​
Proxus Architecture

Proxus is built on a distributed edge architecture designed for scalability, performance, and resilience. The architecture consists of the following key components:
  • Hub: A Blazor server application (C# .NET 9) that serves as the central management and control point for the Proxus platform. The Hub provides a web-based interface for:
    • Device Management: Adding, configuring, and managing devices.
    • Data Flow Configuration: Defining data collection rules, processing pipelines, and data destinations.
    • User Management: Managing user accounts and permissions.
    • Monitoring and Alerting: Monitoring the health and performance of the platform and generating alerts when issues occur.
    • Data Visualization: Creating dashboards and reports to visualize data.
  • Edge Services: Lightweight, distributed services that run at the edge, close to the data source. Edge Services are deployed as Docker containers for easy deployment and management. They are responsible for:
    • Data Acquisition: Collecting data from connected devices using various protocols.
    • Data Processing: Transforming, filtering, and aggregating data.
    • Local Analytics: Performing real-time analysis and generating insights.
    • Data Buffering: Buffering data in case of network connectivity issues.
    • Secure Communication: Encrypting data and authenticating with the Hub.

​
Communication

Communication between the Hub and Edge Services is facilitated by NATS Jetstream, a high-performance, distributed messaging system. NATS Jetstream provides:
  • Reliable Messaging: Ensures that messages are delivered even in the event of network outages.
  • Scalability: Supports a large number of concurrent connections and high message throughput.
  • Low Latency: Minimizes the delay in message delivery.
Edge Services connect to the Hub using leaf nodes, which provide a secure and efficient way to connect to the central NATS cluster.

​
Data Storage

Data storage is handled by TimescaleDB, a time-series database built on PostgreSQL. TimescaleDB is optimized for:
  • High-Volume Data Ingestion: Supports high write throughput for storing large amounts of time-series data.
  • Efficient Querying: Provides optimized query performance for analyzing time-series data.
  • Data Retention Policies: Allows users to define data retention policies to automatically delete old data.
  • Data Compression: Supports data compression to reduce storage costs.

​
Data Flow - Detailed Explanation

  1. Device Configuration: A user configures a device within the Proxus Hub, specifying the following parameters:
    • Protocol: The protocol used to communicate with the device (e.g., Modbus TCP, OPC UA).
    • Connection Parameters: The connection details for the device (e.g., IP address, port number).
    • Data Sampling Rate: The frequency at which data is collected from the device.
    • Data Destination: The destination for the collected data (e.g., TimescaleDB, an external API).
  2. Edge Service Deployment: The device configuration is deployed to an Edge Service that is located near the device.
  3. Data Collection: The Edge Service connects to the device using the specified protocol and collects data at the configured sampling rate.
  4. Data Processing: The Edge Service processes the collected data using configured rules and User-Defined Functions (UDFs). This can include:
    • Filtering out irrelevant data.
    • Converting data to a standard format.
    • Aggregating data over time.
    • Performing calculations on the data.
  5. Data Storage: The processed data is stored in TimescaleDB.
  6. Data Visualization and Analysis: Users can visualize and analyze the data using the Proxus Hub’s built-in dashboards or external tools connected to TimescaleDB.

​
Architecture Highlights

Proxus is built on modern architectural principles that ensure high performance, scalability, and reliability:
  • Actor Model: For concurrent and distributed processing
  • Message-Based Communication: Ensuring loose coupling and scalability
  • Stateless Design: Enabling true horizontal scalability
  • SIMD Optimization: For high-performance data processing
  • Container-Native: Full Docker and Kubernetes support
Next: Key Features Previous: What is Proxus?