OPC UA Server

Integration

Agenda

1. Fetching data from OPC UA
2. Location of OPC UA Client
3. Cloud event ingestion pipeline

Fetching data from OPC UA

Subscription-based approach

Fetching data from OPC UA

Subscription-based approach

Pros:

• No unnecessary calls to OPC UA Server
• Instant granular updates exactly when variable changes

Cons:

• More complex architecture and need for aggregation of single variable updates to project state of the system
• Need to maintain state of the system locally as a part of client
• Needs failover scenario

Fetching data from OPC UA

Subscription-based approach with incremental updates

Fetching data from OPC UA

Subscription-based approach with incremental updates

Pros:

• No unnecessary calls to OPC UA Server
• Instant granular updates exactly when variable changes
• Simple implementation of OPC UA Client

Cons:

• More complex processing architecture on the backend
• Needs failover scenario

Fetching data from OPC UA

Periodic fetch approach

Fetching data from OPC UA

Periodic fetch approach

Pros:

• Simpler architecture
• No need to maintain local state as part of the OPC UA Client

Cons:

• Unnecessary calls to OPC UA Server when no changes are observed
• Needs failover scenario

Placement of OPC UA Client

Co-located with AGV and OPC UA Server

Placement of OPC UA Client

Co-located with AGV and OPC UA Server

Pros:

• Lower latency between OPC UA Server/Client
• Better resilience to outages in Internet connection

Cons:

• More complex architecture at the edge
• More challening maintenance of OPC UA Client at the edge
• Need to create multiple clients if there are AGVs that are not co-located

Placement of OPC UA Client

Located in Cloud environment

Placement of OPC UA Client

Located in Cloud environment

Pros:

• Simpler architecture
• Can support data ingestion from multiple OPC UA Servers

Cons:

• Higher communication latency between OPC UA Server/Client
• Less resilience to Internet connection outages
•  

Ingestion pipeline

IoT-hub based - managed services

Ingestion pipeline

IoT-hub based - managed services

Pros:

• Managed Services
• Elasticity due to additional data store
• Ability to extend via Stream Analytics

Cons:

• Lock-in to Azure-specific services
• Potential limitations of Azure-specific services

Ingestion pipeline

Simplified - managed services

Ingestion pipeline

Simplified - managed services

Pros:

• Managed Services
• Simplicity
• Possibility to extend by evolving architecture

Cons:

• Lock-in to Azure-specific services
• Potential limitations of Azure-specific services
• It is recommended by Microsoft to use IoT Hub 

Ingestion pipeline

HDInsight-based - self-managed services

Ingestion pipeline

HDInsight-based - self-managed services

Pros:

• Elasticity and support for multiple tools
• Extensibility
• Lack of lock-in to Azure-specific tools
• Can still take advantage of Azure IoT Hub

Cons:

• Increased maintenance - lack of Azure-managed services
• Longer time to implement

Potential format of data

• JSON
• CSV
• Apache Parquet (Column-oriented)

Potential shape of data

granular updates to specific variables

Potential shape of data

full state of the system at each event

Jetson Jetbot