In this glossary deep dive, we’ll explain what SIP is (and what it isn’t), how SIP call flows work in plain English, why SIP matters for reliability and scalability, and how related concepts like SIP trunking and SIP channels.
What Is SIP?
SIP (Session Initiation Protocol) is a signaling protocol used to start, manage, and end real-time communication sessions over IP networks, most commonly VoIP calls, but also video and conferencing.
SIP controls the “call setup” and “call control” parts of communication: finding the other party, ringing, answering, putting a call on hold, transferring, and hanging up. SIP does not carry the actual voice or video media itself—media typically flows over RTP (or SRTP when encrypted), based on the media parameters negotiated during setup.
What is SIP used for?
ASIP is used anywhere you need devices and platforms to establish and control live sessions across networks, carriers, or cloud services: phones, softphones, PBXs, contact centers, and SIP trunks.
SIP vs RTP vs SDP
In VoIP, these three work together but do different jobs. SIP is the “call control” protocol: it sets up the call, negotiates how it should work, and tears it down. SDP is not a transport protocol—it’s a description that rides inside SIP messages to agree on media details (codecs, IPs/ports, encryption, etc.). RTP is what actually carries the audio/video packets once the call is established (and SRTP is the encrypted version of RTP).
| Protocol | What it is | Primary job | Used when | What it carries | Common issues |
|---|---|---|---|---|---|
| SIP (Session Initiation Protocol) | Signaling protocol | Sets up, modifies, and ends sessions (ringing, answer, hold, transfer, hang up) | Before and during a call (signaling phase) | SIP messages (INVITE, REGISTER, BYE, etc.) often containing SDP | Registration failures, wrong routing/DIDs, auth issues (401/407), call drops due to timers |
| SDP (Session Description Protocol) | Session/media “offer/answer” description format | Negotiates media parameters (codecs, IP/ports, RTP direction, SRTP settings) | Inside SIP messages during call setup/changes | Codec list, media IP/ports, SRTP crypto, attributes like sendrecv/recvonly | One-way/no audio due to wrong IP/ports (NAT), codec mismatch, SRTP mismatch |
| RTP (Real-time Transport Protocol) | Media transport protocol | Carries the actual voice/video streams | After call is answered (media phase) | Audio/video packets (with timing/sequence) | One-way audio, choppy audio (jitter/loss), firewall/NAT blocking RTP ports |
How SIP Works
Your phone or softphone (the SIP endpoint) uses SIP messages to find the other party, ring them, authenticate if needed, and agree on media details (via SDP). Once the call is answered, the audio/video usually flows over RTP/SRTP, while SIP stays in the background to handle changes like hold, transfer, or hang-up.
SIP endpoints and core roles
-
User Agent (UA)
The device or app that speaks SIP (desk phone, softphone, PBX). A UA can act as:-
UAC (User Agent Client): sends requests (e.g., INVITE to start a call)
-
UAS (User Agent Server): responds to requests (e.g., 180 Ringing, 200 OK)
-
-
SIP servers (what they do)
-
Register: accepts REGISTER requests and stores where a user can be reached (their current IP/Contact).
-
Proxy server: routes SIP requests between endpoints (like a traffic director).
-
Redirect server: tells the caller where to try next (gives a new address, doesn’t forward the call itself).
-
Location service: the database that maps a SIP address to a current Contact location (often used by the registrar/proxy).
-
What is a SIP trunk?
A SIP trunk is a virtual phone line that connects your PBX/UC system (on-prem or cloud) to a telecom provider (ITSP) over the internet. Instead of physical circuits, your calls are carried over IP: SIP handles the call setup/control, and the audio flows over RTP/SRTP once the call is connected. In simple terms: it’s how most modern business phone systems reach the PSTN without traditional phone lines.
SIP trunking vs traditional PSTN lines (PRI/BRI/analog)
Traditional PSTN access uses dedicated physical services:
-
Analog lines (single call per line)
-
BRI/PRI (ISDN) circuits (fixed call capacity per circuit)
SIP trunking replaces those circuits with an IP connection and a capacity model defined by your provider (concurrent calls, usage, or a hybrid), typically with more flexibility and easier scaling.
What is a SIP Channel?
A SIP channel is the unit of capacity that determines how many calls you can have happening at the same time on a SIP trunk. In most business telephony contexts, 1 SIP channel = 1 concurrent call (whether inbound or outbound). If you have 10 SIP channels, you can typically support up to 10 simultaneous calls at peak—regardless of how many total users/extensions you have.
SIP channels vs users
-
Users/extensions = how many people/devices can place or receive calls
-
SIP channels = how many calls can be active at once (concurrency)
So 50 agents might be fine on 15 channels if only ~15 are on calls at the busiest moment—but a 20-agent sales floor could need 20+ channels if everyone dials at the same time.
FAQs
What is SIP in VoIP?
SIP (Session Initiation Protocol) is the signaling protocol used in many VoIP systems to set up, control, and end calls. It handles things like dialing, ringing, answering, hold/transfer, and hang-up—while the actual audio usually travels separately over RTP (or SRTP).
Is SIP the same as VoIP?
No. VoIP is the broader concept of making calls over IP networks. SIP is one specific protocol commonly used to manage (signal) those VoIP calls. You can think of SIP as a “call setup/control” standard inside many VoIP solutions.
What’s the difference between SIP and RTP?
SIP controls the session (call setup and call control). RTP carries the media (the actual voice/video packets) once the call is connected. SIP = signaling; RTP = audio/video transport.
What is a SIP trunk?
A SIP trunk is a virtual connection between your PBX/UC system and a telecom provider over the internet. It replaces traditional PSTN lines (PRI/BRI/analog) and enables inbound/outbound calling through an IP-based service.
How many SIP channels do I need?
In most cases, 1 SIP channel equals 1 concurrent call. Estimate your busiest-hour peak simultaneous calls (inbound + outbound), then add headroom (often 20–30%). If you run bursts (campaigns/dialers), also confirm provider limits like calls-per-second (CPS).
What ports does SIP use?
Most commonly, SIP uses 5060 for UDP/TCP and 5061 for SIP over TLS. Media (RTP/SRTP) uses a separate range of UDP ports that varies by platform/provider, so you typically need to allow an RTP port range in your firewall for audio to work reliably.
