Evaluation

Experimental & simulation results

Evaluations were carried out in Mininet-WiFi (emulation) and MATLAB (simulation), with ns-3 used to select a TCP benchmark. Reported results indicate that A³L-FEC reduces age violations relative to TCP-BBR and ACP+, while keeping packet delay close to the network's propagation delay.

Placeholder data. The charts below use illustrative placeholder values structured to match the figures in the papers. Replace the arrays in assets/data/results.js with your exact measured values — the charts update automatically. Claims here are intentionally careful ("reported results indicate").

Age violation rate

A³L-FEC-VSVB vs TCP-BBR

Age violation probability versus block length (chunk number) under two age-violation thresholds. Lower is better.

Age violation — AVT = 600 ms

Mininet-WiFi · 1 Mbps, 30 ms, 5% loss
Emulation
500 samples / run3 realizations

Age violation — AVT = 150 ms

Stricter freshness threshold
Emulation
500 samples / run3 realizations
AoI & delay

Freshness and latency behaviour

Average Age of Information and average packet delay. A³L-FEC's fire-and-forget UDP design keeps delay near the propagation delay.

Average AoI

Lower means fresher data at the receiver
Emulation

Average packet delay

A³L-FEC ≈ 30 ms (≈ propagation); BBR ≈ 130 ms
Emulation
Coding parameters

Effect of coding rate & chunk number

Redundancy improves decoding reliability but adds load and queuing delay — so there is an optimal coding rate. More chunks per sample also raise age violation as block length grows.

Effect of coding rate

Age violation vs block length n (per k)
MATLAB
50 runs averagedFCFS bottleneck

Chunk number & threshold

A³L-FEC-VSVB under different AVT
Emulation
2000 samples / run3 realizations
Baselines

Comparison with related protocols

A³L-FEC-FSFB versus ACP+ in MATLAB, and the average AoI of TCP variants in ns-3 that motivated BBR as the emulation benchmark.

A³L-FEC-FSFB vs ACP+

Average age violation (AVT = 5) — lower is better
MATLAB

TCP variants — average AoI

BBR's short queues give the best AoI → benchmark
ns-3

Evaluation setup

MATLAB simulation
Deterministic FCFS bottleneck, buffer 5000 chunks, monitoring intervals of 10² slots, runs of 10⁵ slots, 50 repetitions; loss varied over {0, 0.1, 0.2}.
Mininet-WiFi emulation
Source, sink and four switches; bottleneck 1 Mbps with a 1000-packet buffer; 30 ms forward delay, zero-delay feedback path, 5% loss on the bottleneck; C++ UDP client/server.
ns-3 benchmark
One source, two routers, one sink; 10 Mbps bottleneck with DropTail; used to rank TCP variants by AoI (BBR best).

Placeholders for future assets

Slots reserved so you can drop in final material later without redesign.

Raw datasets
CSV / .mat exports
Demo videos
Emulation captures
Source code
Reference implementation
High-res figures
Publication-quality