VoIP voice over IP - that is, voice delivered using the Internet Protocol is a term used in IP telephony for a set of facilities for supervision the delivery of voice information using the Internet Protocol. A major benefit of VoIP and Internet telephony is that it avoids the tolls charged by ordinary telephone service.
"VoIP is consequently telephony using a packet based network instead of the PSTN (circuit switched). The first VoIP application was introduced in 1995 - an "Internet Phone". The application was designed to run on a basic PC. The idea was to compress the voice signal and interpret it into IP packets for broadcast over the Internet. This "1st generation" VoIP appliance suffered from delay (due to congestion), disconnection, low quality (both due to lost and out of order packets) and incongruity."
"Voice over IP is the technology of digitizing sound, compressing it, breaking it up into data packets, and sending it over an IP (internet protocol) network where it is reassembled, decompressed, and converted back into an analog wave form.."
" Voice over IP (VoIP) is a blanket description for any service that delivers standard voice telephone services over Internet Protocol (IP)." Computers to transfer data and files between computers normally use Internet protocol.
The transmission of sound over a packet switched network in this manner is an order of magnitude more efficient than the transmission of sound over a circuit switched network.
As mentioned before, VoIP saves bandwidth also by sending only the conversation data and not sending the silence periods. This is a considerable saving because generally only one person talks at a time while the other is listening. By removing the VoIP packets containing silence from the overall VoIP traffic we can reach up to 50% saving. In a circuit switched network, one call consumes the entire circuit. That circuit can only carry one call at a time.
Requirements of a VoIP
The requirements for implementing an IP Telephony solution to support Voice Over IP varies from organization to organization, and depends on the vendor and product chosen. The following section aims to identify the fundamental requirements in the general case and is split into 3 sections:
Software Requirements
Hardware Requirements
Protocol Requirements
Software Requirements
The software package chosen will reflect the organizational needs, but should contain the following modules as defined in the Technology Guide Series - Voice Over IP Publication, and other sources.
Voice Processing Module. This aspect of the software is required to prepare voice samples for transmission. The functionality provided by the voice processing module should support:
A PCM Interface is required to receive samples from the telephony interface (e.g. a voice card) and forward them to the Voice Over IP software for further processing.
Echo Cancellation is required to reduce or eliminate the echo introduced as a result of the round trip exceeding 50 milliseconds.
Idle Noise Detection is required to suppress packet transmission on the network when there are no voice signals to be sent. This helps to reduce network traffic as up to 60% of voice calls are silence and there is no point in sending silence.
A Tone Detector is required to discriminate between voice and fax signals by detecting DTMF (Dial Tone Multi frequency) signals.
The Packet Voice Protocol is required to encapsulate compressed voice and fax data for transmission over the network.
A Voice Playback Module is required at the destination to buffer the incoming packets before they are sent to the Codec for decompression.
Call Signaling Module. This is required to serve as a signaling gateway which allows calls to be established over a packet switched network as opposed to a circuit switched network (PSTN for example).
Packet Processing Module. This module is required to process the voice and signaling packets ready for transmission on the IP based network.
Network Management Protocol. Allows for fault, accounting and configuration management to be performed.
Hardware Requirements
The exact hardware, which would be required, again, depends on organizational needs and budget. The list below highlights the most general hardware required.
The most obvious requirement is the existence (or installation) of an IP based network within the branch office gateway is required to bridge the differences between the protocols used on an IP based network and the protocols used on the PSTN.
The gateway takes a standard telephone signal and digitizes it before compressing it using a Codec. The compressed data is put into IP packets and these packets are routed over the network to the intended destination.
The PC's attached to the IP based network require the voice/fax software outlined above. They also require Full Duplex Voice Cards which allow both communicating parties to speak at the same time - as often happens in reality.
As an alternative to installing Voice Cards, IP Telephones can be attached to the network to facilitate Voice Over IP. A secondary gateway should be considered as a backup in the event of the failure of the primary gateway.
