Skip to main content

OctoMesh Solution Architecture

This document provides a comprehensive overview of the OctoMesh platform architecture, designed to help developers understand the system components, data flows, and integration patterns.

Platform Overview​

OctoMesh is a cloud-native data infrastructure platform that implements the Data Mesh paradigm. It enables organizations to build decentralized, scalable data architectures specifically engineered for:

  • Industrial Internet of Things (IIoT) scenarios
  • Multi-tenant environments
  • Complex distributed systems
  • Real-time data processing and analytics

Core Philosophy​

OctoMesh treats data as a product with:

  • Domain-oriented ownership: Data is owned and managed by domain experts
  • Self-serve data platform: Infrastructure enables autonomous data operations
  • Federated governance: Centralized policies with decentralized execution
  • Product thinking: Data quality, discoverability, and usability are first-class concerns

High-Level Architecture​

Core Services​

Asset Repository Service​

The central service for managing data products and resources. It provides:

  • Runtime entity storage and retrieval
  • Construction Kit model management
  • GraphQL API for data access
  • Tenant-specific data isolation

Default Port: 5001

Identity Service​

Enterprise-grade authentication and authorization supporting:

  • OpenLDAP and Microsoft Active Directory
  • Azure Entra ID (formerly Azure AD)
  • Google, Meta, and Microsoft OAuth providers
  • Custom identity providers

Default Port: 5003

Communication Controller Service​

The central hub for managing communications between devices and applications:

  • Adapter registration and lifecycle management
  • Secure message routing
  • Device authentication
  • Network expansion capabilities

Default Port: 5015

Bot Service​

Automated assistants for routine tasks:

  • Data preparation and cleaning
  • Aggregation and anonymization
  • User management automation
  • Scheduled reporting

Default Port: 5009

Admin Panel​

Web-based system management interface for:

  • Tenant configuration
  • Service monitoring
  • User and role management
  • System settings

Default Port: 5005

Data Persistence Layer​

MongoDB​

Primary storage for master data with the following characteristics:

  • Dedicated database per tenant for data isolation
  • Stores Construction Kit definitions
  • Stores Runtime Entities (RtEntities)
  • Supports complex queries and aggregations

CrateDB​

Optimized for time-series and streaming data:

  • High-throughput data ingestion
  • Real-time analytics capabilities
  • SQL interface for familiar query patterns
  • Horizontal scalability

RabbitMQ​

Message broker for asynchronous communication:

  • Adapter-to-service messaging
  • Event distribution
  • Pipeline execution coordination
  • Reliable message delivery

Multi-Tenancy Model​

OctoMesh implements strict multi-tenancy at multiple levels:

Tenant Features​

  • Data Isolation: Each tenant has a dedicated MongoDB database
  • Tenant Cascading: Tenants can inherit configurations from parent tenants
  • Adapter Pools: Tenant-specific adapter configurations
  • Role-Based Access: Fine-grained permissions per tenant

Data Model Architecture​

Construction Kit (CK)​

The Construction Kit is OctoMesh's flexible data modeling framework:

ConceptDescription
TypesDefine data entities with attributes, associations, and inheritance
AttributesSimple properties (string, number, boolean) or complex Records
RecordsStructured embedded documents containing multiple attributes
AssociationsRelationships between types (1:1, 1:N, N:M)
EnumsPredefined controlled vocabularies

Construction Kit Libraries​

Libraries are organized hierarchically:

All libraries depend on the System Construction Kit Library, which provides base types and common definitions.

Runtime Model​

Runtime entities are instances of Construction Kit types:

// Example: Runtime Entity structure
{
    "rtId": "unique-identifier",
    "ckTypeId": "Industry.Energy/EnergyMeter",
    "attributes": {
        "name": "Meter-001",
        "location": "Building A"
    },
    "associations": {
        "connectedTo": ["device-rtid-1", "device-rtid-2"]
    }
}

Adapter Architecture​

Adapters connect OctoMesh to external systems for bidirectional data exchange.

Adapter Types​

TypeLocationUse Case
Edge AdapterNear data sourcesLow latency, local processing, offline capability
Mesh AdapterCloud/DatacenterAggregation, large-scale processing, central management

Adapter Terminology​

  • Plugs: Adapters that retrieve data INTO OctoMesh from external sources
  • Sockets: Adapters that provide data FROM OctoMesh to external systems

Data Pipeline Flow​

API Access​

GraphQL API​

Primary API for data access, available at:

https://{host}/tenants/{tenantId}/graphql/playground

Runtime Data Query Example:

query {
  runtime {
    rtIndustryEnergyEnergyMeter {
      items {
        rtId
        ckTypeId
        name
        location
      }
    }
  }
}

Stream Data Query Example:

query {
  streamData {
    tsIndustryEnergyEnergyMeter(
      first: 100
      after: "cursor"
    ) {
      items {
        rtId
        timeStamp
        voltage
        current
      }
    }
  }
}

Authentication​

OctoMesh uses OAuth 2.0 / OpenID Connect for authentication:

  1. Obtain access token from Identity Service
  2. Include token in Authorization header
  3. Token contains tenant and scope information

Deployment Architecture​

Kubernetes Deployment​

OctoMesh services are designed for Kubernetes deployment:

Edge Deployment (K3s)​

Edge components run on lightweight K3s clusters:

  • Minimal resource footprint
  • Offline operation capability
  • Local data processing
  • Automatic synchronization with cloud

Security Model​

Authentication Flow​

Security Features​

  • TLS/SSL: All communications encrypted
  • JWT Tokens: Stateless authentication
  • Scope-Based Authorization: Fine-grained API access control
  • Tenant Isolation: Complete data separation between tenants
  • Audit Logging: Comprehensive activity tracking

Next Steps​