Technical data

Optimizing Voice Over IP
© 2012 Meru Networks, Inc. Configuring Quality of Service 241
Optimizing Voice Over IP
Transmitting voice over IP (VoIP) connections is, in most senses, like any other
network application. Packets are transmitted and received from one IP address to
another. The voice is encoded into binary data at one end and decoded at the other
end. In some sense, voice is just another form of data. However, there are a few
special problems.
The requirements for quality voice traffic are not exactly the same as the require-
ments for most data traffic:
 If a data packet arrives a second late, it is usually of no consequence. The data
can be buffered until the late packet is received. If a voice packet arrives a
second late, it is useless and might as well be thrown away.
 If a data packet takes a third of second to arrive at the destination, that is usually
fast enough. If voice packets routinely take a third of a second to arrive, the users
will begin to take long pauses between sentences to make sure that they don’t
interfere with the other person’s speech.
Quality VoIP calls need voice data to be delivered consistently and quickly. Meeting
the requirements of VoIP data requires either a connection with plenty of bandwidth
all along the data route or a means of ensuring a certain quality of service (QoS) for
the length of the call.
Even if the bandwidth is available, setting up the phone call can be a nontrivial task.
When a phone call is initiated, the destination of the call might be a standard tele-
phone on the public switched network, an IP-to-voice device at a particular IP
number, or one of several computers (for example, systems at home or the office and
a laptop used by an individual). If the destination device is a phone on the public
network, the initiation protocol must locate a gateway between the Internet and the
telephone network. If the destination is a person, the initiation protocol must deter-
mine which computer or device to call.
After the destination device has been found, the initiating and the destination
devices must negotiate the means of coding and decoding the voice data. This
process of finding a destination device and establishing the means of communication
is called session initiation.
The two main standards for initiating voice sessions:
 Session Initiation Protocol, or SIP, used for most VoIP telephone calls.
 H.323v1, used for multimedia communication, for example by Microsoft
NetMeeting.
In both cases, the initiating device queries a server, which then finds the destination
device and establishes the communications method.
After the two devices have been matched and the communication standards chosen,
the call proceeds. The server may remain in the communication loop (H.323v1) or it
may step out of the loop (SIP).