A queueing model (commonly referred to as Erlang C) for systems with c identical servers, Poisson arrivals and exponential service times (M/M/c, infinite buffer, first‑come first‑served) that provides the steady‑state probability an arriving job must wait and related waiting‑time metrics used to size server pools to meet delay targets.

Principle

The probability of delay and expected waiting time are determined by the traffic intensity ρ = λ/(cμ); stability and finite steady‑state moments require ρ < 1, and as ρ approaches 1 both delay probability and mean waiting time grow rapidly.

Demonstration

Illustrative scenario: a contact centre estimates offered arrival rate and average handling time, computes the Erlang‑C delay probability and average wait, and increases staffing (c) until the predicted average wait meets the service‑level objective.

Misapplication

Applying Erlang‑C when callers abandon, balk, or when service times and arrivals deviate markedly from exponential/Poisson assumptions (or when priorities and skill‑based routing are present) misstates expected delays; ignoring abandonment typically overestimates actual queue lengths.

Consequence

Used appropriately, the model yields staffing rules and expected delay figures for capacity planning; misuse leads to staff under/over‑provisioning and mismatches between predicted and realized service levels.

Reversal

If customer impatience (abandonments), finite population effects, non‑exponential service distributions, priority schemes, or time‑varying arrival rates are material, models that include patience (M/M/c+G), time‑varying queues, simulation or more detailed analytic approximations should replace Erlang‑C.

Boundary

Within: homogeneous servers, FCFS discipline, Poisson arrivals, exponential services, and no abandonment. Boundary case: modest abandonment rates or weak non‑exponential service where corrections/approximations may suffice. Outside: systems with significant impatience, skill‑based routing, priorities, or strongly non‑Poisson arrivals.

Semantic Tension

Analytical clarity and simple staffing formulas ↔ deviation of real operational behaviors (abandonment, skill mixes, non‑stationary arrival patterns); Erlang‑C is a baseline that must be adapted when operational complexities matter.

Synthesis

Erlang‑C provides a tractable baseline for delay and staffing estimation in many service systems, but operational features like abandonment, priorities, and non‑exponential dynamics commonly require model extensions or simulation to obtain reliable predictions.