Multimedia Traffic Scheduler Simulator

Internet & Multimedia Course — Interactive Teaching Tool

Configure a network traffic scenario, choose a queue scheduling discipline, and observe how it affects latency, throughput, drop rate, and fairness. Use Compare All Mechanisms to see FIFO, Priority, and Round Robin side-by-side on identical traffic.

Scheduling Mechanism

Traffic Parameters

5 pkt/s

Mean number of packets arriving per second (Poisson process)

7 pkt/s

Mean number of packets the server can process per second

50 pkts

Maximum packets allowed in the waiting queue (tail-drop when full)

30 %

Proportion of packets classified as high-priority (e.g. video/VoIP)

300

Total packets to inject into the simulation

Fixed seed ensures identical packet streams across runs (good for comparisons)

Traffic intensity ρ = λ/μ = 0.71 — Queue is stable. Packets will mostly be served in time.

Scheduling Mechanisms Explained

Each mechanism represents a different policy for deciding which packet to serve next when the link is busy. The choice has measurable effects on delay, fairness, and packet loss — especially under congestion.

FIFO — First In, First Out

Simple. Fair by arrival order. Blind to urgency.

Packets are served in strict arrival order with no differentiation between traffic types. This is the simplest possible discipline and forms the baseline for comparison. Under light load (ρ < 1) it performs well. Under heavy load, urgent multimedia traffic must wait behind large background data packets.

Advantages

  • + Simple to implement
  • + No starvation
  • + Predictable order
  • + Fair across flows

Limitations

  • - No QoS differentiation
  • - High delay for real-time traffic
  • - Head-of-line blocking

Priority Queue

Real-time traffic first. Background traffic waits.

High-priority packets (e.g. VoIP, video) jump ahead of low-priority ones (e.g. file transfer). Service is non-preemptive: a packet already in service is not interrupted. This dramatically reduces latency for high-priority flows under moderate load. The risk is starvation of low-priority traffic when the link is heavily loaded.

Advantages

  • + Low delay for high-priority flows
  • + Supports QoS differentiation
  • + Suitable for real-time multimedia

Limitations

  • - Low-priority starvation risk
  • - Less fair overall (lower Jain index)
  • - No benefit if all traffic is high-priority

Round Robin

Take turns. No class goes hungry.

The server alternates between a high-priority queue and a low-priority queue, serving one packet from each in turn. If one queue is empty, the server continues from the other. This eliminates starvation and provides more equitable sharing of link capacity, at the cost of slightly higher delay for high-priority traffic compared to strict Priority.

Advantages

  • + No starvation
  • + Better fairness (higher Jain index)
  • + Predictable service for all classes

Limitations

  • - Higher delay for high-priority vs strict Priority
  • - More complex than FIFO
  • - Less optimal for very time-sensitive flows
Key insight: All three mechanisms perform similarly under low load (ρ = λ/μ < 0.7). Differences become visible as load increases (ρ → 1) and become critical under congestion (ρ > 1). Try the Congested Network preset to observe the trade-offs clearly.