Protocol Requirements
There are many protocols in existence but the main ones are considered to be the following:
H.323 is an ITU (International Telecommunications Union) approved standard which defines how audio /visual conferencing data is transmitted across a network. H.323 relies on the RTP (Real-Time Transport Protocol) and RTCP (Real Time Control Protocol) on top of UDP (User Datagram Protocol) to deliver audio streams across packet based networks.
G.723.1 defines how an audio signal with a bandwidth of 3.4KHz should be encoded for transmission at data rates of 5.3Kbps and 6.4Kbps. G.723.1 requires a very low transmission rate and delivers near carrier class quality. The VoIP Forum as the baseline Codec for low bit rate IP Telephony has chosen this encoding technique.
G.711. The ITU standardised PCM (Pulse Code Modulation) as G.711. This allows carrier class quality audio signals to be encoded for transmission at data rates of 56Kbps or 64Kbps. G.711 uses A-Law or Mu-Law for amplitude compression and is the baseline requirement for most ITU multimedia communications standards. This is the basic protocol that is being used now in this simulations.
Real-Time Transport Protocol (RTP) is the standard protocol for streaming applications developed within the IETF (Internet Engineering Task Force).
Resource Reservation Protocol (RSVP) is the protocol which supports the reservation of resources across an IP network. RSVP can be used to indicate the nature of the packet streams that a node is prepared to receive.
Main 'types' VoIP
VoIP has broadly three main branches, which can and do overlap.
VoIP over the Internet This is probably the best known and most publicized, talking PC to PC. Basically free telephone calls. The call is only free if both parties to the call have access to the public Internet at zero cost..
Advantage... free calls regardless of distance or length of call.
Disadvantage.... often the voice quality is bad due to the lack of bandwidth available for the call.
Other factors. Have to use a PC or other computer running VoIP software.
Office to Office A large multinational company will have offices across the whole country. This allows every computer access to every other computer in the company.
Advantages. Interoffice calls are free, since the company already has the bandwidth between offices. The technology is transparent to the user, and requires minimum training. The only new equipment required is a gateway at each office. Voice quality is good, because the company has control over the bandwidth.
Disadvantage. Extra bandwidth may be required between offices, which offset the savings.
Other factors... The carrier providing the interoffice bandwidth will almost certainly offer an alternative solution including management of the internal telephone traffic.
DESCRIPTION
Here the simulations are run for a single scenario having two routers, high-priority source, low-priority source, high-priority destination, low-priority destination and low priority traffic routes. It can be shown as follows:-
Task 1:-Effect of voice packet size on link loads, voice packet dealy and voice packet delay variation
Task 2:-effect of speech activity detection on link loads, voice packet dealy and voice packet delay variation
The network and simulation details for task 1 are given in the tabular form below:-
Network, Traffic and Simulation run details for large and small frames per packet
Routers
2
High priority Source
1
High priority destination
1
Low priority source
1
Low priority destination
1
Number of rows
2
Number of frames per packet
1(small),10(large)
Frame size
4 m sec
Coding type
PCM G7.11
Speech activity detection
disabled
Simulation run time
900 simulation seconds
The network and simulation details for task 2 are given in the tabular form below:-
Network, Traffic and Simulation run Details for with and without speech activity detection
Routers
2
High priority source
1
High priority destination
1
Low priority source
1
Low priority source
1
Low priority destination
1
Number of rows
1
Number of frames per packet
1
Speech activity detection
Enabled, disabled
Simulation run time
900 simulation seconds
Results using graphs and tabular form
Task 1:-
Voice Packet End-to-End Delay Graph Comparison between large and small voice packets
Voice Packet End-to-End Delay for small voice packets with 1 frame per packet
Voice Packet End-to-End Delay for large voice packets with 10 frames per packet
This graph shows the comparision for voice packet end-to-end delay between small voice packets and large voice packets of 1 and 10 frames per packet respectively.From the graph we can see that the voice packet End-to-End delay is flat and very close to null without any variations for large number of packets and the average end-to-end delay is 0.138212154 which is almost close to 0.Where as for small number of frames per packet it has great raise initially and later at one point becomes stable and the average value of end-to-end delay is 3.21339727. From the above graph and average values we can conclude that the Voice Packet End-to-End delay trend is null for large voice packets ie;10 and is more for small frames per packet ie;1.
