IoT Cores in Public Clouds - A Review

Cloud IoT core services are a new class of cloud services, a network that connects and manages IoT devices. Although many people assume edge computing and IoT cores are synonymous terms, they’re actually different things.

Edge computing refers to the specific architecture utilised by IoT cores and workflow, whereas IoT is used within the edge computing architecture. To put it simply, edge computing pulls data storage and computation closer together, saving time and bandwidth. This kind of tool is used for developing device data exchange and flow from connected IoT sensors and embedded devices with analytics tools.

Many specific cases like car fleets, robots fleets, and distributed sensor networks need one system as a service for storing telemetric device data, management, upgrades, monitoring, and analytics.

With cloud IoT core services using edge computing, it’s possible to develop the Internet of Things solutions quickly and with the highest quality. Moreover, different telementrics data from specific business areas combined with cloud IoT core services allow scaling and deploying independent solutions in devices fleet.

What is IoT core?

Using edge computing and cloud services, IoT cores allow you to connect as many IoT devices as you need to the cloud to exchange data without the need to provision or manage servers. Cloud IoT core services can support billions of devices and trillions of IoT data messages.

Businesses can process and route the data exchanged using internal dedicated components in the public clouds and redirect data into data analytics and monitoring ready-to-use solutions.

All IoT data streams from devices like vehicles, meters, data loggers, production processes, or wearables with IoT cores do not need special services outside the public clouds for data analytics and machine learning.

Popular cloud services managing cloud IoT cores

There are a few popular cloud service platforms to manage cloud IoT core systems using cloud computing for general availability, for example:

• AWS IoT Core
• Azure IoT
• Google Cloud IoT Core / GCP IoT Core

All public cloud service providers, including the GCP (Google Cloud Platform) IoT core and AWS (Amazon Web Services) IoT core, offer free plans with high limits for developing prototypes. All building blocks from cloud computing platforms give more power than is needed to develop IoT devices fleets and dedicated integrations.

Using platforms designed for edge computing such as the Google Cloud Platform as well as other Google Cloud Services allows businesses to securely connect devices across a network using cloud IoT core services.

This allows for seamless IoT deployment. Many services also include cloud pub/sub. Cloud pub/sub being a method of communication for exchanging relevant business data generated between IoT devices.

If you’re not sure whether IoT cloud solutions are for your business, many platforms offer a free tier of services. This free tier usually allows for a limited number of devices to be hosted within the cloud, and upgrading the free tier usually removes this cap.

Features of fully managed IoT solutions and services

Below is the list of features of IoT solutions and services, including device shadow management, end-to-end security with threats detection, and device deployment at scale.

Device shadow management

The device manager allows individual IoT devices to be configured and managed securely, usually through a console. The device manager establishes the identity of a device and provides the mechanism for authenticating a device when connecting.

It also maintains a logical configuration of each IoT device and can be used to remotely control the device from the cloud. This allows you to effectively manage cloud IoT devices.

Protocol bridge for popular stacks

The protocol bridge provides connection endpoints for protocols with automatic load balancing for all device connections. The protocol bridge has native support for secure connection over industry-standard protocols such as MQTT and HTTP protocols.

The protocol bridge publishes all device telemetry to the cloud public-subscribe, which can then be consumed by downstream analytic systems.

End-to-end security with threats detection

Enable end-to-end security using asymmetric key authentication over TLS 1.2; CA-signed certificates can be used to verify device ownership. Connected devices supporting a cloud IoT Core security requirements can deliver full-stack security, reducing overall security risk.

A single global system with gateways for connecting our all devices fleets

Cloud services and edge computing allows you to connect as many IoT devices and gateways as you need to the cloud over standard protocols, such as MQTT and HTTP, through the protocol endpoints.

It also allows you to manage all your cloud IoT core devices as a single global system, helping them exchange data more efficiently. The service uses public-subscribe concepts, which retain data for seven days.

