Define IoT service architecture
Adatum's management team wants to use IoT capabilities to extend the scope of services it delivers to its customers, including automated monitoring, alerting, and the self-servicing of smart appliances. Your development team plans to provide the desired functionality by implementing a robust and feature-rich IoT pipeline in a manner that adheres to the foundational design principles of cloud-native applications. To accomplish this objective, you decide to explore the concepts of IoT services architecture.
What is IoT services architecture?
The term Internet of Things (IoT) represents a network of interconnected devices involved in generating, processing, or storing sensor-generated telemetry. At a high level, an IoT solution consists of two primary groups of components:
- Device-side. This group includes devices that serve primarily as sources of telemetry but might also perform initial telemetry processing and analytics.
- Cloud-side. This group includes cloud-based services that are optimized for data collection, persistence, and analytics.
In your intended design, smart fridges represent the device-side, while Azure IoT Central and Azure Cosmos DB accommodate data collection, persistence, and analytics.
Another approach to the IoT services architecture divides it into Things, Insights, and Actions. Things represent IoT devices generating telemetry that, through Insights-based analysis, is turned into Actions, which deliver meaningful business benefits. As an example, consider a sensor on a smart refrigerator (the thing), which sends temperature data. That data serves as the basis of evaluation whether the refrigerator is fully operational (the insight). The insight, in turn, can trigger an alert that leads to an updated maintenance schedule (the action).
A more detailed consideration of Azure IoT architecture allows you to distinguish between its core and optional components. The core components include:
- IoT devices that can securely register with the cloud and support the sending and receiving of data.
- A cloud gateway service that serves the role of a hub, relays data from devices, and provides device management capabilities.
- Stream processors that consume the device data, integrate with data processing components, and deliver data into storage.
- A range of data stores that provide data persistence and aggregation. Storage can be part of the hot or warm path, or the cold path. The hot or warm path consists of the data required for near-real time reporting and visualization. The cold path consists of data stored for a longer term and used for batch processing.
- Analytics services that retrieve insights from raw telemetry data and integrate with business processes and systems, such as Customer Relationship Management (CRM).
- Programming and management interfaces that simplify the administration of IoT solutions.
Optional components include:
- IoT Edge devices that serve as brokers for management and data flow for IoT devices, whenever direct communication with the cloud is restricted or not possible. They might, for example, play a role in device provisioning, data filtering, aggregation, buffering, protocol translation, or event rules processing.
- Data transformation services that perform ingress tasks such as protocol translation, data conversion, and data aggregation.
- A user management subsystem that implements role-based access control (RBAC) and allows for the delegation of management tasks.
- Machine learning that analyzes telemetry data to detect trends and correlations and provides functionality such as predictive analytics.
- Bulk device provisioning services that accommodate scenarios involving large-scale deployments.
IoT services data pipeline
The flow of device-generated data typically consists of several stages, including:
- Storage. This stage includes preserving data for a short term or a longer term, relying on technologies such as in-memory caches, temporary queues, databases, and data lakes.
- Routing. This stage involves delivering data to one or more storage endpoints, analysis processes, and actions.
- Analysis. This stage consists of evaluating and processing data records based on customizable criteria.
- Action. This stage involves responding to customizable rules to address a condition indicated by the state or value of collected data.
Azure IoT services and technologies
Microsoft offers a comprehensive portfolio of services that deliver various types of IoT functionality, including:
- Azure IoT Central. This service implements a wide range of IoT capabilities, including telemetry collection, processing, analytics, and secure device management. Azure IoT Central includes the following functionality.
- Helps you minimize custom development efforts and administrative overhead.
- Lets you use cloud agility and scalability.
- Supports built-in integration with several other Azure services.
- Simplifies the development and implementation of custom IoT solutions and supports ad-hoc testing with built-in mock devices.
- Integrates with Time Series Insights and can define rules for generating alerts. This integration might help you eliminate the need for dedicated stream processing services, such as Stream Analytics.
- Azure IoT Hub. This service is optimized for reliable and secure bidirectional communications between IoT devices and cloud services. Just as Azure IoT Central, it incorporates Device Provisioning Service (DPS) as a helper service for in-bulk, zero-touch, just-in-time provisioning of IoT devices.
- Azure Time Series Insights. This highly performing analytics, storage, and visualization service for time series data provides capabilities such as filtering and aggregation. Its visualization capabilities include support for overlays of different time series, dashboard comparisons, accessible tabular views, and heat maps. Azure Time Series also offers a JavaScript controls library that facilitates embedding time series charts into custom applications. The service is optimized for scenarios that require aggregation over large sets of data.
Azure also provides a large array of options for data analytics, event handling, and storage, that all integrate easily with IoT components.