Factors Affecting Quality Of Service Of Voip Information Technology Essay

This project proposal offers the thoughtful of Quality of Service using the Voice over IP network, which allows you to understand that how to improve quality of service and Mean Opinion Score (MOS) of voice signal. The project proposal explain what are the factors that affect the quality of voice signal when travelling over the IP network and, which are methods or techniques that can be used as solution of quality of service.

Introduction

To many, Quality of service (QoS) is one the major, if not the largest issue facing Voice over IP (VoIP). This is the reason why many of the VoIP solutions on the market today that provide voice over the Internet are giving free services by sending voice over the Internet and not worrying about the quality of service. Quality of Service for IP traffic is a difficult topic to characterize in today’s networking environment. There is no even a single standard for implementing end- to- end QoS on IP internets, so many sellers and standard organizations are busy creating fresh protocol to fill the void. The opportunities to take a big taste can be even greater if a network can be implemented at a lower cost than traditional circuit switching and if voice and data services can be integrated to offer a more complete service package. VoIP offers this opportunity. IP is compatible to non-real time communication and can be found for the QOS issue. Fortunately, solutions are available and are being developed.

In designing a VoIP network that transport voice over packet network, frame, or cell infrastructure, it is important to consider all the major factors like codec, frame loss, echo cancellation, and delay, which affect voice quality. Of course there are some standards required to get better quality of voice, which includes [1][2]

To avoid audible errors, codec requires packet loss far less than a percent.

The International Telecommunication Union (ITU) considers less than 300 millisecond (ms) two way end -to- end delay for high quality real time traffic.

Jitter buffers, which is used to compensate for varying delay only effective on delay variations less than 100 ms.

Internet Service Providers, Internet Telephony Service Providers and Public Switched Telephone Network are implementing VoIP because it’s including toll bypass, network consolidation and service convergence. Like other real-time application, VoIP is extremely bandwidth – and delay sensitive. So designing a VoIP network requires careful planning to ensure that voice quality can be properly maintained. In other words, VoIP have a guarantee high-quality voice transmission only if the voice packets are given priority over network traffic. For successful VoIP transmission to the receiver, voice packets should not be dropped, should not be delayed, or suffer varying delay, which demand by user to receive an acceptable level of voice quality. QoS ensures that VoIP voice receive the preferential service, which they require. In general, QoS provides better service by providing features like Supporting dedicated bandwidth, QoS queuning Mechanisms, Resource Reservation, Shaping traffic network, Maintaining sources of delay, VOIP QoS using Point-to-Point Protocol, Compressed RTP, Differentiated Services for VoIP, Avoiding and managing network congestion, Shaping network traffic and setting traffic priorities across network.

Background

What is QoS?

QoS, or Quality of Service, is techniques to control network resources to guarantee that delay-receptive information pass through the network in a suitable manner. For VoIP, this implies prioritizing voice packets over the network to evade delay and packets loss. It is achievable to arrange the hardware and software on your network to achieve more predictable service by: [1][3]

Supporting dedicated network bandwidth

Avoiding and managing network congestion

Shaping network traffic

Setting traffic priorities across the network

Why Do I need QoS?

A modern study by Network computing showed that network latency and QoS are the biggest problems to implementing an IP telephony solution. Most existing data network were designed for “bursty” applications that are not delay-sensitive. However, Voice is a “real-time” application. For example, if a voice packet comes more than approximately 200 milliseconds (ms) after it is transmitted, it will appear too late! Receivers will criticize about echo, delay and calls that sound distorted.

Also, a conventional circuit switched network has its personal offered bandwidth all over the network and that’s why delay is rarely a problem and they achieve 99.999% reliability. This is the usual that VoIP calls are judged against. But by using QoS tools, it is achievable to get good enough quality using shared and packet switched data/voice networks. [3]

Getting QoS is a bigger challenge when packets of video and voice are travelling through a Wide Area Network. Nearly all Local Area Network (LAN) run at 10 Mbps to 100 Mbps. Because of bandwidth using the Wide Area Network is much more expensive, so many WANS works at T1 speeds (1.45 Mbps) or at slower then that speed, can create a jam at the LAN/WAN interface. In such a case, Quality of Service techniques are serious to providing expected reliability and voice quality. [6]

State of the Art for this technology

Quality Measures: [12]

In designing a Voice over IP network, it is important to think about all the factors, which can affect quality of voice. A quick summary of those major factors are given below.

