays.

End to end delay can be caused by queues and buffers created in the routers. End-

to-end delay can be sub divided into various categories, and these are; propagation,

processing and queueing delays.

The quality of service required by multimedia data is regulated by many parameters.

Bandwidth is one of the most important. Bandwidth refers to the amount of information

that can be transmitted over a network in a given amount of time, usually expressed

in bits per second or bps [1.]

The traffic generated by these application requires more bandwidth than the traditional

applications. In terms of quality of service, metrics such as available bandwidth,

4 | P a g e

transfer delay, and packet loss should be within the acceptable ranges for multimedia

traffic.

The data rates in these applications is high and often has a lower bound to ensure

continuous playback. The user should be able to play the audio or video with the

content being fully downloaded. This is called continuous or streaming media. The

traffic generated by multimedia applications is of a higher bit rates as compared to

other usual network applications [2.]

In terms of real-time and interactivity, these applications demand a low start up delays

and synchronization between audio and video.

CLASSIFICATION OF MULTIMEDIA NETWORK APPLICATIONS

Multimedia applications differ in the way the traffic is generated, transmitted and

received at the destination. Each class has its own set of service requirements.

Multimedia network applications can be divided into three classes with regard to the

timeliness requirements they possess; Stored information, real-time interactive and

streaming live audio/video.

Stored Audio/Video Multimedia

The first class of multimedia application is streaming of stored audio and/or video.

These applications consist of pre-recorded audio/video which is stored on servers and

are available to the clients. Streaming stored audio/video refers to on-demand

requests for compressed files [3.]

Streaming involves the client application receiving the content in parts while playing it

continuously. This media become interactive when the user is able to pause and fast-

forward through the video content. The user experience should be continuous video

and audio play without any breaks or delays in receiving the content.

Stored multimedia is implemented in broadcasting audio and video and Video On

Demand (VOD.) because they do not have to deal with interactive communication,

5 | P a g e

stored multimedia is less demanding when it comes to delay requirements as

compared to real-time interactive and living streaming multimedia traffic.

Real-Time Interactive Audio/Video Multimedia

Real-time interactive multimedia has conversational voice and video capabilities.

These conversational systems enable users to conference with three or more parties.

The interactivity allows users to communicate with others in real-time. This category

of applications allow a bidirectional traffic to flow in both ways simultaneously. An

example of this is Voice Over Internet Protocol (VOIP) and Skype. These applications

are very sensitive to jitter or the delay between packets.

Streaming Live Audio/Video

This is similar to broadcasting of radio and Television, but it is done via the internet.

Streaming live media and stored media are both sensitive to delay and do not accept

retransmissions. However, store media is on-demand and unicast while live streaming

is multicast and is live [4.]

THE TRANSPORT LAYER PROTOCOLS AND SERVICES

Transport layer protocols have features that can be of benefit or cause problems to

the performance of some types of multimedia applications.

The transport layer provides a logical communication between applications or

processes running on different machines. This process to process means that the

transport layer is in charge of the delivery of packets between the actual applications.

Port Numbering

To identify these applications, the transport layer deploys port numbers to uniquely

identify the sending and receiving applications or processes. Every TCP and UDP

connection has a source and a destination port. When a service starts, a program

listens on a port and informs the operating system it wishes to receive data directed

to that port [5.]

6 | P a g e

Port numbers are in three categories; well-known port numbers, registered port

numbers and dynamic port numbers. Well-known port numbers range from 0 to 1,023

and these are assigned and controlled. Registered port numbers range from 1024 to

49,151 and are not assigned or controlled. Dynamic Port numbers are from 49,152 to

65, 535 and they are used temporary for private port numbering.

Encapsulation and Decapsulation

When a packet is processed at the transport layer, additional information such as

socket addresses is added to the packet header and this is called encapsulation.

This takes place at the sender side.

At the receiver side, the data that was added during the encapsulation process is

read and the packet can be passed on to the appropriate process. This is called

decapsulation.

TRANSPORT LAYER PROTOCOLS (TCP, SCTP UDP)

The transport layer has three protocols that provide different services. The widely used

protocol is the Transmission Control Protocol (TCP). The second protocol is User

Datagram Protocol (UDP). The third protocol is Stream Control Transmission Protocol

(SCTP) and it has features of both TCP and UDP.

TRANSMISSION CONTROL PROTOCOL (TCP)

TCP has a lot of service to offer but some are not required by multimedia applications.

These services add processing delays of packets and other overheads which can

increase jitter cause problems for multimedia applications.

