Skip to Content

PLATFORM ARCHITECTURE


A TECHNICAL DEEP DIVE FOR IT ARCHITECTS, INTEGRATORS, AND CTO-S


This page explains what sits inside Vela Context Data Hub, why each component exists, and how the pieces fit together. If you're evaluating Vela for a production deployment — or you need to convince your technical team — start here.

WHY ORGANIZATIONS STRUGGLE WITH DATA TODAY

Cities, utilities, industrial operators, and integrators already generate massive amounts of data — from open datasets, sensors, utilities, ERP exports, and public services. The problem isn't data volume. It's that data speaks different languages across every system, every vendor, every department.

  • Data is usually published in raw CSVs or Excel files without machine-readable semantics.
  • External developers must spend too much time cleaning and transforming data before they can build apps.
  • Each new project requires intensive back-and-forth communication between city IT teams and vendors.
  • Data silos exist: traffic, energy, water, manufacturing, environment, and GIS -all managed separately, not shared across departments.
  • Data comes from too many sources (IoT, CSV, ERP exports, APIs) with inconsistent structures.
  • Every integration invents its own JSON, naming, units, and timestamps.
  • Without a standard context model, datasets do not connect (energy + building + mobility remain separate worlds).
  • Teams waste time mapping and re-mapping instead of building applications. 

The result is high integration cost, fragile solutions, and vendor-specific data silos that compound with every project.



 WHAT IS A CONTEXT DATAHUB (AND WHY IT MATTERS)




A Context DataHub is a standardised “context layer” that stores and serves real-time entities (assets, sensors, meters, rooms, vehicles,...) through one open API.

What it ensures:

  • Data is structured as linked entities (JSON-LD, NGSI-LD) instead of arbitrary payloads.
  • All systems speak the same language - Smart Data Models, consistent attributes, timestamps, and units.
  • Applications integrate faster because the interface stays stable even when sources change.
  • You can scale from one use case to many without rebuilding integrations.

What every Context Data Hub needs at minimum:

  • A NGSI-LD Context Broker (real-time context API).
  • Connectors/ingestion pipelines (streams + batch imports).
  • Smart Data Model mapping (so meaning stays consistent across sources).
  • A security gateway (auth + API protection).
  • Monitoring and evidence logs ( so operations teams can sleep at night ).

COMPONENTS OF vela CONTEXT DATA HUB

Vela Context Data Hub is built on ETSI NGSI-LD standards and ships with every component a production-grade context platform needs. No assembly required. Components:

  • NGSI-LD Context Broker - the real-time engine for context queries, subscriptions, and entity management. Based on Scorpio.
  • Smart Data Models  - standard formats for energy, buildings, mobility, water, environment, industry, and logistics. Import, extend, or define your own.
  • Metadata Catalog (DCAT-AP, CKAN) - makes your datasets discoverable and harvestable by EU and national data portals.
  • Security Gateway -  OAuth2/OIDC authentication, API access control, reverse proxy. Powered by LIBRA.
  • Connectors & Harvesters - periodic imports (CSV, Excel, APIs) and real-time streams (sensors, MQTT, OPC-UA). Powered by CASSIOPEIA.
  • Time-series Storage - historical data queries and analytics.
  • Dashboards & Developer tools - Grafana, custom dashboard builder, and debugging UIs for exploration and operations.
  • Evidence Logs -​ who changed what and when, with a full audit trail.
  • Validation and Normalization - enforce units, shapes, and timestamps so partners can't create local dialects.
  • Temporal Rules - ObservedAt, late-data handling, retention policies — so history is predictable.


WHO THIS PAGE IS FOR



For Decision Makers

A low-risk path to interoperability:

  • Start with one system (one dataset if you need to).
  • Reduce integration cost for every subsequent project.
  • Avoid vendor lock-in by committing to an open standard API (NGSI-LD), not a proprietary platform. 

For IT Architects & Integrators

A ready-to-use platform, no assembly required:

  • No need to build a context platform from  components and glue.
  • Standard models, production connectors, and repeatable onboarding patterns.
  • Secure, documented, and maintainable -  built to hand off to operations teams.

technical highlights

STACK COMPONENTS
  • Context Broker -  Scorpio (NGSI-LD)
  • API Gateway -  LIBRA (policy enforcement, rate limiting, RBAC)
  • Data Transformation -  CASSIOPEIA (no-code NGSI-LD mapping)
  • Identity & Access -  Keycloak (OAuth2/OIDC)
  • Database -  PostgreSQL + time-series extensions, ClickHouse (option)
  • Workflow / Integration -  Node-RED, Python
  • Dashboards -  Grafana & Custom
  • Logging - Loki
  • Metadata Catalog -  CKAN + DCAT-AP
  • Operating System -  SandBoxOS (130+ pre-loaded packages)
INTERFACES & STANDARDS
  • ETSI NGSI-LD / JSON-LD as the primary interface for all applications and integrations
  • Smart Data Models (smartdatamodels.org) as the default semantic vocabulary
  • DCAT-AP for metadata catalog exchange
  • Built-in monitoring and alerting
  • Optional validation and normalization rules for strict governance environments

READY TO SEE IT WITH YOUR DATA?

The free 30-minute strategy call maps your data sources and delivers a concrete integration roadmap.

BOOK YOUR FREE STRATEGY CALL

Or see pricing and plans.