Out-of-box business data insights, analytics and dashboards

Use downstream analytic systems by integrating with Big data analytics and ML services such as Big Data orchestrators like Apache Spark, Apache Airflow, or partner Business Intelligence tools.

Fully managed service and scalable with high availability on the box

The fully managed service is serverless and doesn’t require any up-front software installation. It scales instantly without limits, using horizontal scaling of cloud platforms.

Role-level access control with IAM and system-specific roles

Apply IAM roles to device registries to control user access to devices and data.

Device deployment at scale with many deployment strategies

Use REST APIs to automatically manage cloud IoT devices, including the registration, deployment, and operation of the devices at scale. Also, use the APIs to retrieve and update device properties and states even when the devices are not connected.

High-frequency, low-latency communication - important for IoT critical systems

Send command or configuration directives to devices connected to cloud IoT core. Commands are fast, frequent, and one-time directives sent to devices. Configurations are persistent directives that, when using MQTT, are delivered to all subscribed devices, including ones added at a later date.

Offline operation and support for resource-constrained devices

Gateways give resource-constrained devices capabilities offline as well as online. This gateway can perform tasks on the device’s behalf. This includes communicating with cloud IoT cores as well as connecting to the internet and authenticating credentials.

Real-time metrics with monitoring

Use integrated monitoring to create dashboards that show data like the total number of active devices in a registry. You can also set up alerts based on metric thresholds, such as devices in a registry exceeding a preset billable data limit.

All your device logs in one place

See connection and error logs in logging alongside audit logs. Configure user-defined metrics to gain insights like the number of devices that published data to a specific Cloud public-subscribe topic.

Build and train ML models

All cloud IoT cores have integrated connectors where we can build and train ML models for i.e. dedicated predictions for our production system.

Additional services to IoT cores

Businesses can select additional cloud services dedicated for the cloud IoT core faster adoption process with hardware stacks. All cloud platforms have their own hardware markets with a list of certificated devices.

Of course, it's a very comfortable situation for developing phases or building prototypes, when you can buy devices supported by IoT core. Probably if you need something special and battery-powered, but for other things connected to the IoT core, it's a very good choice.

Lists with popular hardware markets in IoT cores from public clouds:

• GCP IoT Kit
• Amazon AWS Device Catalog for IoT Core
• Azure Certified Device catalog

You can select many criteria from IoT device and hardware catalogs, such as:

• Connectivity (Bluetooth, 5G, LAN, WAN, LTE, LoRaWAN, Narrowband IoT, 3G, WiFi)
• Device type (Developer Kits, Finisher Products)
• Geo availability (APAC, America, EMEA, Worldwide)
• Hardware interfaces (COM, GPIO, I2C ISP, USB)
• Industrial protocols (Can BUS, EtherCAT, Modbus, OPC, Profinet, ZigBee)
• Industries (healthcare, automotive, education, government, hospitality, retail, smart buildings)
• Integrated sensors (gas, temperature, humidity, LEDs, pressure, GPS, accelerometers, touch, vibration, vision, weight)
• Ingress protection (IP) value
• Operating systems (Android, Arduino, Free RTOS, Linux, mbed, RTOS)
• Operating temperatures (Commercial: 0 ° to 70 °C, Industrial: -40 ° to +85 °C)
• Processor manufacturers (Broadcom, Espressif, Intel, Raspberry Pi, STM, Mediatem, Microchip, Texas Instruments)
• Secure hardware (by internal cloud security components or dedicated hardware)

It’s very easy to select a cloud IoT device based on destination criteria and environment. If your project is based on open source you have few possibilities and hardware platforms to choose from. Many connectivities and industrial protocols are based on IEEE specifications, so you can build industrial-level devices with IoT core additional services.

How does it work from an architectural point of view?

