Siemens Industrial Edge on AWS Reference Architecture (Part 1 of 2)
This reference architecture provides a baseline of integrating Siemens Industrial Edge with AWS services.
Siemens Industrial Edge Devices run
Factory AWS Cloud RedShift Spectrum 1 Southbound Connector which connects to field devices and gateways through 4 Amazon Amazon AWS IoT Notification Redshift QuickSight different protocols. Connectors are ML Inference Events deployed and distributed from Siemens Siemens (DWH (BI & Visualization) Siemens (Complex Events) Industrial Edge Management to Industrial Edge Industrial Edge 5 Structured data) Cameras Device Management 3 Siemens Industrial Edge devices. 2 Data Lake AWS AppSync OPC-UA Cloud connector application Amazon AWS Lambda 2 establishes connectivity to AWS and AWS IoT Amazon Kinesis Athena Industrial Cloud Amazon S3 AWS Glue Amazon S3 transfers data using MQTT protocol Equipment/ Core Data Firehose Connector (Raw data) (ETL job) (Processed over a secure TLS layer. Machines data) Amazon API 1 Amazon Gateway Data ingested into AWS IoT Core is Profinet App Neptune 3 streamed to AWS IoT SiteWise, PLC/DCS Repository Amazon Kinesis Data Streams, and + additional AWS IoT Amazon AWS IoT Events. Southbound Amazon AWS Notification Notifications industrial SiteWise SNS Connector Kinesis Fargate protocols Amazon Data (Analytics) AWS IoT Events processes the data Streams SageMaker Web App 4 and sends notifications based on 6 business logic. Historian AWS CodePipeline App 1..N VPC – Business Apps Amazon Amazon Kinesis Data Firehose streams Cognito 5 data directly into an Amazon S3 bucket, then AWS Glue extracts and I.E API or Private subnet transforms data to be consumed by AWS CodeCommit AWS CodeBuild Docker via IoT Jobs Amazon SageMaker or Amazon images Amazon MES / Config files CloudFront Redshift. Elastic Load Balancing The industrial Edge API and AWS IoT 6 jobs processed in the CI/CD pipeline S3 Bucket publish Docker images and AWS Transfer AWS Glue Amazon EC2 configuration files to Siemens for SFTP Amazon ECR Industrial Edge Management. Local Data Amazon Amazon Amazon AWS DataSync AWS Database Route 53 RDS EBS Migration Service
Web App ML Inference Asset Data Siemens Industrial Siemens Industrial Fleet Asset Perf Closed-Loop Manage my Control Perf Visual Visual Process Event Connector ingests data directly into Cameras Edge Edge Manager Manager Monitor Simulation Explorer Machines Analytics Explorer Anayzer Anayzer 2 MindSphere through Rest API of Device Management Amazon Siemens MindSphere (MindConnect) 1 MindSphere IoT MindSphere Metadata MindSphere Cognito OPC-UA & Access Control Analytics
Industrial Cloud AWS IoT Core
(MindSphere Amazon Kinesis Amazon S3 Amazon IoT Modeler Equipment/ Connector 2 Data Firehose DynamoDB App MQTT) Aurora APIs Amazon Siemens MindSphere application Machines Repository CloudFront 3 container registry uses Amazon Elastic Asset Modeler Predictive Container Registry (Amazon ECR) to Profinet Learning deploy analytics apps and models to MDSP PLC/DCS Connectivity Siemens industrial Edge. + additional Rules Notification Amazon SES Data Sharing S3 Bucket Southbound Semantic industrial Connector Queries protocols Ontologies Batch and bulk upload to MindSphere 4 BYO Amazon 4 MindSphere Data Lake & Processing Data lake is possible through Industrial AWS IoT Model Route 53 Edge applications. Greengrass Catalog & Historian Search IDL SDI Data VFC Eng. Digital Twin Closed-Loop 6 Customers train their own models in App 1..N Docker 5 Amazon SageMaker from MindSphere images MindSphere App Container Registry 5 MES Data Lake exported data. ML inference / Config files can then be deployed on MindSphere 3 Model Amazon ECR or on the edge. Deploy ML Model
MindSphere AWS Cloud MindSphere Data lake integration with
Edge Amazon SageMaker 6 AWS services makes it possible to Local Data Streaming import and export data into or out of (ML) AWS Amazon S3 AWS IoT MindSphere. AWS Transfer AWS IoT AWS Glue Amazon Redshift Amazon QuickSight for SFTP DataSync Core SiteWise (ETL job) (DWH (BI & Visualization) Structured data)
