1. SIP vs. RTP: The Control and the Media

Every VoIP call consists of two different types of traffic:

• SIP (Session Initiation Protocol): The "Control Plane". Like a operator, it finds the other person, rings their phone, and asks them if they want to talk. Usually runs over TCP or UDP port 5060.
• RTP (Real-time Transport Protocol): The "Data Plane". Once the call is established, SIP "gets out of the way" and the two phones send raw audio data directly to each other using RTP over UDP (random port range 16384-32768).

2. Metrics of Quality

We measure VoIP quality using the MOS (Mean Opinion Score), a scale from 1 to 5.

1. Latency: Must be below 150ms (one-way). Above 200ms, people start talking over each other.
2. Jitter: The variation in packet arrival time. Must be below 30ms.
3. Packet Loss: Anything above 1% causes noticeable audible gaps.

Common Issues: NAT and One-Way Audio

The most common VoIP problem is "One-Way Audio" (User A can hear User B, but not vice versa). This is almost always caused by a NAT/Firewall issue. The firewall allows the outgoing RTP stream from User A but blocks the incoming RTP stream from User B because it doesn't recognize it as a related session.

Conclusion

VoIP engineering is the art of minimizing jitter and protecting real-time flows. By understanding the separation of SIP and RTP, engineers can accurately diagnose whether a call failure is a connection issue (SIP) or a quality issue (RTP/Network).