Connection-Oriented

TCP is connection-oriented protocol. This means that before data is exchange, there

must be a connection establishment to be made between the sender and the receiver.

Establishing a connection involves a process known as the Three-Way-Handshake.

The sender begins by sending an ACK (Acknowledgement) packet to the receiver

requesting for a connection to be established. After the receiving the ACK packet, the

7 | P a g e

receiver sends back a SYN-ACK (Synchronization Acknowledgement) packet which

tells the sender that a connection can begin. Lastly, the sender receives the SYN-ACK

and sends back an ACK and then data transfer can commerce.

For a connection to be terminated, TCP has the Four-Way process to close a

connection. Either the sender or the receiver can initiate the termination of the

connection. It starts with sending a Fin Flag to the party. The other part then sends

back an ACK of Fin and also sends a Fin flag back to the connection termination

initiator. The initiated send a final ACK of Fin to completely terminate the connection.

This process of establishing and terminating a connection can cause unnecessary

delays for multimedia applications.

Flow Control

TCP’s flow control mechanisms prevents the discarding of packets. Flow control

creates a balance between production and consumption of packets [6.] Flow control

uses buffering at the receiver’s end to hold packets pending to be processed.

However, when these buffers are packets are discarded. Multimedia applications like

real-time interactive video/audio streaming do not need buffers to store the packets;

as this can cause unnecessary delays. The packets should be processed as soon as

they get to the destination node.

Error Control

TCP is a reliable protocol. This means that it ensures that packets are delivered to the

destination and are undamaged. This is achieved via the error control services in

which TCP detects and discards corrupt packets, keeps track of lost and discarded

packets and resends them. Multimedia applications have a tolerance to packet loss

and resending packets can have a negative effect on the streaming of real-time

packets [7.]

Error control can be very problematic when it comes to streaming media on a network.

TCP uses a retransmission mechanism to resend the packets that have not been

acknowledged at destination host. This process can be complex and cause delays on

8 | P a g e

the network while making the resent packets unacceptable when it comes to real-time

interactive media because it increases end-to-end delay

Congestion Control

Congestion control mechanisms are deployed by TCP to share bandwidth resources

on the internet [8.]

Congestion is a situation where the number of packets sent to the network id greater

than the capacity the network can handle. TCP congestion control regulates and keeps

the load below the capacity. Like flow control, congestion control deploys buffers at

routers and switches. This simply means that the packets have to be stored or delayed

in the buffer.

TCP congestion control reduces the rate at the sender to a lower rate that the drain

rate at the receiver. This has a severe impact on voice intelligibility at the receiver.

STREAM CONTROL TRANSMISSION PROTOCOL (SCTP)

Stream Control Transmission Protocol is a hybrid of TCP and UDP and it offers

features of both these protocols. It is a reliable, message oriented transport layer

protocol. However, it is a more sophisticated protocol compared to TCP and UDP.

SCTP is a multiple stream oriented protocol; meaning that multiple streams are

allowed in each connection, and these streams are called associations.

USER DATAGRAM PROTOCOL (UDP)

UDP is a simple protocol that only offers basic features and hence adds very little

overhead during packet processing.

This protocol is connectionless protocol and does not require a connection to be

established for data transfer to begin. Apart from this, the packets are not numbered;

UDP does not track down lost or delayed packets. In other words, UDP offers an

unreliable service in which packets are not guaranteed of delivery. The protocol does

9 | P a g e

not offer error control; the packets lost or corrupted during transmission are not

required to be retransmitted. Multimedia applications like real-time video/audio

streaming have a tolerance of packet loss. It can be of no use to retransmit packets

lost in a live multimedia session

It relies on the application layer to offer a reliability of packet delivery in case the

service is required.

UDP does not offer congestion control mechanisms. Real-time applications react very

poorly to congestion control. For this reason, developers of multimedia applications

often choose to run their applications over UDP instead of TCP [9]

The UDP frame is smaller and the header has a fixed size of 8 bytes, while TCP has

20 bytes. This makes processing of packets very fast. As soon as an applications

processes the data, it passes it to the transport layer via the UDP protocol, which will

package the data inside the UDP segment and immediately pass the segment to the

network layer without delay [10.]

UDP supports multicasting and this feature is very cardinal in multimedia networking.

This simply means that the data on a network segment is multiplied and sent to a

group of clients. It is applied in applications where there is need to broadcast a stream

of media to more than one user.

For multimedia streaming applications to perform accordingly, there should be very

minimal or no delays during packet processing. Both TCP and SCTP are connection

oriented protocol, with flow control, error and congestion control services slows down

packet processing.