Bandwidth [3]

Bandwidth is the entire capacity of a transmission medium, which allows transfer of data. Bandwidth can make an effect on voice quality of service using the parameter like speed and capacity of the network. More bandwidth gives more ability to the network to transmit data. Capability of a network to convey information reduces if the network is full of users.

Latency (End-To-End Delay) [1][3]

It is a time necessary to send data from one end (source) to the other end (destination). It is a delay that happens in order to exchange data or information between source and destination nodes. Purely, it can be defined as the network’s speed, which can make an effect on quality of the voice service. One solution of delays in data packets can be reduced if the overall packet size can be reduced. Mainly there are couple different types of delay in a network, which are called variable and fixed. Fixed delay arises from components on the link, which can be added directly to the overall delay. Variable delays comes from queuing delays, uses buffers on the connected interface to the Wide Area Network. Those buffers generate variable delays across the network named jitter delay. Such Variable delays can be handled by using de-jitter buffer at the other end i.e. the receiving router/gateway end. There are some delays, which have an affect the QoS are mention below.

Coder (Processing) Delay [12]

PCM signal is compressed by the processor and time require for that is known as codec delay. It is varies over the codec and processor’s speed. For example, ADPCM coded using 32kbps speed generate variable delay depending processor. While decompression delay is approximately one tenth of the compression delay and proportional to the no. of sample per frame. Decompression delay occurs at the end router or gateway.

Algorithmic delay [12]

Compression algorithm must have knowledge about the samples of N block. Because during the compression of signal, it is necessary that voice signal passes correctly signal blocks whether it’s an N or N+1, it should be processed in correct manner. So it takes time for running such algorithm, which is called algorithmic delay. G729 coder generates 5 ms algorithmic delay while G726 generates 0 ms algorithmic delay. So it’s all depending on, which coders are going to be used. Lumped coder delay parameter is a function (addition) of coder delay, decompression delay and algorithmic delay.

Packetization delay [1][12]

This delay is a time requires producing entire packet with encoded signal. Packetization delay is also called Accumulation delay. Because of voice samples are accumulate in a buffer before they released to a network, it sometimes called as accumulating delay. For example, coder PCM G.711 having a rate of 64 kbps and having a payload size of 160 bytes generates Packetization delay of 20 ms. don’t forget about CPU load, because to get the Packetization delay minimum, frame rate should be maximum, which enhance to increase the CPU load.

Serialization Delay [12]

It is a fixed delay required to put the voice frame or data frame on a network interface. It is depend on clock rate of trunk. For example, a frame size of 64 bytes having line speed of 64 kbps takes time about 8 ms while having a line speed of 2.048 Mbps require time of 0.25 ms.

Queuing/Buffering Delay [3]

Once the frame is ready after adding a header of compressed voice payload, it is queued for transmission on network connection, where frames have given priority in the router or gateway. It is depend on state of queue and speed of trunk. Voice frame has to wait for another voice frame or data frame ahead of it. In other words, frame has to wait for serialization delay of previous frame in the buffer. As the number of user increases, the probability of packet waiting in a queue will increase. However voice frame has a higher priority than data frame, voice frame doesn’t need to wait in a queue behind data frame.

Network Switching Delay

ATM or the public frame relay network, which interconnects the source and destination endpoints locations, is the main source of the biggest delays for audio means voice connections. Network Switching Delays (ωn) are difficult to calculate. It is nearly possible to recognize the individual delay of components. In common, the fixed components are from propagation delays on the trunks within the network, and variable delays are from queuing delays regulating frames into and out of such network intermediate switches. To orderly calculate transmission delay, a popular known approximation of (G.114) 10 microseconds (ms)/mile or 6 microseconds (ms)/km, is used widely. But, intermediate multiplexing equipment, microwave links, backhauling, and other factors found in carrier networks are generating many problems.

De-Jitter Delay [12]

De-jitter is a fixed delay, transform by the router using variable delay. It is necessary for a network that jitter need to be suppressed because of the constant bit rate of voice. Using Cisco router and gateway at the end point, this is accomplished. It works on initial play out. When a sample received by a router or gateway, the sample has to wait for the time, which is defined as fixed play out time. The required initial play out time for a de-jitter buffer is the same as the all variable delay during connection. The de-jitter buffer should be handled properly because if samples are hold for small period of time then variation comes in delay and produces break in speech signal while samples are coming to fast then it is require, buffer has to over-run and it drops packet from queue, which also makes gaps in output signal. Normally maximum size of buffer before it overflows is normally kept as 1.5 of initial time out delay.