Voice Packet Delay Variation Graph Comparison between large and small voice packets
Voice Packet Delay Variation for small voice packets with 1 frame per packet
Voice Packet Delay Variation for large voice packets with 10 frames per packet
This graph shows the comparision for voice packet delay variation between small voice packets and large voice packets of 1 and 10 frames per packet respectively.From the graph we can see that the voice packet delay variation is flat and null without any variations for large number of packets and the average delay variation is 0.Where as for small number of frames per packet it has great raise initially and later at one point decrease at the same speed and the average value of delay variation is 0.741509354. From the above graph and average values we can conclude that the Voice Packet delay variation trend is null for large voice packets ie;10 and is more for small frames per packet ie;1.
Point-to-Point throughput comparision graph between large and small voice packets
Point-to-Point Throughput for small voice packets with 1 frame per packet
Point-to-Point Throughput for large voice packets with 10 frames per packet
This graph shows the comparision for point-to-point throughput between small voice packets and large voice packets of 1 and 10 frames per packet respectively.From the graph we can see that the Point-to-point throughput for large frames per packet is low compared to small frames per packet.The average point-to-point throughput for large frames per packet is 96639.81886 .Where as for small number of frames per packet the average value of Point-to-point Throughput is 220612.6379. From the above graph and average values we can conclude that the Point-to-point throughput trend is less for large voice packets ie;10 and is more for small frames per packet ie;1.
Point-to-Point utilization comparison graph between large and small voice packets
Point-to-point utilization for small voice packets with 1 frame per packet
Point-to-point utilization for large voice packets with 10 frames per packet
This graph shows the comparision for point-to-point utilisation between small voice packets and large voice packets of 1 and 10 frames per packet respectively.From the graph we can see that the Point-to-point utilisation for large frames per packet is low compared to small frames per packet.The average point-to-point throughput for large frames per packet is 49.79753005 .Where as for small number of frames per packet the average value of Point-to-point Throughput is 87.17681168. From the above graph and average values we can conclude that the Point-to-point throughput trend is less for large voice packets ie;10 and is more for small frames per packet ie;1.
Tabular form for displaying results of task 1:-
Results for large and small voice packet simulation run details for link loads,voice packet delay and delay variation.
Large voice packets
Small voice packets
Average value of voice packet end-to-end delay
0.138212154
3.21339727
Average value of voice packet delay variation
0
0.741509354
Average value of point-to-point throughput
96639.81886
220612.6379
Average value of point-to-point utilisation
49.79753005
87.17681168
Task2:-
Voice Packet End-to-End Delay Graph between enabled and disabled speech activity detection
Voice Packet End-to-End Delay with speech activity detection enabled for 1 frame per packet
Voice Packet End-to-End Delay with speech activity detection disabled for 1 frame per packet
Here is the comparison of Voice-packet End-to-End delay for 1 frame per packet between enabled speech activity detection and disabled speech activity detection.From the values we know that with speech activity detection enabled the voice packet End-to-End delay is very low and is just above null and the average voice packet End-to-End delay is 0.074346845.Whereas with speech activity detection disabled the voice packet delay variation is increasin gradually and remains same at one point of time and the average voice packet End-to-End delay is 3.21339727.This average value along with the graph shows that the voice packet delay variation is low for enabled speech activity detection simulation compared to disabled speech activity detection simulation.
