What Is Telecommunication Traffic?

Telecommunication traffic refers to the aggregate of messages, calls, and data flows carried by a communication network over a given time period. It describes both the volume and the temporal pattern of demand placed on a network, and its analysis is the basis for network dimensioning, capacity planning, and quality of service management. Understanding how traffic behaves under normal conditions and under load is essential to designing networks that serve users reliably without either under-provisioning resources or wasting capacity.

The field draws on queuing theory, probability theory, and operations research. Agner Krarup Erlang established the mathematical foundations of telecommunication traffic analysis in 1909 when he derived the probability distributions governing call blocking in telephone exchanges. The CCITT named the standard unit of telephone traffic intensity the erlang (E) in his honor: one erlang represents a continuous occupancy of one circuit for one hour.

Traffic Characterization and Measurement

Telecommunication traffic is characterized by its intensity, its statistical distribution, and the burstiness of its arrival pattern. Traffic intensity, expressed in erlangs for circuit-switched systems and in bits per second for packet-switched networks, describes the average load on a resource. Arrival processes are typically modeled using Poisson distributions for telephone traffic and more bursty self-similar or heavy-tailed distributions for internet traffic, which tends to exhibit long-range dependence due to the aggregation of many simultaneous flows with variable file sizes. Measuring traffic in operational networks involves collecting call detail records, interface counters, and flow records (using protocols such as NetFlow or IPFIX) and analyzing them to identify usage patterns, peak demand periods, and anomalous conditions. The ITU-T Teletraffic Engineering Handbook provides a systematic treatment of these measurement and modeling methods.

Erlang Theory and Capacity Planning

The Erlang-B formula computes the probability that an arriving call will be blocked when all circuits in a group are occupied, assuming no queuing: blocked calls are cleared and lost. The Erlang-C formula models systems where blocked calls wait in a queue rather than being lost, which is relevant for call centers and packet queues. These formulas allow engineers to determine how many circuits, trunk lines, or server instances are required to achieve a target grade of service, defined as the acceptable probability of blocking or delay. The erlang unit and its role in teletraffic engineering are foundational concepts that remain in active use for cellular network dimensioning and call center staffing. Capacity planning applies these results iteratively: traffic forecasts inform dimensioning decisions, and actual measurements are fed back to validate and update the models. For packet networks, the equivalent planning exercise uses bandwidth allocation algorithms and quality of service scheduling disciplines to allocate capacity among competing traffic classes.

Traffic Management and Congestion Control

When demand exceeds available capacity, networks must apply traffic management mechanisms to maintain acceptable service. In circuit-switched networks, call admission control prevents new calls from entering an overloaded exchange, accepting a small number of blocked calls rather than degrading all active calls. In packet-switched networks, active queue management algorithms such as Random Early Detection (RED) signal congestion to TCP flows before buffers overflow, allowing the TCP congestion control mechanism to reduce sending rates. Differentiated services architectures assign traffic to priority classes, ensuring that latency-sensitive flows such as real-time voice receive preferential treatment over bulk data transfers. Research published through IEEE Xplore on teletraffic modeling and congestion spans both classical queuing-theoretic approaches and more recent machine-learning-based traffic prediction methods. Communication system traffic analysis, in the broader sense of studying how information flows across any type of communication infrastructure, extends these methods to satellite links, industrial control networks, and sensor networks.

Applications

Telecommunication traffic has applications in a wide range of disciplines, including:

• Telephone exchange and cellular network capacity dimensioning
• Internet service provider backbone bandwidth planning
• Call center staffing and service level optimization
• Data center network traffic engineering
• Quality of service design for multimedia streaming and real-time communication