๐Ÿ—๏ธ

Advanced and Carrier-Level

Glossary / Advanced and Carrier-Level
๐Ÿ—๏ธ Advanced and Carrier-Level covers the infrastructure, software, and engineering concepts used by telecom carriers and large-scale VoIP providers. This section contains 21 terms. Most business users will never need to configure these directly, but understanding them helps when evaluating provider quality and reliability.
On this page: Class 4 Switch ยท Class 5 Switch ยท SIP Server Cluster ยท Load Balancer ยท Kamailio ยท OpenSIPS ยท Homer SIP Capture ยท SBC Transcoding ยท Multicast Paging ยท BroadSoft/Cisco BroadWorks ยท ASR ยท ACD (Average Call Duration) ยท PDD ยท Call Capacity Planning ยท Erlang ยท Erlang B Formula ยท Erlang C Formula ยท Geo-Redundancy ยท BGP ยท SIPREC ยท SRTP Key Management

Class 4 Switch
A high-capacity switch designed to route calls between different carriers or networks over long distances. Class 4 switches handle call routing decisions (which carrier to use, least-cost routing) but do not provide end-user features like voicemail or IVR. They are the backbone of wholesale voice traffic.
Related: Class 5 Switch ยท SIP Trunk
Class 5 Switch
A switch that serves end users directly, providing features such as voicemail, call waiting, caller ID, and call forwarding. Your Cloud PBX is essentially a modern, software-based Class 5 switch. While Class 4 switches move calls between networks, Class 5 switches deliver the features people use every day.
Related: Class 4 Switch ยท Cloud PBX
SIP Server Cluster
A group of SIP servers working together to handle calls as a single system. If one server in the cluster fails or becomes overloaded, the others take over its traffic. Clustering provides both redundancy (protection against failure) and scalability (ability to handle more calls by adding servers).
Related: Load Balancer ยท High Availability ยท Kamailio
Load Balancer
A device or software that distributes incoming SIP traffic evenly across multiple servers. Without a load balancer, one server could be overwhelmed while others sit idle. Load balancers monitor server health and route new calls only to servers that are functioning correctly. They are essential for large-scale VoIP deployments.
Related: SIP Server Cluster ยท Geo-Redundancy
Kamailio
An open-source SIP proxy server used by many telecom carriers and large VoIP providers. Kamailio can handle thousands of calls per second, making it suitable for carrier-grade deployments. It acts as a traffic director: receiving SIP messages, applying routing logic, and forwarding them to the correct destination server.
Related: OpenSIPS ยท SIP Server Cluster ยท SIP Proxy
OpenSIPS
Another open-source SIP proxy server, similar to Kamailio. Both projects originated from the same codebase (SIP Express Router) and share many features. OpenSIPS is known for its modular architecture and script-based routing logic. Providers choose between Kamailio and OpenSIPS based on community support, specific module availability, and team expertise.
Related: Kamailio ยท Load Balancer
Homer SIP Capture
An open-source tool that captures, stores, and visualises SIP messages flowing through a VoIP network. Homer lets engineers search through millions of SIP messages to diagnose call problems: failed connections, one-way audio, or authentication errors. It is the most widely used SIP troubleshooting tool in the VoIP industry.
Related: SIP Protocol ยท Kamailio
SBC Transcoding
The process where a Session Border Controller (SBC) converts audio from one codec format to another in real time. For example, if your PBX uses the G.711 codec but the carrier requires G.729, the SBC translates between the two formats on the fly. Transcoding uses significant processing power and can introduce slight audio delay.
Related: SBC ยท Codec ยท SRTP Key Management
Multicast Paging
A method of sending a live audio announcement to many phones simultaneously using IP multicast. Instead of the PBX sending a separate audio stream to each phone (which wastes bandwidth), multicast sends one stream that all phones on the network receive at the same time. It is commonly used for overhead paging in offices, warehouses, and schools.
Related: Paging ยท QoS
BroadSoft / Cisco BroadWorks
A commercial Cloud PBX platform used by many large telecom operators worldwide to deliver hosted phone services. BroadSoft (acquired by Cisco in 2018) provides the underlying software that powers the PBX features offered by carriers to their business customers. It is one of the most widely deployed hosted PBX platforms globally.
Related: Class 5 Switch ยท Cloud PBX
ASR (Answer-Seizure Ratio)
A quality metric that measures the percentage of call attempts that are successfully answered. An ASR of 50% means half of all attempted calls result in a connection. Low ASR can indicate network problems, incorrect routing, or issues with the destination carrier. Providers monitor ASR closely to maintain call quality.
