Друкувати книгуДрукувати книгу

Хмарні рішення для IoT та IIoT (лек)

Сайт: Школа автоматики
Курс: Технології індустрії 4.0
Книга: Хмарні рішення для IoT та IIoT (лек)
Надруковано: Гость
Дата: Friday 22 November 2024 8:20 PM

Зміст

1. Сценарії використання АСУТП в хмарах

•Monitoring & visualization
•Alerting & notifications
•Analytics & insight
•Proactive & intelligent is preferred

2. IoT vs IIOT

•Internet of Things
  • –HTTP / REST
  • –TCP/IP
  • –XML/JSON/Text
  • –STOMP
  • –WebSockets
  • –Simple binary
  • –Simple or no SLA

Industrial Internet of Things

  • AMQP
  • MQTT (Message Queuing Telemetry Transport
  • CoAP (Constrained Application Protocol)
  • Sophisticated binary protocols (Protocol Buffers, …)

3. Key players

4. IoT Cloud

5. Architecture sample IoT Cloud

6. Azure Services Sample

7. Process

  • Capture & store
  • Queueing
  • Streaming
  • Dispatch
  • Kafka
  • IoT Hubs’
  • Amazon Kinesis
  • Azure Event Hub
  • Cloud DataFlow

8. Capture

9. sample

10. Analytics

•Provides high level analytical dashboards
•Has pre-build set of reports and tools
•Simple and fast to use
•Business oriented data
  • HDFS (Spark, Hadoop)
  • Amazon DynamoDB, Amazon S3, Elastic MapReduce, Amazon RedShift
  • Google BigQuery, Google DataStore,
  • Azure HDInsight, Azure Cosmos DB, Azure Table Storage

11. Machine learning

•Processes data using different machine learning algorithms
•Has a collection of already implemented algorithms
•Part of data processing pipeline
•Numeric, Textual, Visual, Audio, Signal
  • Numpy, Keras & TensorFlow and many other
  • Amazon SageMaker
  • Google Cloud Machine Learning Engine
  • Azure Machine Learning

12. IoT Edge Purpose

•Management of devices in the field
•High level features and configuration
•Significant simplification for typical scenarios and devices
•Moves part of compute to the device
•
•Cloud offerings
–AWS Greengrass 
–Azure IoT Edge
–Cloud IoT Edge

13. IoT Edge Key goals

•Cloud vs Local compute
•Local machine learning
•Works during offline
•Near real-time
•Integration with non-OS capable sensors

14. Sample architecture (Azure)

15. Key modules (Google)