With IoT cores, you can easily build analytics for your data. Based on the example below diagram there are few building blocks in today's IoT cores:

• Things with SDK from the cloud platform provider
• Message brokers or publish-subscribe solutions for receiving data streams and building dedicated topics for specific data messages
• Mitigating security risks and identity layer for devices fingerprints and access keys for operators
• Thing shadows - this is a concept from digital twins, where all devices states have their shadow copy in the cloud, so any failures on devices are accepted because our data state is frozen in the cloud and restarting device is enough to high availability
• Thing registry it’s a database for all connected IoT devices called Things
• Rules engine is the very important part when we need to send our data or partial data from measurements into analytical level or into databases or data warehouses
• IoT application - this is the application layer when we build business cases

Image source: Research Gate

Scenarios having applications of IoT

Here are some real-world examples of IoT application:

Smart parking systems

It's a sensor-based system, used in environments such as shopping centers, airports, commercial parking operations, universities, and municipal streets.

Products include ground sensors that use an advanced combination of infrared and magnetic technologies to reliably register a vehicle’s arrival and communicate to a gateway.

Smart parking systems are useful as they can provide automatic guidance to inform customers of how many vacant parking spaces are available for each level of parking throughout an area of a city.

Beyond this, more advanced functionality would allow parking enforcement to identify whether a vehicle has overstayed in a limited time spot. These operators may notify authorities who can undertake enforcement actions and issue infringement notices.

Another smart parking mobile product enables parking operators to photograph license plates to identify vehicles that have overstayed. Applications are also being developed to enable users to view numbers of vacant parking spots in nearby parking facilities.

Benefits of smart parking systems

• Reduced smart parking, smart city and smart home IoT installation and operational support effort by more than half
• Enabled development of a Smart Cloud IoT platform in just four months
• Democratised data access and use across the organization

Smart factory cloud platform for manufacturing analytics

The giant network of “things” (including people) connected to each other via the internet has the potential to reduce waste, increase efficiency, and improve safety. Many things in manufacturing like combining wireless connectivity, big data, and cloud computing give us real benefits and low costs in the Industry 4.0 era.

The use of data and cloud computing to improve manufacturing is practically as old as manufacturing itself. However, the computerization of manufacturing has resulted in rapid and widespread changes to the way data is collected and processed, as well as the sheer volume of data available.

The main goal is to help manufacturers tap into this data to quickly identify process trends and even warning signs of machine breakdown. This kind of foresight can reveal opportunities to improve the manufacturing process as well as the maintenance processes that have the potential to reduce waste and increase profit margins.

Benefits of smart factory cloud platform

• Increased profitability and decreased waste with real-time data analysis.
• Decreased the number of VM instances to manage almost in half, freeing resources
• Reduced storage resources and analytics services costs by tens of percent

Fleet management systems

Fleet management systems can be defined as the process used by fleet managers to monitor activities and make various decisions from asset management to dispatching and vehicle acquisition. This helps companies ensure compliance, improve efficiency and reduce costs.

The process of fleet management is to ensure the smooth and continuous operation of a fleet of vehicles, no matter the size. This includes tracking the condition and location of the vehicles, maintaining schedules and fuel usage to manage costs and optimize the life of the equipment.

Fleet management tools allow fleet managers insight and real-time visibility into their various operations. At the same time, this increases driver satisfaction and decreases fuel usage via analytics and accurate reporting.

Benefits of fleet management systems

• Ease of use - often simple implementation based on IoT Cores
• Mobility
• Scalability due to the cloud
• Reporting
• Expense management and total cost of ownership
• Performance metrics

Why are IoT cores so valuable?

IoT cores open industry standards enabling integration with a wide range of 3rd party sensors to extend the range of data provided to our analytics platform. We recommend the managed cloud IoT cores platforms for building Industry 4.0 solutions for businesses.

All dedicated solutions need dedicated architectures and IoT cores are the best choice for realizing the Internet of Things ambitious projects. With IoT cores you can build and ship IoT apps faster with a high level of security for all the important data.