Why Photaxon exists

The starting question is simple: how resilient can a network be when it depends entirely on a few central nodes?

Hardware failures, power interruptions, human error, cyber attacks and natural events can disrupt important parts of an infrastructure.

FiberSpider studies a different model in which control, operational state and reconstruction capability are distributed across multiple nodes.

The research focuses on:

• distributed and decentralized network architectures;
• operational continuity for FTTH infrastructure;
• intelligent nodes that can be rebuilt or replaced quickly;
• automatic network-state monitoring;
• AI-assisted diagnostics;
• defensive cybersecurity for telecommunications infrastructure;
• controlled automation and deterministic validation.

In a strongly centralized infrastructure, the loss of a main node can affect the entire network.

FiberSpider studies an architecture in which operational intelligence is distributed across multiple peripheral nodes, represented in the laboratory as intelligent cabinets.

• each node maintains the information needed for its operation;
• an isolated node should not necessarily stop the remaining network;
• a damaged component can be replaced and reconfigured quickly;
• software agents can analyze telemetry, logs and anomalous events;
• AI-generated proposals are checked by the backend before any authorized action.

Network resilience also requires security by design.

• strong authentication for technicians, services and devices;
• least-privilege access;
• separation between public, control-plane and data-plane layers;
• encrypted communication between nodes;
• continuous monitoring and anomaly analysis;
• deterministic validation of AI-agent suggestions;
• isolation of compromised components;
• rapid service recovery after an incident.

No LLM agent is treated as an absolute authority. AI can analyze, classify and propose; independent controls and deterministic rules remain responsible for authorization.

Photaxon studies AI as a support component for technical infrastructure.

Agents can analyze data from networks, sensors, cameras and radio systems to help classify events.

• automated network diagnostics;
• log and telemetry analysis;
• anomaly detection;
• computer vision;
• object recognition and tracking research;
• sensor fusion;
• radio-signal analysis;
• automation of distributed systems.

One experimental research line concerns autonomous platforms and sensor systems for the detection, recognition and monitoring of potential threats to strategic infrastructure.

The work combines telecommunications, resilient communications, computer vision, artificial intelligence and embedded systems.

Fiber-linked platforms are studied as a communication concept for stable, low-latency data transmission that is less dependent on radio links and their interference environment.

The research considers, at a high level:

• sensor-based event detection and classification;
• computer-vision and telemetry analysis;
• resilient data links and infrastructure monitoring;
• Edge AI and distributed processing;
• software simulation of defensive scenarios;
• human supervision, authorization and safety controls.

This is public research information only. Photaxon does not publish procedures for weaponization, targeting, engagement or operational deployment.

• Python for AI, automation, computer vision, data analysis and prototyping;
• C / C++ for real-time and embedded systems;
• Rust for reliable system software and networking;
• OpenCV for image processing and vision research;
• PyTorch and Transformers for AI and computer vision models;
• CUDA for GPU acceleration;
• ROS 2 for robotics and distributed systems;
• PX4 / MAVLink for UAV platform research and simulation;
• Linux for laboratory nodes and experiments;
• FastAPI and PostgreSQL for backend, APIs, telemetry and state management.

Photaxon is an experimental laboratory. FiberSpider, AI agents, computer-vision systems and autonomous-platform studies are proof-of-concept work used to evaluate architectures and technologies before any possible real-world implementation.

Experiments can begin in virtual environments using simulators, containers, software-defined networks and Linux nodes. This makes it possible to study complex systems without requiring dedicated hardware immediately.

The objective is to understand how telecommunications, distributed networks, cybersecurity, artificial intelligence and autonomous systems can be combined into more resilient infrastructure.

The research focuses on systems that can detect problems, adapt to events and continue operating when part of the system is unavailable.

For the FiberSpider architecture, visit FiberSpider Research Laboratory.