Discrete Event Simulation

The system state only updates when a scheduled event occurs, not continuously. An event represents an instantaneous occurrence that may change the system state, such as a packet arrival at a router queue or a timer expiry triggering a retransmission. Between events, the state is assumed to remain constant, allowing the simulation clock to jump directly to the next scheduled event time. This next-event time advance mechanism is the defining characteristic of DES, enabling it to skip over periods of inactivity and achieve high computational efficiency for network models.

The core engine of a DES is the future event list (FEL), a priority queue ordered by event timestamp. The simulation loop repeatedly:

• Removes the event with the smallest timestamp from the FEL.
• Advances the simulation clock to that event's time.
• Executes the event's associated logic, which may generate new future events. This event scheduling approach is used in simulators like ns-3, where a central scheduler manages millions of packet transmission and reception events efficiently.

An alternative to pure event scheduling is the process-interaction approach. Here, entities (like a TCP connection) are modeled as threads or processes that execute a sequence of actions, including waiting for a specified simulation time. When a process issues a wait, it is suspended, and control returns to the scheduler. This makes modeling complex protocol state machines more intuitive, as the code structure mirrors the protocol's sequential logic. The simulator kernel maps these process waits into underlying event scheduling.

DES relies on pseudo-random number generators (PRNGs) to model stochastic behavior. Key applications include:

• Inter-arrival times: Generating packet arrival intervals from an exponential distribution.
• Service times: Modeling processing delays with a normal or Pareto distribution.
• Error events: Triggering packet loss based on a uniform distribution and a target Block Error Rate (BLER). Reproducibility is critical; simulators use deterministic PRNGs with configurable seeds to enable exact reruns of experiments.

During a simulation run, the system must continuously collect performance metrics without interfering with the model logic. This is done by instrumenting the code to record observations, such as queue length over time or end-to-end packet delay. Because the initial state is often empty (a transient phase), data collection typically starts after a warm-up period to ensure steady-state analysis. Final results, like mean throughput or 99th percentile latency, are computed with confidence intervals to account for the stochastic nature of the inputs.

The fundamental mechanism of DES where the simulator clock jumps directly from one scheduled event to the next, ignoring idle periods. This contrasts with time-stepped simulation. An event represents a change in system state, such as a packet arrival, a timer expiry, or a user mobility update. The core loop involves an event list (or calendar) that is continuously processed: the earliest event is removed, the simulation clock advances to that event's timestamp, and the associated handler function is executed, potentially scheduling new future events.

A simulation methodology that models a multi-cell network with numerous User Equipment (UE) instances to evaluate overall network performance. It relies on a DES engine to manage the discrete interactions between the MAC scheduler, handover logic, and traffic generators across many cells. Key performance indicators (KPIs) like average cell throughput, user fairness, and call drop rates are measured. This abstracts the physical layer details to focus on resource management and mobility algorithms.

A simulation methodology that models a single communication link between a transmitter and receiver to evaluate physical layer performance. It uses DES to process transport blocks and Hybrid Automatic Repeat Request (HARQ) feedback. The primary metric is the Block Error Rate (BLER) versus Signal-to-Noise Ratio (SNR). Results from link-level simulations are often abstracted into lookup tables for use in system-level simulators to balance accuracy and computational load.

A testing method where recorded real-world network data is injected into a discrete event simulator to recreate a specific field event. This involves feeding RF measurements, call traces, and user mobility logs into the DES engine. The simulator replays the exact sequence of events, allowing engineers to debug algorithms against a known, reproducible ground truth. It is a critical bridge between live network anomalies and offline root cause analysis.

A software or hardware tool that creates synthetic data packets to load a network simulation. It models application-layer behavior by scheduling discrete events for packet generation based on statistical distributions. Common models include:

• Constant Bit Rate (CBR) for voice
• Variable Bit Rate (VBR) for video
• Full-buffer for stress testing
• FTP/HTTP models for web traffic These generators provide the stimulus that drives the DES engine.