📡 What Is SIP Trunking? A Quick Refresher
Before diving into the architecture, let's ground ourselves in the basics. SIP stands for Session Initiation Protocol — a communications standard that manages the setup, maintenance, and termination of real-time voice, video, and messaging sessions over IP networks.
A SIP trunk is a virtual telephone line. Instead of running copper wires from your office to the public switched telephone network (PSTN), a SIP trunk carries your calls over your broadband internet connection. Think of it as a "virtual cable" connecting your business phone system to the global telephone network — without any physical infrastructure.
SIP trunking is not a single piece of hardware or software — it's a stack of technologies working in harmony. Understanding how each layer works helps businesses make smarter decisions about telecom infrastructure, security, and cost management.
🏗️ Core Components of SIP Trunking Architecture
SIP trunking architecture consists of several distinct layers and components that each play a critical role. Here's a breakdown of every key piece in the puzzle:
🔧 SIP Trunking Architecture — End-to-End Flow
IP-PBX (Private Branch Exchange)
The brain of your phone system. The IP-PBX manages internal extensions, call routing, voicemail, and connects your office phones to SIP trunks. Examples: Asterisk, FreePBX, 3CX.
Session Border Controller (SBC)
Acts as a security gatekeeper between your network and the SIP provider. The SBC handles NAT traversal, encryption, protocol normalization, and DDoS protection.
Internet Telephony Service Provider (ITSP)
Your SIP trunk provider — the company that connects your calls to the PSTN. NetViaVoice is an ITSP offering high-availability SIP trunks globally.
DID Numbers (Direct Inward Dial)
Virtual phone numbers assigned to your SIP trunks. DIDs let you have local numbers in multiple cities or countries without physical offices.
SIP Proxy Server
Routes SIP messages between user agents. Acts as an intermediary that forwards INVITE, BYE, and other SIP requests to the correct endpoints.
Firewall / NAT Router
Your network edge device. Proper SIP ALG (Application Layer Gateway) configuration ensures SIP packets traverse NAT correctly without media issues.
📜 The Protocols Behind SIP Trunking
SIP trunking doesn't run on a single protocol — it's a coordinated stack. Understanding each protocol's role helps diagnose issues and optimize performance.
SIP — Session Initiation Protocol
Handles signaling: setting up, modifying, and tearing down calls. Operates over UDP (port 5060) or TCP/TLS (port 5061). Text-based, similar to HTTP.
RTP — Real-Time Transport Protocol
Carries the actual voice media (audio packets) between endpoints. Uses even-numbered UDP ports, dynamically negotiated via SDP. Provides timestamps for synchronization.
SDP — Session Description Protocol
Embedded in SIP messages (INVITE, 200 OK). Describes the media session: codecs, IP addresses, ports, and bandwidth requirements. The "blueprint" for the call.
SRTP — Secure RTP
Encrypted version of RTP. Provides confidentiality and integrity for voice media. Used with TLS-secured SIP (SIPS) for end-to-end encryption.
| Protocol | Function | Transport | Default Port | Encrypted Version |
|---|---|---|---|---|
| SIP | Call signaling & control | UDP / TCP | 5060 | SIPS (TLS) — port 5061 |
| RTP | Voice media transport | UDP | 10000–65535 (dynamic) | SRTP |
| SDP | Session parameters negotiation | Embedded in SIP | N/A | Via SIPS |
| RTCP | RTP quality monitoring | UDP | RTP port + 1 | SRTCP |
| DNS SRV | SIP server discovery | UDP / TCP | 53 | DNSSEC |
🔄 Step-by-Step SIP Call Flow Explained
A SIP call might seem instantaneous to the caller, but behind the scenes it involves a precise sequence of messages exchanged in milliseconds. Here's exactly what happens when you dial a number over a SIP trunk:
Your IP phone or softphone sends a SIP INVITE message to the IP-PBX. This message contains the destination number, your IP address, and an SDP body describing your preferred codecs and media ports.
The IP-PBX routes the INVITE through the SBC to your SIP provider (ITSP). The SBC performs NAT translation, security checks, and may re-negotiate codec parameters.
The ITSP sends a "100 Trying" immediately, acknowledging receipt. Once the call reaches the destination carrier, a "180 Ringing" is sent back — this triggers the ringback tone you hear.
When the remote party picks up, a "200 OK" message is sent back through the chain. This message contains the remote party's SDP — their IP, port, and codec selection. Both sides now know where to send media.
Your PBX sends an ACK message to complete the SIP three-way handshake. This confirms the session parameters. At this point, signaling is complete.
Voice packets start flowing directly between the two endpoints via RTP — completely separate from the SIP signaling path. Audio is encoded using the negotiated codec (e.g., G.711, G.729).
