IoT (Internet of Things) refers to a system in which various physical objects are embedded with sensors and communication capabilities to collect and transmit data over wireless networks。 The Azure Architecture Center proposes a reference architecture for IoT that outlines best practices for building scalable, secure, and reliable IoT solutions on Microsoft Azure. This architecture typically includes the following key components:
Azure IoT Reference Architecture
Microsoft's suggested architecture with Azure PaaS components is as follows
we convert the format of data from edge devices and IoT devices, divide the data into cold and warm paths, and apply machine learning to combine pre-collected data with device data to derive business insights.
Cold path or warm path? It can be categorized according to how the measurement data is processed. Hot path analyzes measurement data in near real-time. It is implemented with an engine that processes streaming data. The results of processing are recorded in a structured format that can be triggered by alarms or queried using analytics tools. The cold path sends measurement data to a batch job and performs actions at intervals. It is typically applied to large amounts of data.
Architecture Component descriptions
Azure IoT Edge: IoT devices that perform some data processing on the device itself or on a field gateway can be managed through Azure IoT Edge. You can batch deploy code to devices or manage devices remotely.
IoT Hub: Acts as a cloud gateway. It hosts cloud services that process events coming from devices. It can be understood as a broker between IoT devices and backend services. It provides features such as secure connectivity, event collection, two-way communication, and device management.

IoT Hub Delivery Protocols
▶MQTT 3.1.1
▶ MQTT over WebSockets
▶ AMQP 1.0
▶ AMQP over WebSockets and
▶ HTTPS
IoT Hub Device Provisioning Service (DPS): With DPS, you can assign devices to specific Azure IoT Hub endpoints and auto-provision IoT devices in a highly scalable way.
Azure Stream Analytics: Aggregate stream data and combine external data sources to run complex analytics.
Cosmos DB: A PaaS database that can be used as warm path storage is Cosmos DB. Cosmos DB is globally distributed and easy to scale. You can easily find another scenario where Cosmos DB is used as a warm path in the IoT Architecture proposed by Azure. Architecture utilizing CosmosDB as a warm path DB
Azure Blob Storage : Use Azure Blob Storage as the cold path storage. With IoT Hub message routing, you can store IoT device messages in Azure Blob Storage, allowing Blob Storage to serve as a low-cost, long-term cold data store. See > Storing data with Azure Blob Storage at the IoT Edge
protocol gateway: You might need to manipulate or aggregate telemetry stream data. For example, protocol transformations, such as converting binary data to JSON or combining data points. If the data needs to be transformed before it reaches the IoT Hub, you can use a protocol gateway to handle it. If the data needs to be transformed after it reaches the IoT Hub, you can also trigger an Azure Functions function with an event that occurs in the IoT Hub.
Azure Logic Apps: These are services that help you schedule and automate tasks or business processes and workflows that are required when you need to integrate data into your services. You can also define actions that raise alerts or send emails or SMS messages.
Azure Active Directory: Manage who can access your services through Azure Active Directory.
Device Connectivity
There are four ways to connect your device to the cloud gateway.
↓ ↓ ↓ ↓ ↓ ↓ ↓ ↓ ↓ ↓ ↓ ↓ ↓ ↓ ↓ ↓ ↓ ↓ ↓ ↓ ↓ ↓ ↓ ↓ ↓ ↓ ↓ ↓ ↓ ↓ ↓ ↓ ↓ ↓ ↓ ↓ ↓ ↓
① Connecting devices directly to the cloud gateway: This means a secure connection over the internet.
② Connecting via a field gateway (IoT Edge Device): Use to connect devices that use industry-specific standards, short-range communication technologies (Bluetooth, ZigBee), resource-constrained devices that cannot host the TLS/SSL stack, or devices that are not exposed to the internet. This option is useful when running streams and data aggregations on a field gateway.
③Connect through a custom cloud gateway: Use when protocol conversion or custom processing is required before reaching the cloud gateway endpoint.
④Connecting through a field gateway and a custom cloud gateway: When some protocol enforcement or customization is required on the cloud side, you can choose to connect to a custom gateway running in the cloud. Some scenarios require integration with field and cloud gateways that use VPN technology or isolated network tunnels using application-level relay services.
◈ ◈ ◈ ◈ ◈ ◈ ◈ ◈ ◈ Additional IoT security articles are available. ◈ ◈ ◈ ◈ ◈ ◈ ◈ ◈ ◈ ◈ ◈ ◈ ◈ ◈ ◈ ◈ ◈ ◈ ◈
'🐳Azure' 카테고리의 다른 글
Getting started with Azure IoT Hub(2) (3) | 2025.06.29 |
---|---|
Getting started with Azure IoT Hub (1) (6) | 2025.06.24 |
Azure AD Cloud Provisioning (0) | 2025.06.17 |
Controlling resource access with Azure AD Identity Governance (1) | 2025.05.19 |
The uncomfortable truth about data file size and IOPS in Azure SQL (4) | 2025.05.18 |