karimelmasry42/ns3-rba-vanet-reproducibility

Purpose and Integration Model

This is an overlay for a pinned ns-3-dev revision, not a standalone application. It reproduces VANET scaling measurements for proof-bearing vehicle-to-RSU authentication over IEEE 802.11p. The overlay is copied into an ns-3 tree; no src/ changes are required. Primary experiment source: scratch/rba-paper-scaling.cc. Build integration depends on scratch/CMakeLists.txt (referenced by the README but absent from the supplied tree).

Entry Points

• Simulation executable: the ns-3 scratch target built from scratch/rba-paper-scaling.cc; accepts topology, timing, beacon interval, and CSV-output arguments.

• Sweep orchestrator: tools/rba/run-scaling.sh; builds if needed, selects vehicle-count schedules, repeats runs, and coordinates aggregation/plotting.

• Analysis/visualization: tools/rba/plot-scaling.py; consumes run-level CSV data to produce grouped node-count summaries and SVG figures.

Simulation Topology and Wireless Boundary

scratch/rba-paper-scaling.cc creates only wireless vehicle OBUs and RSUs. Vehicles move on a multi-lane highway, choose the nearest RSU from initial position, and retain that association. Communications use 802.11p ad hoc networking, fixed communication range, and an OFDM 6 Mbps/10 MHz PHY mode. The modeled network boundary is a single vehicle-prover → RSU-verifier wireless hop.

Authentication Workload Mapping

The simulation represents each vehicle transmission as a fixed-size authenticated payload containing pseudonym identifiers, a compressed Sigma-style ZK proof, timestamp, and signature. Vehicle-side proof/signing and RSU-side verification/eligibility work are modeled as fixed processing-delay constants rather than executable cryptography. Logical TA, pseudonym list, RSU key ledger, blockchain, consensus, and feedback/reputation-update paths are explicitly outside ns-3 node and airtime modeling.

Relevant implementation path: scratch/rba-paper-scaling.cc.

Measurement and Result Semantics

FlowMonitor-derived packet counts and one-way wireless delay drive per-run metrics. Packet loss is computed from transmitted versus received packets; reported RBA end-to-end delay adds fixed prover and verifier processing delays to measured network delay. The simulator appends one row per run to a configured CSV; the result schema also captures topology, radio/payload settings, timing windows, offered load, and wall-clock duration.

Committed example artifacts establish expected output contracts:

• results/rba/rba-scaling-summary.csv

• results/rba/rba-node-count-summary.csv

Sweep, Aggregation, and Plot Pipeline

tools/rba/run-scaling.sh is the control-plane pipeline: it resolves explicit, uniform, or segmented vehicle-count schedules; executes repeated seeded simulations; manages FlowMonitor XML retention; and writes/merges run data. It then invokes tools/rba/plot-scaling.py to group observations by vehicle count, calculate descriptive statistics, and render packet-loss and authenticated-delay scaling figures. Plot artifacts are generated outputs, not simulation inputs.

Runtime and External Dependencies

The material runtime is C++ on ns-3-dev at the documented pinned commit, configured and built with CMake. Batch execution requires a POSIX shell; post-processing requires Python and its plotting/data dependencies as used by tools/rba/plot-scaling.py. There are no live TA, blockchain, cloud, database, or web-service dependencies: those roles are abstract protocol context only.