Either party hangs up, sending a SIP BYE message. The receiving side responds with "200 OK." Both ends stop sending RTP packets. The session is cleanly terminated.
💡 Key Insight: Signaling vs. Media Are Separate
One of SIP's most important architectural features is the separation of signaling and media. SIP (signaling) just sets up the call and negotiates parameters. RTP (media) carries the actual voice. This separation enables advanced features like call transfer, conferencing, and media re-routing without interrupting the voice stream.
🎵 Codecs & Audio Quality
A codec (coder-decoder) compresses and decompresses voice data for transmission. The codec you use directly impacts call quality and bandwidth consumption. SDP negotiation during the INVITE/200 OK exchange determines which codec both sides will use.
| Codec | Bit Rate | Quality | Bandwidth/Call | Best For |
|---|---|---|---|---|
| G.711 (PCMU/PCMA) | 64 kbps | Excellent | ~87 kbps | LAN, high-quality calls |
| G.729 | 8 kbps | Good | ~31 kbps | Low bandwidth environments |
| Opus | 6–510 kbps | Excellent | Variable | WebRTC, HD voice, adaptive |
| G.722 | 64 kbps | HD Audio | ~87 kbps | Wideband / HD voice calls |
| iLBC | 13.3/15.2 kbps | Good | ~27 kbps | Lossy networks, packet loss tolerance |
| GSM | 13 kbps | Acceptable | ~26 kbps | Legacy compatibility |
Most modern SIP trunks support G.711 as the baseline and G.729 for bandwidth-constrained scenarios. At NetViaVoice, we support all major codecs and provide HD voice capability via G.722 and Opus for crystal-clear call quality.
🔒 Security Architecture in SIP Trunking
Security is a critical consideration in SIP trunking. VoIP systems are targets for toll fraud, eavesdropping, and DoS attacks. A properly architected SIP trunk deployment includes multiple security layers:
🛡️ SIP Security Layers — What Every Deployment Needs
- TLS (Transport Layer Security): Encrypts SIP signaling messages in transit. Uses port 5061. Prevents eavesdropping on call setup data.
- SRTP (Secure Real-Time Protocol): Encrypts the actual voice media stream. Prevents call interception even if packets are captured.
- Session Border Controller (SBC): Acts as a hardened edge device — hides your internal topology, enforces policies, and blocks malformed SIP messages.
- IP Whitelisting: Restricts SIP registrations and INVITE messages to known, trusted IP addresses from your ITSP.
- Fail2Ban / Rate Limiting: Automatically blocks IPs that send excessive authentication failures — stops brute-force SIP attacks.
- VLAN Segmentation: Isolates voice traffic on a dedicated VLAN, preventing voice and data networks from interfering with each other.
- Strong SIP Credentials: Complex usernames and passwords for SIP authentication. Weak credentials are the #1 cause of toll fraud.
⚖️ SIP Trunk Architecture vs Traditional PSTN
| Feature | SIP Trunking | Traditional PSTN (PRI/BRI) |
|---|---|---|
| Infrastructure | Virtual — IP-based, no physical lines | Physical copper/fiber circuits |
| Scalability | Add channels instantly, software-based | Fixed increments (PRI = 23 channels) |
| Cost per call | Very low — often flat-rate bundles | Higher per-minute rates |
| International calls | Low cost, global number support | Expensive international rates |
| Disaster recovery | Failover to alternate IP in seconds | Requires physical rerouting |
| Setup time | Minutes to hours (remote provisioning) | Days to weeks (physical install) |
| HD Voice support | ✅ Yes (G.722, Opus) | ❌ No (narrowband only) |
| Advanced features | UCaaS, video, presence, chat integration | Basic voice only |
| Redundancy | Built-in geographic failover | Requires duplicate circuits |
🚀 Key Business Benefits of SIP Architecture
- Unlimited Scalability: Add or remove SIP channels in real time — no hardware upgrades, no waiting for an engineer.
- Geographic Flexibility: Get local numbers in hundreds of cities worldwide. Appear local, call globally.
- Business Continuity: Automatic failover to backup SIP servers or mobile numbers if your primary connection fails.
- Cost Reduction: Eliminate expensive PRI circuits and reduce international calling costs by up to 70%.
- Unified Communications: SIP integrates with video conferencing, instant messaging, CRM systems, and contact centers.
- Easy Management: Cloud-based portals let you manage DIDs, call routing, and billing from any browser.
- HD Voice Quality: G.722 and Opus codecs deliver wideband audio — callers sound like they're in the room with you.
- Compliance & Recording: Built-in call recording, encryption, and audit trails for regulated industries.