Related: PDD ยท ACD (Average Call Duration)
ACD (Average Call Duration)
The average length of answered calls over a given period. ACD is used alongside ASR to assess traffic quality. Unusually short ACD (a few seconds) may indicate fraudulent traffic or routing errors, since legitimate calls typically last at least 30 seconds. Carriers use ACD trends to detect anomalies in their networks.
Related: ASR ยท Call Capacity Planning
PDD (Post-Dial Delay)
The time between when a caller finishes dialling and when they hear a ringing tone or busy signal. PDD is measured in seconds, and anything under 5 seconds is considered acceptable. Longer delays frustrate callers and may cause them to hang up. High PDD can be caused by complex routing, overloaded servers, or slow carrier interconnections.
Related: ASR ยท Load Balancer
Call Capacity Planning
The process of calculating how many simultaneous calls your phone system and network need to support. Capacity planning considers the number of employees, average call duration, peak calling hours, and expected growth. Under-provisioning leads to busy signals; over-provisioning wastes money. Erlang formulas are the standard tool for these calculations.
Related: Erlang ยท Erlang B Formula ยท Concurrent Call
Erlang
A unit of measurement for telephone traffic load. One Erlang equals one phone line being used continuously for one hour. If your office makes 30 calls per hour and each call lasts 4 minutes, the traffic load is 30 x 4/60 = 2 Erlangs. This unit is named after the Danish mathematician A.K. Erlang, who developed the formulas in the early 1900s.
Related: Erlang B Formula ยท Erlang C Formula ยท Call Capacity Planning
Erlang B Formula
A mathematical formula that calculates how many phone lines (trunks) you need based on your expected traffic and an acceptable probability of a caller getting a busy signal. For example, if you have 5 Erlangs of traffic and want less than 1% busy probability, Erlang B tells you that you need 11 trunks. It assumes blocked calls are lost (the caller hangs up).
Related: Erlang ยท Erlang C Formula ยท Call Capacity Planning
Erlang C Formula
A formula that calculates staffing requirements for call centres. Unlike Erlang B (which assumes callers hang up if lines are busy), Erlang C assumes callers wait in a queue. Given the traffic load, number of agents, and target answer time, the formula tells you the probability of a caller having to wait and the expected wait time.
Related: Erlang ยท Erlang B Formula ยท Call Queue
Geo-Redundancy
Running your PBX infrastructure in two or more geographically separated data centres. If one data centre goes offline due to a power failure, natural disaster, or network outage, the other data centre takes over automatically. Geo-redundancy provides the highest level of disaster protection for business communications.
Related: High Availability ยท Disaster Recovery ยท BGP
BGP (Border Gateway Protocol)
The routing protocol that directs traffic across the internet. In a geo-redundant PBX setup, BGP can automatically reroute calls to a backup data centre if the primary one becomes unreachable. BGP is managed at the network infrastructure level and is typically handled by the Cloud PBX provider, not the customer.
Related: Geo-Redundancy ยท Load Balancer
SIPREC (SIP Recording)
A standard protocol (RFC 7865/7866) for recording SIP-based calls. SIPREC creates a copy of the audio stream and sends it to a dedicated recording server without affecting the original call. It separates the recording function from the call-handling function, which improves reliability and makes it easier to comply with recording regulations.
Related: Call Recording ยท Call Recording Consent ยท SRTP Key Management
SRTP Key Management
The process of securely exchanging encryption keys between two endpoints so they can set up an encrypted voice call using SRTP (Secure Real-time Transport Protocol). Common methods include SDES (keys exchanged in the SIP signalling) and DTLS-SRTP (keys exchanged directly between endpoints). Proper key management is essential for truly secure voice communication.
Related: SRTP ยท TLS ยท SBC Transcoding

Related Sections

๐Ÿ”— SIP Protocol โ€” The signalling protocol behind all SIP-based communication
๐ŸŒŠ Networking for VoIP โ€” QoS, SBC, NAT, and network fundamentals
๐Ÿ”’ Security โ€” TLS, SRTP, encryption, and fraud prevention
โš™๏ธ Deployment and Administration โ€” Failover, high availability, and disaster recovery

๐Ÿ“… Ready to explore Cloud PBX for your business?
Start with the provider comparisons or feature guides. If you want expert help, book a short call with a consultant.