Introduction

PARA//EL is a multi-layer, edge-centric energy management system designed to operate across buildings, devices, and energy infrastructures under one non-negotiable rule:

Occupant comfort and energy flexibility must be enforced simultaneously, never traded off.

It is an operational architecture integrating physical building systems, edge intelligence (Jetson Nano), cloud infrastructure (AWS), user interfaces (web + mobile), and external orchestration (EdgeWare).

Four tightly integrated layers

• Far-edge (Jetson Nano): Connects to devices and executes control decisions locally, even offline.
• Cloud backend (AWS): Aggregates data, stores time-series + registry, serves APIs and apps.
• Web dashboard: Deep analytics, multi-building views, configuration and diagnostics for installers/managers/admins.
• Mobile app: Real-time status, comfort indicators, basic control and alerts for owners/occupants.

Where it operates

• Inside the building: Jetson connected to DERs, HVAC, EV chargers, sensors, and sub-metering — where real decisions happen.
• AWS cloud: Data persistence, user access, APIs, device communication.
• External energy ecosystem: Dispatch orchestration via EdgeWare, plus vendor APIs and LoRaWAN ingestion.

Boundary: PARA//EL does not communicate directly with higher-level platforms like FLEXUS — only via EdgeWare.

What PARA//EL does

• Data acquisition: Modbus/CAN/GPIO, LoRaWAN, cloud APIs, smart meters — normalised into one schema.
• Real-time monitoring: Power flows, SoC, environmental conditions, occupancy, device state.
• Edge intelligence: Comfort (PMV), forecasts, NILM, flexibility envelope, anomaly detection — running locally.
• Control & optimisation: Batteries, heat pumps, EV chargers, inverters — with strict priority order.
• Flexibility execution: Executes dispatch for grid services / DR / P2P, but does not decide market participation.

Next: Open the dashboard demo to review layout and information hierarchy before we build the mobile apps.

Core design principles

• Comfort-first (but not only): PMV quantified and enforced at all times; never overridden by external commands.
• Edge-first: critical decisions occur locally; cloud is supportive, not required.
• Source-agnostic data: Jetson, LoRaWAN, cloud APIs, manual inputs treated equally.
• Actor-based access: everything scoped by Actor, Role, and Building.
• Graceful degradation: local control continues; data buffers; system remains safe.

Security model

• Cognito authentication + JWT tokens
• Device certificates (mTLS to IoT Core)
• Role-based access control (Actor-scoped building permissions)

Deployment model

• Device: Docker containers, OTA updates via AWS IoT.
• Cloud: Serverless backend, auto-scaling.
• Frontend: Web delivery and mobile store distribution.