Telecoms glossary • I
IP (Internet Protocol)
The addressing scheme everything else runs on. The IP in VoIP, and the reason a call is treated like any other traffic unless you tell the network otherwise.
What is IP (Internet Protocol)?in networking and VoIP
1. IP, the Internet Protocol, is the addressing and delivery scheme every device on a network uses. Data is split into packets, each stamped with a source and a destination address, and routers pass them along hop by hop. The IP in VoIP is this protocol: voice travels as ordinary packets.
2. IP makes no promise that a packet will arrive, arrive in order or arrive on time. That indifference is the reason voice needs QoS, and the reason loss, delay and variation are measured on a call rather than assumed.
How a call becomes packets
A VoIP call is two streams over IP. SIP carries the signalling that sets the call up, and RTP carries the audio, fifty packets a second holding 20 ms of speech each. Every one of those packets carries addressing before it carries any sound.
What is inside one packet of a G.711 call
That overhead is why a 64 kbit/s codec costs 87.2 kbps on the wire, and why compressing the audio saves less than the codec names suggest. The headers stay the same size whatever is inside them.
Public, private and why a phone stops registering
Your handsets do not hold public addresses. They sit on one of the private ranges reserved by RFC 1918, and the router translates between those and the single public address your ISP gives you.
- The private ranges
- 10.0.0.0/8, 172.16.0.0/12 and 192.168.0.0/16. Anyone may use them internally without asking a registry, which is why every office in the country has a 192.168 network and why they cannot reach each other directly.
- Translation
- NAT rewrites addresses on the way out and matches replies on the way back. It works invisibly for web traffic and causes most of the one-way audio faults in VoIP, because the addresses inside a SIP message are not the ones NAT rewrote.
- A changing public address
- Handsets register outward and the platform replies to whatever address it saw. When a dynamic public address changes, the platform is briefly answering an address that no longer reaches you, and calls fail until the phones register again.
- Carrier-grade NAT
- Fixed wireless and mobile services often place many customers behind one public address. Outbound registration still works; anything that needs a port opened inward does not.
IPv4, IPv6 and what it means for your phones
IPv4 addresses are 32 bits long, which is where the shortage came from and why almost everyone lives behind NAT. IPv6 addresses are 128 bits, published as an internet standard in RFC 8200, and remove the shortage rather than the complexity.
For a business buying a phone system this is a question to ask rather than a decision to make. If your ISP has moved you to a dual-stack or IPv6-first connection, check whether your voice platform accepts registrations over IPv6. Where it does not, a dual-stack router simply uses IPv4 for the handsets and nothing changes. What matters is that somebody has checked, rather than discovering it during a cutover.
When a static IP is worth paying for
You need one if you run an on-premise PBX that must accept calls inward, if you have a site-to-site link between branches, or if your provider restricts your account to a known address. In each case the cost of the static address is trivial against the alternative.
You do not need one for a hosted system where handsets register outward. Providers who insist on a static address for a standard cloud extension are describing their own platform’s limits, and it is fair to ask why.
About this entry
Definitions are written for South African business buyers and checked against primary sources – provider documentation, standards bodies and ICASA – not vendor marketing. Reviews are independent and sponsors are always disclosed. Read our editorial policy and scoring methodology.
Spotted something out of date, or think we have this wrong? Tell the editors and we will re-check it.