How UDP Covers For The Missing Services

Multimedia applications need limited delay and jitter that is offered by UDP, but require

more transport layer services that are not offered by UDP.

10 | P a g e

UDP does not offer congestion control, but this is needed to prevent the network from

entering a congested state in which very little useful work can be done. UDP relies on

the application layer to offer this service and other required services.

Apart from depending on the application layer to offer additional services, UDP

deploys other protocols such as Real Time Protocols (RTP) to provide more services.

These protocols provide services for data with real-time characteristics.

Real-Time Transport Protocol

RTP is an IP protocol that supports real-time transmission of voice and video over the

internet. It is used for media on-demand as well as interactive services such as IP

telephony and audio/video streaming [11.]

RTP usually runs on top of UDP and resides in between the transport layer and the

application layer, and bridges them for real-time multimedia transmission [12.]

RTP has two other protocols that allow it to offer the services needed by multimedia

applications. These protocols are; Real-time Transmission Control Protocol (RTCP)

and Real-time Streaming Protocol (RTSP.)

Multimedia applications use different encoding schemes and each of these have

various trade-offs between quality, bandwidth and computational cost. RTP is

responsible to negotiate and select which scheme the sending and receiving

applications should deploy. RTP adds the payload type in the header of each packet.

This indicates the media data type and the encoding scheme. This allows the receiver

to know how to decode the data.

For applications to perform well, the jitter caused by routers and other network

constrains should be handled. Jitter changes the original spacing between the packets

and this cause the packets not to play at the right time. RTP is responsible to sort out

this spacing and advises the receiver about the timing relationships among the

received packets. To achieve this, RTP deploys timestamps. The timestamps records

the instant when the first part of the packet was sampled and is set by the sender. The

11 | P a g e

receiver can then play the audio/video in the correct timing order and synchronize

multiple streams when required.

Multimedia data packets, like any other packets, required to be delivered in an orderly

manner. However, UDP does not offer this facility. On the other hand, RTP offers the

sequence numbers for the packets. This allows the receiver to ensure that the packets

can be reconstructed in order by the receiver and also to know which packet comes

first and last or in case of buffering, the packets can be stored and played according

the sequence number given. Sequence numbers also allow the completion of the

function of time stamping.

RTP Control Protocol (RTCP)

Being the control part of RTP, RTCP is responsible for monitoring the quality of service

of multimedia application by providing feedback to the source on the quality of data

transmission.

RTCP delivers quality of service feedback for media distribution through recurrent

statistical data transmission to streaming multimedia contributors [13.] This feedback

data contains information such as; packet loss, jitter, and round-trip delay time. Quality

of service come into play when the application receives this information and changes

its parameters by regulating packet flow or using a different encoding scheme. RTP

and RTCP protocols provide a controlled delivery of multimedia traffic over the

internet. The protocols focus is to provide out-of-band statistics and control information

for RTP sessions.

The RTCP packets contain direct information for quality of service monitoring. There

are five types of RTCP packets, these are; Receiver Reports (RR), Sender Reports

(SR), Sender Description (SDES), Goodbye (BYE) and Application-defined Packet

(APP) [14.]

The Sender Reports and Receiver Reports packets provide information regarding the

packet losses, delay and jitter. Using adaptive encoding, these types of packets can

be used to deploy a flow control mechanism resembling TCP at the application level.

The Source Description provides information about the source, this includes the email

address, phone number, and full name of the source.

12 | P a g e

The BYE packet indicates the end of participation while the Application-defined Packet

provide for future extension of new features.

To keep track of each session, RTCP deploys the transport-level identifier known as

the canonical name.

Real Time Streaming Protocol (RTSP)

Although RTSP is an application layer protocol, it plays a major role in providing

service for multimedia applications. This protocol controls multimedia servers and

streams delivery. RTSP is used for establishing and controlling media sessions

between client and server. In other words, RTSP sets up and controls streams of

continuous audio and video media. This protocol work hand in hand with RTP, RSVP

and RTCP to offer client services that facilitate real-time control of playback such as

play, random-seek, pause multimedia files from the server and also offer a complete

streaming server over a network.

The Real Time Streaming Protocol establishes and controls either a single or several

time-synchronized streams of continuous media such as audio and video [15.]

There are four main steps in creating an RTSP connection. In the first step, the client

sends a Describe request to the server. This requests the server to issue out

information regarding the media type (image, video or audio), frame rate, codec and

other parameters. The description request is sent via HyperText Transfer Protocol

(HTTP) or any other means available.

