Aurora

Aurora is an advanced C++20 research framework focused on bio-inspired adaptive transport and cross-layer control for resilient heterogeneous networks. It addresses crucial questions in networking by asking if a transport system can effectively observe failures, prioritize information according to its importance and deadlines, adapt its redundancy, and recover without exceeding designated safety and resource constraints.

Key Features

• Transport System: Implements generation-based forward error correction (FEC) transportation, allowing both fixed and adaptive policies.
• Simulation and Emulation: Offers deterministic simulations and UDP emulation with authenticated process separation for robust testing.
• Cross-layer Control Models: Integrates various models that manage RF, optical, and backscatter links while adhering to energy constraints.
• Feedback Loop Mechanism: Enforces a structured feedback loop where outcome data influences future generation plans, enhancing adaptive capabilities.

intent + deadline + traffic class
              ↓
      runtime-selected policy
              ↓
 generation plan + FEC protection
              ↓
 authenticated sender → receiver
              ↓
 receiver-local delivery outcome
              ↓
 terminal feedback → next plan

Research Insights

Aurora separates policy from execution, which allows for stable comparators and various operational modes, including a biological-adaptive approach that evolves based on performance feedback after failures. Initial experiments indicate that all policies maintained critical delivery before and after imposed network failures, with observable adjustments in the biological policy enabling improved protection and efficiency.

System Structure

The repository is organized into several directories, each serving distinct purposes, including:

• apps/: Executables and benchmark front ends
• include/aurora/: Transport and policy APIs
• src/: Compiled implementations and codec adapters
• benchmarks/: Frozen traces, designs, and measurement contracts
• tools/: Planning and evidence visualization tools
• tests/: Unit and regression tests
• docs/: Study protocols and research notes

Research Evidence

Aurora operates within a disciplined evidence framework that categorizes results from basic unit tests to comprehensive field data collection, ensuring rigorous validation.

• For more detailed experimentation and evidence, refer to the technical reference and the live evidence explorer.

This project is not intended as a production network stack; its primary contributions are grounded in controlled simulation and emulation evidence, paving the way for future developments in adaptive networking technologies.