3. IIoT Reference
Architechure & usecases
• Industrial Data for Predictive Quality
• Industrial Data for Asset Condition
Monitoring
• Industrial data for Predictive Maintenance
• Industrial Data for Asset Tracking TPT
Monitoring
• Industrial Data for Pricing Model
4. Sources of Machine Data
Sensors Caluclations
Predictions
Automation
Metadata Decisions
Transactions
10. AWS IoT Greengrass Architechure
OPC-DA machine
OPC UA machine
Other Assets
Gateway Server
Ethernet
AWS IoT Greengrass
Industrial Datacenter
Other protocols
OPC UA protocols
OPC DA protocols
OPC UA STREAM
OWN STREAMING
PROTOCOL
Machine to cloud
data publisher
Streaming Manager
Connection Layer
OPC DA STREAM
IOT Connectivity HUB
13. IIoT ML Layer
1. Setting up storage
2. Moving data
3. Preparing and
cataloguing data
4. Configuring
security policies
5. Making data
available for
consumption
17. Data Driven Architectures
• Data Mesh vs. Data Fabric
• Data Lake vs Data Hub
• How to Choose ?
Data Mesh: API-driven [solution] for developers, unlike [data] fabric
Data Fabric: low-code, no-code, which means that the API integration is
happening inside of the fabric without actually leveraging it directly
Data hub as an architectural concept is different from data hub as a database
Finally, they are architectural frameworks, not architectures. You don’t have
architecture until the frameworks are adapted and customized to your needs, your
data, your processes, and your terminology
18. Technologies Involved and Justification
Technologies Tools
Architechure Framework for IIOT Data Mesh
Streaming IOT layer rules AWS MSK and Kinesis
Serverless Lambda
ML Engine workspace Sagemaker, EMR, AWS ECS,
Database DynamoDB, Redshift
Dashboards Amazon Quick Sight
ETL & ELT Aws Glue, EMR
Binary Management Jfrog Artifactory
Monitor Cloudwatch & Prometheus
CI/CD Pipelines Jenkins
Graphdatabases Neptune
Serverless Query AWS Athena
Infrastructure as Code Terraform
Cataloging AWS LakeFormation
Why Kafka ?
1. Supports replay, analytics, and
machine learning based use cases.
2. Reduced Cost
3. Integration with Enterprise Data
Lake
4. Optimize Machine Learning
Workloads
5. Compliance Backloading data to
other system