In the traditional world of network monitoring, a single software crash can blind an entire measurement system. During the AIORI-2 Hackathon, team SYN/ACK from Amity University Greater Noida challenged this status quo by building the Modular Internet Measurement Network over NATS.
By implementing the core principles of RFC 2330 (IPPM Framework) using a cloud-native message bus (NATS.io), the team created a “hot-pluggable” architecture where measurement modules—like pings or DNS lookups—can be added, removed, or updated without ever restarting the central management service.
1. The Architecture of Decoupling
Most measurement systems are “monolithic,” meaning all parts are tightly linked. Team SYN/ACK broke this mold by separating the Anchor (Management Plane) from the Measurement Modules (Data Plane).
- The Anchor: Acts as the brain, discovering new modules and managing job schedules via NATS subjects.
- The Modules: Independent workers that “plug in” to the message bus. If a module crashes, the Anchor remains unaffected, and other modules continue their work.
- NATS.io Message Bus: The central nervous system that routes jobs to modules using Queue Groups for automatic load balancing.
2. Technical Implementation: Cloud-Native Measurement
The team utilized a modern stack to ensure the system was ready for distributed deployment across the AIORI Testbed.
| Component | Technology | Role |
|---|---|---|
| Messaging | NATS JetStream | Ensures persistent queuing and reliable job delivery. |
| Discovery | FS Watcher / NATS | Enables “Hot-Plugging” (Auto-discovery of new modules). |
| Security | JWT / TLS 1.3 | Securely authenticates modules using RFC 7519 standards. |
| Visualization | WebSockets | Streams live measurement results to a real-time dashboard. |
3. Key Findings & Performance Metrics
The primary goal was to prove that a decoupled system is more resilient than a traditional one. The results confirmed that modularity does not come at the cost of speed.
- Hot-Plugging Speed: New modules were discovered and began processing jobs within 2–3 seconds.
- Fault Isolation: Injections of module crashes resulted in 0% impact on the rest of the system.
- Processing Latency: Local job overhead remained under 50ms, proving NATS is efficient enough for high-speed telemetry.
- Horizontal Scaling: By spinning up multiple instances of a “Ping” module, the workload was automatically shared via NATS queue groups.
4. Overcoming “The Discovery Race”
A major technical hurdle was a race condition where modules would try to announce themselves before the Anchor was ready to listen.
“Initially, we struggled with modules ‘shouting’ into a void. We solved this by implementing a heartbeat-based discovery protocol, ensuring the Anchor and modules perform a reliable handshake before jobs are dispatched.” — Aman, Backend Architect
5. Impact and Future Horizons: The “Sky-Path”
This work provides a blueprint for the AIORI-IMN framework to scale globally. The team is already looking toward 2026 with plans to:
- Expand the Ecosystem: Build modules for QUIC performance (RFC 9000) and BGP monitoring.
- Federated NATS: Explore connecting multiple NATS clusters across different geographic regions for a global measurement view.
- Internet-Draft: Formalize the NATS subject schema used in this project as a potential best practice for the IPPM Working Group.
6. Team Reflections
The name SYN/ACK represents the perfect handshake between research and execution.
“Seeing our modular vision come to life on the dashboard, where a module could simply ‘appear’ and start working, was incredibly rewarding.” — Pragyensh Pritiman Panda, Team Lead
Read the full report