In the second step, the client send a Setup request to the server. This simply tells the

server of the destination IP address, port number, protocols and TTL. The session is

only setup when the server sends back a session ID to the client. This identifier is what

the server uses to maintain the connection.

The third step involves requesting and receiving the actual media. At this stage the

server transmits the content using RTP. The client can then store, play, pause, fast-

forward and rewind the media. The client sends feedback RTCP packets to the server

to offer quality of service information.

13 | P a g e

The fourth and last step is the session closure via the teardrop packet. On the client

is done, it sends this packet to the server requesting to terminate the session.

RSTP is responsible for setting up conferencing communication. A media server can

be invited to join an existing conference where there are more than three media

servers participating in the session. This includes both audio and video conferencing

applications such as skype.

RTSP is able to keep track of the concurrent communication sessions using a state

identifier and for this reason it is a statefull protocol.

Secure Real-time Transport Protocol (SRTP)

UDP, RTP and RTCP do not provide any security or authentication. This job is given

to the SRTP. This protocol provides encryption, message authentication and

integrity, and replay protection to the RTP data [16.]

Reservation Protocol (RSVP)

RSVP is a communication protocol that signals a router to reserve bandwidth for real

time transmission. This protocol is used by the client to request specific quality of

service from the network for particular application data streams or flows. [17]

14 | P a g e

CONCLUSIONS

Multimedia network applications generate traffic that is complex and requires the appropriate transport layer to be deployed. The quality of service requirements differ according to the type of multimedia class.

The transport layer plays a huge role in networking and can affect the end user experience of multimedia applications. However, from the protocols offered in the transport layer, UDP proves to be the best choice. It does not provide a reliable connection and no flow-control mechanisms and the lack of these features is an advantage to multimedia data and traffic.

The protocol’s ability to support multicasting allows the broadcasting a single video to more than one client. This is very essential for multimedia networking. TCP and SCTP do not support this very important feature.

It is evident that TCP and SCTP have more features as compared to UDP. But this does not make them to be the best choice to be deployed in multimedia networking. Even though UDP provides very few services, it creates a near perfect environment for multimedia traffic. It relies on RTP and other protocols in the application layer to deliver the services required.

RTP has no way of ensuring the delivery of packets or quality of service, but it work together with RTCP to provide a way of monitoring quality of service and can manipulate their parameters accordingly.

15 | P a g e

REFERENCE

[1] Vicomsoft

URL: http://www.vicomsoft.com/glossary/bandwidth/

[2] Yevgeni Koucheryavy, Multimedia Traffic Delivery over Next-Generation

Telecommunications Networks. Tampere University of Technology.

[3, 4] Behrouz A. Forouzan,Data (2008) Communications and Networking, Fourth Edition,

The McGrew-Hill Companies.

[5] The Transport Layer, University of Bath, Department of Computer Science.

URL: http://www.cs.bath.ac.uk/ag/CM30078-50123/09.pdf

[6] New Paltz State University of New York, TCP Connection Termination,

URL: http://www.cs.newpaltz.edu/~easwaran/CCN/Week6/tcpconnectiontermination.pdf

[7, 8] Behrouz A. Forouzan,Data (2010) TCP/IP Protocol Suite, Fourth Edition, The

McGrew-Hill Companies.

[9] White Paper: TCP Congestion Control with a Misbehaving Receiver

Stefan Savage, Neal Cardwell, David Wetherall, and Tom Anderson

University of Washington, Seattle.

URL: http://cseweb.ucsd.edu/~savage/papers/CCR99.pdf

[10, 11] Jim Kurose and Keith W. Ross, (2014) Computer Networking - A Top Down

Approach. 6th Edition. Pearson Education.

[12] Tommi Koistinen, RTP and RTCP, Nokia Telecommunications.

[13] Ze-Nian Li, Mark S. Drew, Jiangchuan Liu (2004) Fundamentals of Multimedia.

Simon Fraser University. Pearson Education International.

[14] Techo Pedia, RTP Control Protocol

URL: http://www.techopedia.com/definition/25807/rtp-control-protocol-rtcp

[15] Zoran Bojkovic, Bojan Bakmaz, Miodrag Bakmaz, Multimedia Traffic in New

Generation Networks: Requirements, Control and Modeling. University of Belgrade

URL:http://www.wseas.us/elibrary/conferences/2009/rodos/COMMUNICATIONS/CO

MMUNICATIONS19.pdf

[16] SRTP

URL: http://en.wikipedia.org/wiki/Secure_Real-time_Transport_Protocol

[17] Colombia University, RTSP

URL http://www.cs.columbia.edu/~hgs/rtsp/