Issue 2.0
5-14 SV9100 IP K-CCIS
1.5.3 Bandwidth
The bandwidth required for VoIP calls depends on several factors.
Layer 2 media
CODEC
Packet Size
RTP Header Compression
Voice Activity Detection (VAD)
Number of simultaneous calls
Possibly add encryption after research.
Layer 2 media is concerned with moving data across the physical links in
the network. A few of the most common layer 2 media types are Ethernet,
PPP, and Frame Relay.
CODEC stands for Coder/Decoder and is the conversion of the TDM signal
into an IP signal and vice versa. A CODEC can also compress/decompress
the voice payload to save on bandwidth.
Packet Size is the amount of audio in each PDU (protocol data unit)
measured in milliseconds. The larger the packet the less bandwidth used.
This is because sending larger packets (more milliseconds of voice)
requires, overall, less packets be sent. The downside of this practice is if a
packet is dropped/lost a larger piece of voice is missing from the
conversation as the system waits the additional delay for the next packet
arrival.
RTP Header Compression compacts the RTP header from 40 bytes in size
to 2 ~ 4 Bytes in size. RTP header compression is used only on low speed
links. Regularly on every voice packet there is an IP/UDP/RTP header that
is 40 bytes in length. Compressing this header, down to 2 ~ 4 bytes, can
save a considerable amount of bandwidth. The following is an example of a
VoIP packet without RTP header compression and one of a packet with
RTP header compression.
Notice that the overall packet size, when using RTP header compression, is
considerably smaller.
VOIP packet without RTP header compression
IP Header
20-bytes
UDP Header
8-bytes
RTP-Header
12-bytes
Voice-Payload
To Next Page
Loading...
Need help?
General
