Skip to content

Virtual Parameter Service#

The Virtual Parameter Service (VPS) simplifies the calculation of virtual parameters — parameters defined as FDL (Function Definition Language) functions in ECU logging configurations. Traditionally calculated on demand by downstream clients such as ATLAS, the VPS performs these calculations in real time and writes the results back to the broker, removing the need for client-side recalculation.

Key Features#

Feature Description
Real-time virtual parameter calculation Automatically computes FDL-defined virtual parameters as live telemetry data arrives.
Stream API integration Reads source sessions from Kafka and writes computed results back as new virtual sessions.
Session lifecycle management Detects live sessions, creates associated virtual sessions, and manages their full lifecycle.
Prometheus metrics Exposes detailed operational metrics for monitoring via Prometheus and Grafana.
Docker-first deployment Ships as a Docker image with a provided docker-compose.yaml for full-stack orchestration.
Configurable buffering Tunable buffering window and sliding window percentage to balance latency and throughput.

How It Works#

At a high level, the VPS operates as a background service that:

  1. Connects to a Kafka broker via the Stream API.
  2. Listens for live telemetry sessions published by ADS through the Bridge Service.
  3. Reads configuration packets containing parameter definitions and FDL expressions.
  4. Compiles FDL expressions into executable functions at runtime.
  5. Subscribes to source parameter data streams and buffers incoming samples.
  6. Calculates virtual parameter values in real time using the compiled functions.
  7. Writes the computed results back to the Stream API as an associate session.

Downstream clients can then consume these pre-calculated virtual parameters directly from the stream, instead of computing them locally.

flowchart LR
    ADS["ADS (ATLAS)"] -->|Telemetry| BS["Bridge Service"]
    BS -->|Publishes to| K["Kafka"]
    K -->|Sessions & Data| VPS["Virtual Parameter Service"]
    VPS -->|Computed Virtuals| K
    K -->|Virtual Data| C["ATLAS Viewer"]
    VPS -->|Metrics| P["Prometheus"]
    P --> G["Grafana"]

Use Cases#

  • Eliminating redundant calculations — Instead of every client independently computing the same virtual parameters, the VPS computes them once and publishes the results.
  • Ensuring consistency — All consumers receive identical computed values, removing discrepancies between different client implementations.
  • Reducing client load — Downstream applications no longer need to include FDL evaluation logic, simplifying their architecture.
  • Centralised monitoring — Operators can track the health and throughput of virtual parameter processing from a single Grafana dashboard.