Data Storage & Processing: Azure Cloud

• Data storage and processing platforms include Microsoft Azure Cloud, Google Cloud Platform (GCP), and Amazon Web Services (AWS).

Microsoft Azure Cloud

• It is an end-to-end IoT platform for connecting, monitoring, and controlling billions of IoT devices using cloud services.
• It has back-end service capabilities to receive and process telemetry data at scale, analyze data in real-time or batch, send commands to IoT devices, and monitor/control device states.
• Azure IoT Hub serves as a central communication point between IoT devices and the cloud, supporting secure, bidirectional messaging and providing device libraries for multiple platforms/languages.
• It includes an identity registry and device identity management.
• Stream Analytics monitors data streams for patterns/anomalies and executes actions based on event conditions, using SQL-like rules.
• Azure Machine Learning Studio enables pattern recognition, and trained models can be deployed in Stream Analytics for advanced detection, useful for predictive maintenance and anomaly detection.
• Storage options include SQL Database, Data Warehouse, and DocumentDB.
• Visualization is done via PowerBI for dashboards and insights.
• All Azure services work together within the same infrastructure, enabling fast, secure, and scalable IoT solutions.

Google Cloud Platform (GCP)

• Cloud service suite for storage, analytics, and machine learning.
• Operates on the same infrastructure as Google Search and YouTube, where Google handles infrastructure, scaling, and storage.
• GCP's advantages include a fast global network, competitive pricing, broad service support, and strong IoT integration with Firebase and Cassandra.
• It has scalable infrastructure to handle millions of devices and large data volumes.
• Android Things provides a lightweight OS for IoT with OTA updates.
• Cloud IoT Core manages device connectivity and data ingestion at scale.
• Devices publish data to Google Pub/Sub via MQTT.
• Cloud Functions react to events and route data.
• Data is stored in Firestore (real-time NoSQL DB) and BigQuery (massive-scale SQL querying).
• Cloud ML trains models using collected data, which then can be exported for inference on other platforms.
• Pokémon Go used a predecessor to Firestore for real-time, location-aware features.

Amazon Web Services (AWS)

• Comprehensive cloud computing platform with services like EC2 (compute) and S3 (storage).
• It provides scalable, low-cost infrastructure with automatic provisioning, load balancing, and monitoring.
• Devices publish their state to a message broker via topics and device info is stored in a 'Thing Registry'.
• AWS IoT Device SDK enables secure device connection and messaging, supporting C, JavaScript, and Arduino.
• Device Gateway manages connections using MQTT, WebSockets, and HTTP, supporting over a billion devices.
• Message Broker is a pub/sub system for secure and timely message delivery with fine-grained access control at the topic level.
• Security is achieved through mutual authentication and encryption using SigV4, X.509 certificates, and token-based authentication.
• Amazon Cognito connects IoT Core with mobile apps and provides temporary AWS credentials.
• Device Shadow stores last reported and desired future device states, allowing apps to read/write device state even when the device is offline.
• Rules Engine automates actions based on incoming IoT data, supporting SQL-like syntax, and can route data to multiple AWS services.
• Integration with AWS Lambda allows execution of custom code (Java, Python, Node.js) for advanced IoT workflows.