The Effect of Packetization on VoIP Performance
When speech is transported over the Internet with VoIP (Voice over Internet Protocol) the digital signal is broken up into packets with a header added. Different transport protocols generate different packet sizes.
Other People Are Reading
-
Transport Packetization
-
Voice data is chopped into segments--called packets--by a codec program, then passed to a Transport program, which in VoIP is usually based on the Real-time Transport Protocol (RTP), the User Datagram Protocol (UDP), and the Internet Protocol (IP). Each of these adds a header. Finally, the Link Layer program adds another header.
Overhead
-
The payload size is a multiple of the codec's sample size. With a sample size of 10 bytes, for example, the data packet body can contain 10, 20, 30 bytes, etc. If the packet body is 10 bytes, the IP/UDP/RTP headers will add 40 bytes. An Ethernet Link Layer header adds another 18 bytes. A transfer 10 bytes of data, therefore, needs a total packet size of 68 bytes.
-
VoIP Performance
-
A larger data payload is more efficient. Also, the protocol used to transport the data makes a big difference. For example, using Multilink Point-to-Point Protocol instead of Ethernet adds six bytes rather than 18. If Compressed RTP is used, the 40-byte IP/UDP/RTP header is reduced to two or four bytes. An 8-byte header on a 40-byte body results in a lower overhead and produces faster voice data transfer.
-
References
- Photo Credit cigarette packet image by Warren Millar from Fotolia.com
