How MeshCloud
is architected.

Devices connect through Azure IoT Hub, with provisioning at scale handled by Device Provisioning Services and certificate lifecycle management built in. The architecture supports multiple device protocols and can consume from any broker, protocol, or data format with minimal effort, so brownfield fleets connect alongside greenfield deployments, with private networking available out of the box. Telemetry flows through a Container App and Event Hubs pipeline load-tested to one million devices, with commands routing back to devices through a parallel pipeline orchestrated by the same patterns.

Time-series data lands in Azure Data Explorer or Microsoft Fabric for analytics, while operational state, asset metadata, and digital twin data live in Cosmos DB, each store tuned for what it does best. Device models follow DTDL v3, with configuration management and full historization baked in, so you can query not just what a device reports now, but what it reported and how it was configured at any point in time. A secure HTTP API exposes all device data over service-principal authentication, simplifying KQL while preserving its querying power, so any app or cloud-to-cloud integration can consume it without learning a custom query language.

An Engineering Portal built in React ships out of the box, with reusable UI components and dynamic forms for device operations and health monitoring. Grafana, Power BI, and Microsoft Fabric handle dashboards and analytical surfaces. A YARP-based Backend-for-Frontend layer with JWT auth and standardized error normalization gives your team a clean integration point. Extension to source code where scoped lets you build exactly the customer experience your product needs.

Modern auth patterns run throughout the platform: service principals, managed identities, and Azure Entra integration, with customer IDPs extending the auth layer where scoped. OpenTelemetry-native instrumentation covers structured logging, distributed tracing, and alerting, with the ability to swap monitoring backends without code changes. Audit trail is on by default. Deployment automation and infrastructure-as-code ship with the platform, alongside automated test suites that exercise the core paths.