Voice Packet Delay Variation graph between enabled and disabled speech activity detection
Voice-Packet Delay Variation with speech activity detection enabled and frames per packet is 1
Voice-Packet Delay Variation with speech activity detection disabled and frames per packet is 1
Here is the comparison of Voice-packet-delay variation for 1 frame per packet between enabled speech activity detection and disabled speech activity detection.From the values we know that with speech activity detection enabled the voice packet delay variation is very low and almost null and the average voice packet delay variation is 0.001989961.Whereas with speech activity detection disabled the voice packet delay variation is very high and keeps on changing the trends and the average voice packet delay variation is 0.741509354.This average value along with the graph shows that the voice packet delay variation is low for enabled speech activity detection simulation compared to disabled speech activity detection simulation.
Point-to-point throughput graph between enabled and disabled speech activity detection
Point-to-Point throughput with speech activity detection enabled and frames per packet is 1.
Point-to-point throughput with speech activity detection disabled and frames per packet is 1.
Here is the comparison of point-to-point throughput for 1 frame per packet between enabled speech activity detection and disabled speech activity detection.From the values we know that with speech activity detection enabled the point-to-point throughput is variable and is maximum 16000bits/sec and the average Point-to-Point throughput is 96639.81886.Whereas with speech activity detection disabled the voice packet delay variation is increasing steadily and follows stability at one point of time and then decreases and the average Point-to-Point throughput is 220612.6379. This average value along with the graph shows that the Point-to-Point throughput is low for enabled speech activity detection simulation compared to disabled speech activity detection simulation
Point-to-point utilisation graph between enabled and disabled speech activity detection
Point-to-Point utilisation with speech activity detection enabled and frames per packet is 1.
Point-to-point utilisation with speech activity detection disabled and frames per packet is 1.
Here is the comparison of point-to-point utilisation for 1 frame per packet between enabled speech activity detection and disabled speech activity detection.From the values we know that with speech activity detection enabled the point-to-point utilisation is variable and the average Point-to-Point utilisation is 37.74992924.Whereas with speech activity detection disabled the voice packet delay variation is increasing steadily and follows stability at one point of time and then decreases and the average Point-to-Point utilisation is 86.17681168. This average value along with the graph shows that the Point-to-Point utilisation is low for enabled speech activity detection simulation compared to disabled speech activity detection simulation
Tabular form for displaying results of task-2:-
Results for speech activity detection enabled and disabled for link loads,packet delay and delay variation
Enabled
Disabled
Average value of voice packet end-to-end delay
0.074346845
3.21339727
Average value of voice packet delay variation
0.001989961
0.741509354
Average value of point-to-point throughput
96639.81886
220612.6379
Average value of point-to-point utilisation
37.74992924
86.17681168
Verification/Validation of Results:-
Here ,
The voice coding rate is-64 kbps
Packet (voice frame) size is- 4 milli sec
Size of voice frame -0.004*64000 is 256 bits or 32 bytes
We used 1 frame per packet and 10 frames per packet
a.)For small voice packet the number of frames per packet is 1.
Inter arrival time = 4 ms
Packet size is 32+48 = 80 bytes
Packet overhead is 48/80 = 60 %
Now 640/0.004=16000bps/160kbps
Load along inter router link=2*160kbps=320kbps
But link speed is 256 kbps therefore over head
But the calculated average values are different. This is because that all simulations cannot be run exactly and there may be different scenarios .Thus we can justify that the expected results are not exact but are close to the results.
B.)For large voice packets the number of frames per packet is 10.
Inter arrival time=4 *10=40 ms
Packet size is 32*10=320+48=368bytes
Overhead is 48/368=13%
Now load along router link=2*8*368/0.04=147.2kbps
And utilisation =0.575or 57.5%
But the obtained average is 49.23 which is different from the expected result
This is because all simulations cannot be run precisely and exactly because different atmospheres have different overheads. Thus the obtained results are not exact but close to the expected results.
Such variations may occur in simulations .
Interpretation/conclusion
With the simulations run over and observing the results obtained we can conclude that the voice over IP has its own prominence in digital voice telephony. It is the next generation application which is going to pre occupy all the existing systems and solutions. It is no doubt because the on coming days are going to be ruled by VOIP.