Figure 1. Shows all types of delay during end-to-end transmission of packet [12]

Jitter [4][3]

Information, which is transmitted or transferred from one ends (sources) to another ends (destinations) via little or small messages called packets. These packets are experiencing different kinds of delays to arrive at their final destinations. Such a variation in those delays is recognized as jitter, which negatively affect voice quality of the offered service. It makes break in sound or create certain sounds due to loss of packets but can be controlled via jitter buffers.

Packet Loss [1][2]

Data or voice packets are dropped because of congestion happens in the network or not having required buffer size at receiver side on router or gateway. If these packets are going too lost, they can’t be able to obtain unless they are again retransmitted. In this way, it affect the speed, which enhance degrade the quality of the network. To minimize such data loss in network, QoS monitoring and guarantees congestion and queuing management using different tools like Custom Queuing (CQ), First In First Out (FIFO) Queuing, Priority Queuing (PQ) etc. Queueing management avoids queues by not filling and regulates space for packets, which have high priority.

Techniques can improve Quality of Service (QoS)

Sufficient Bandwidth [3]

Allocating sufficient bandwidth is look like a basic and costly move towards the Quality of Service because of not demanding major changes to system development, but it’s a pricey way. This problem should not to appear so that all resources of network could be accessible with traffic bursts. Unhappily, this overbuilds would many times stay idle for almost of all the time. This strategy would not be more an ineffective way of resolving the Quality of Service problem, but it provides some good merits and so it should not be send away absolutely. For instance, an 80 kbps G.711 VoIP call, which includes 64 kbps payload and 16-kbps header will be unfortunate over a 64-kbps link because at least 16 kbps of the packets (which is 20 percent) will be dropped. This paradigm also presumes that there is no additional traffic is travelling through the entire interface. After providing sufficient bandwidth to voice traffic, further steps can give a guarantee that voice or data packets have a certain percentage of the total bandwidth and get priority.

QoS Queueing Mechanisms [3]

For efficient flow of traffic through network, the crucial mechanism involve is Queuing Mechanism. According to QoS requirement, all traffic has been placed into QoS classes, which provide bandwidth guarantees and priority servicing through queuing mechanism. Queuing provide router and switches to handle burst of traffic, measure network congestion, prioritize traffic, and allocate bandwidth. Every queuing has its own configuration commands, which help to control network traffic on routers and switches. Cisco routers offer four basic output queuing schemes for the handling of network traffic. They are first-in, first-out (FIFO), priority queuing, custom queuing, and weighted fair queuing. [1]

Figure 2 shows queuing mechanism.

References

[1] Colin Perkins, Orion Hodson, and Vicky Hardman. A survey of packet loss recovery techniques for streaming audio. IEEE Network, 12(5):40-48, September 1998.

[2] Wenyu Jiang, Henning Schulzrinne, Modeling of packet loss and their effect on real-time multimedia service quality. NOSSDAV 2000, Chapel Hill, NC, June 2000.

[3] Weibin Zhao, David Olshefski and Henning Schulzrinne. Internet Quality of Service: an Overview. Columbia Technical Report, CUCS-003-00. Columbia University, Computer Science Department, Feb. 2000.

[4] R. Guerin and V. Peris. Quality-of-service in packet networks: Basic mechanisms and directions. Computer Networks, 31(3):169–189, February 1999.

[5] Peuhkuri M., IP Quality of Service, Helsinki University of Technology, Laboratory of Telecommunications Technology, 1999.

[6] Leonard Franken. Quality of Service Management: A Model-Based Approach. PhD thesis, Centre for Telematics and Information Technology, 1996.

[7] Oram, Andy (2002-06-11). “A Nice Way to Get Network Quality of Service?”. O’Reilly Net.com.

[8] Kevin Wallace, CCVP QOS Quick Reference Sheets, Cisco Press.

[9] Teck-Kuen Chua and David C. Pheanis, QoS-aware end-to-end adaptive congestion detection and control for VoIP, ACTA Press

[10] “Supporting Differentiated Service Classes: Queue Scheduling Disciplines”. Chuck Semeria. Juniper Networks. December 2001.

[11] “Quality of Service Solutions Configuration Guide”. Cisco Systems Inc.

[12] RFC 2475 – “An Architecture for Differentiated Services”. M. Carlson, W. Weiss, S. Blake, Z. Wang, D. Black, and E. Davies. December 1998.

[12] Mansour J. Karam and Fouad A. Tobagi, Analysis of delay and delay jitter of voice traffic in the Internet.

[13] Robert Caputo, Cisco Packetized Voice And Data Integration, McGraw Hill 2000.