VoIP: Putting your voice across

In this era of convergence, sending voice over networks–like IP, ATM and

frame relay–has become the hottest technology facilitating real time

communication. The basic element in the technology of Voice over Internet

Protocol (VoIP) is the packet. It uses the packet switching network to transfer

voice across the network. But the challenge that lies before VoIP technology is

that it has to route its service through the traditional circuit-switching

network. To address this issue, the International Telecommunication Union (ITU)

has defined a standard–the H.323 for packet based multimedia networks.

A circuit switched network or the public switched telephone network creates a

dedicated path between the original and the destination call. Although the same

two-end points of the call can follow different paths, the circuit is fully

available for the duration of the call, making it unavailable to other network

users. But, in the packet switched networks there is no dedicated path between

the users. Here the content is broken up into packets and sent across the

network. Each packet carries a header, which determines the destination of the

packet. Here too the packets take different paths towards the destination but

they share their paths with other packets.

Now let’s see how H.323 works in this process. They are connected to the

ISDN, PSTN and your local landline or wireless devices. The components of H.323

are:

• H.323 terminals which are end points to the LAN
• Gateways
• Gatekeepers
• Multipoint control units.

The H.323 are the LAN based end point terminals. These terminals include the

Codecs and accomplish the voice transmission functions. Codecs are

compression/decompression devices that encode or decode a signal. When the

signal is determined to be voice, it is usually compressed by a DSP from 64K PCM

to one of the compression/decompression (Codec) format. The voice packet is

constructed as an IP packet, to avoid TCP/IP to attempt to correct a corrupted

packet by re-transmitting the packet. Any attempt to retransmit the packet would

result in unnecessary delays. They are hardware or software components providing

digital encoding and decoding of analog signals. They send and receive

packetized voice. These terminals also need to support signaling functions that

are used for call setup. The terminals are connected to the gatekeepers through

their RAS (Registration, Admission, and Status) point.

Gateways are interfaces between the LAN and switched circuit network. The

gateway serves as the interface between the two different physical networks. It

forms a connection between the packet switched network and the traditional

circuit switched network. It performs the task of translating the signaling

messages between the two sides and also compresses and decompresses the voice.

The originating gateway establishes connection to the destination gateway,

exchanges call set up and compatibility information, and performs any

negotiation and security handshake. At the same time, the gateway is also

receiving packets from other IP gateways and decompressing voice information

back to the original format, to be connected to the appropriate telephony

interface. When the IP gateway is required to place a call, it receives a called

party phone number from the calling phone and converts it to the IP address of

the far-end or the called party’s gateway (destination gateway).

Gatekeepers are usually present to perform a set of functions like provide

address translation (routing) for devices in their zone. The translation can be

between external and internal numbering systems. They also specify what devises

can call what numbers thus providing admission control. A gatekeeper can take

part in a variety of signaling models as commanded by the gatekeeper. The

signaling models determine what signaling messages pass through the gatekeeper,

and what can be passed directly between entities such as the terminal and the

gateway. There are two such kinds of signaling models. A direct signaling model

does not involve a gatekeeper for exchange of signaling messages, while in a

gatekeeper routed call signaling model, all signaling passes through the

gatekeeper.

The Multipoint Control Units can be standalone devices, or be integrated into

a terminal, a gateway or a gatekeeper. They have two parts, each performing

certain functions. The Multipoint controller handles the signaling and control

messages essential to set up and manage conferences. The Multipoint processor

accepts messages from endpoints, replicates them and forwards them to the

correct participating points.

But the most important aspect of the technology is the voice quality. This

can be affected primarily due to the type of Codec used and also because of

latency, jitter and packet loss.

But the technology enables the transfer of your voice in the form of packets

through a traditional circuit-switching network to its right destination.