Skip to content

Scenario: large-scale

Class: large / mobile

← Benchmark index · Metrics · Methodology

Provenance — regenerated on every merge. Unlike the sweep pages, this table is not pinned to a release: benchmarks.yml re-runs the discrete taxonomy on every merge to the default branch and rewrites the generated block below, so these numbers track the current tip of main and already include the pheromone-weighting change (#327) and the per-seed RNG stream assignment (#352). They are a per-merge regression signal at the 10-run publication floor (#318); the release-pinned, 20-seed numbers live on the sweep and headline pages. See Provenance and #365.

What it stresses

100 nodes. The paper base scenario grown in both node count and terrain, so paths get longer and control traffic has more nodes to reach — the discrete counterpart of the scale sweep.

Configuration

nodes area long edge (m) max speed (m/s) pause (s) flows propagation
100 2100 20 30 20 range

Defined as DISCRETE["large-scale"] in ns3/tools/run-scenarios.py (scenario=paper nNodes=100 areaX=2100 areaY=700). The per-merge benchmarks workflow runs the taxonomy with --quick.

Results

Caveat — large-scale and heavy-load currently reflect an open defect. #173 (P1, open): with enableMultipath on, reactive forward ants are bounded neither by a broadcast budget nor by (src,seq) duplicate suppression, so route discovery floods combinatorially in dense graphs. Read those two rows as a bug report, not as protocol character. See docs/fidelity.md.

Mean of 10 run(s), every baseline on identical realisations; ± is the 95% CI half-width (#293). Generated by run-scenarios.py.

Measured at 1a7c32b, run.

protocol PDR % ±95 mean delay (ms) ±95 99th delay (ms) ±95 throughput (kbps) NRL ±95 jitter (ms) dOff90 (ms)
anthocnet 81.1 ± 1.5 120.5 ± 13.4 1500.7 ± 211.1 4.33 109.517 ± 8.00 171.92 inf
aodv 68.1 ± 3.3 126.3 ± 23.0 1827.3 ± 356.4 3.64 156.438 ± 9.73 189.35 inf
olsr 62.5 ± 4.1 43.8 ± 5.7 1019.5 ± 3.4 3.31 18.866 ± 1.57 65.85 inf
dsdv 57.4 ± 4.0 64.7 ± 9.2 1139.0 ± 163.2 3.07 172.200 ± 13.86 102.64 inf