Technology Internet of Things (IoT) Edge: Computing for Devices

Technology

Internet of Things (IoT) Edge: Computing for Devices

January 9, 2023

A popular trope in action movies involves characters throwing an entire city’s traffic into chaos by flipping a few switches at some central control room. Thanks to Internet of Things (IoT) edge architecture, which includes edge computing and IoT devices, this is no longer that simple (if it ever was). IoT edge: 1; movie villains: 0.

IoT edge refers to devices at the periphery of an IoT network that process data close to the point of origin, rather than at a centralized location. This type of IoT framework is known as “IoT edge computing”, and the portion of a network that consists of edge devices is called the “edge layer”.

Dgtl Infra provides a comprehensive overview of the Internet of Things (IoT) edge, including how the architecture works, as well as use cases, examples, benefits, and platforms of IoT edge computing and edge devices.

What is IoT Edge?

Internet of Things (IoT) edge is a set of technologies and services that enable a level of real-time data processing and decision-making that would otherwise be impossible or extremely difficult to achieve. It also enables a host of benefits, ranging from improving the reliability and efficiency of IoT systems, to enhancing security, and reducing server bandwidth and computational requirements.

IoT edge comprises various hardware and software systems within the overall framework of edge computing. This includes edge devices, edge gateways, cloud platforms, and more.

What is IoT Edge Computing?

IoT edge computing is the practice of processing data on IoT networks as close to its source as possible. This is achieved through the use of edge devices and edge gateways, which filter and process data locally, instead of relying exclusively on a central server or cloud server for data processing.

Edge computing is particularly valuable in IoT use cases where the speed of data processing, or reducing network usage, are critically important. For example, in autonomous vehicles, edge computing facilitates real-time decision-making, which is essential for the safe and smooth functioning of vehicles.

Before detailing the architecture of Internet of Things (IoT) edge computing, the following are further real-world use cases and examples.

IoT Edge Computing Use Cases

There are myriad applications for edge computing in the Internet of Things (IoT), especially in scenarios where low-latency is critical. Some of the most notable use cases of edge computing in IoT include:

Industrial Control: IoT-enabled edge computing helps process data from sensors in real-time in industrial control systems, allowing for the monitoring and control of industrial equipment, such as high-speed conveyor belts, or critical motors and pumps. This use case forms part of the Industrial Internet of Things (IIoT)
Military and Defense Systems: from drones and tanks to missile defense systems, IoT edge computing in modern military equipment has radically transformed operational awareness and analytics, making operations far safer and more effective
Energy: operators of wind farms and oil & gas facilities are achieving unprecedented levels of energy efficiency with the localized data filtering and processing enabled by IoT edge computing
Smart Cities: cities are increasingly relying on IoT edge computing with thousands of IoT sensors collecting data on traffic flow, air quality, energy meters, smart lighting, and parking systems. Processing all of this data at the edge allows for dynamic adjustments to be made to various systems

IoT Edge Computing Examples

Edge computing in combination with the Internet of Things (IoT) is already being implemented by a number of leading companies. Below are a few examples:

Manufacturing: Vantage Power, a UK-based automotive supplier, is using Amazon Web Services’ IoT edge computing solution to gain more data-backed insights from its products. The data not only helps with predictive and preventive maintenance of its existing products, but also enables improvements in their future products, shortening their development cycle by as much as six months
Oil & Gas: Schneider Electric used Microsoft’s Azure IoT Edge service to create tools for its own customers to analyze their data from pumps in oil fields, detect deviations from optimal conditions, predict impending failures, and change the pump’s parameters in real-time to mitigate the impact and reduce downtime
Agriculture: U.S.-based winemaker, Deep Sky Vineyard, used the Google Distributed Cloud Edge platform to monitor and automate irrigation and soil moisture tracking. This IoT edge computing solution reduced their energy consumption by 15%, drove down human error by 75%, raised labor efficiency 30%, and boosted crop yields by 50%

IoT Edge Computing Market Size

Verizon has stated that, by 2025, multi-access edge computing (MEC) services will unlock over $30 billion in total addressable revenue across private networks, edge computing, IoT, and enterprise solutions.

What is IoT Edge Architecture?

IoT edge computing architecture refers to the overall design and structure of an Internet of Things (IoT) system that utilizes decentralized computing. By contrast, IoT systems based on a cloud architecture rely on centralized or cloud-based servers to process data.

Without an edge architecture, data needs to be transmitted from distributed IoT devices to a central location for processing, and the results are then used to transmit commands back to the IoT devices to adjust their operating parameters. This adds considerable time and network latency to the process, and becomes exponentially more challenging in bandwidth-intensive contexts, such as in solutions that involve the transmission of high-definition video, where transfers often reach terabytes or petabytes of data.

Notably, edge and cloud architectures are not mutually exclusive. Presently, most edge architecture deployments include some use of cloud servers or distributed cloud computing. The key factor that determines if a system is based on an edge architecture, is whether data is processed at or close to its origin.

Dave Mosley, Chief Executive Officer of Seagate Technology, a provider of data storage technology and infrastructure solutions, recently described the relationship between edge and the cloud, in the context of IoT, as follows:

“The cloud has centralized that data by putting forward a good value proposition. I think there’s a coming push to really decentralize some of the data again. It’s not necessarily at the expense of the cloud, but it’s just because the size of the data set is growing so much. When you think about it, mobile cloud was designed as a kind of a push model to push data out to the edge. And what we’re going to see in IoT and with a bunch of AI and ML, there’s a lot of data being created on the edge.”

IoT Edge Architecture Layers

Building on the aforementioned example of a network of autonomous vehicles, the image below illustrates the simplified architecture of an IoT edge system.

While advanced flow control, traffic management, and route optimization for an entire network of autonomous vehicles does require a certain degree of centralized data processing, it would be impractical to transmit and centrally process the thousands of data points collected each second by the array of sensors in every single vehicle. This is where edge devices and edge gateways become relevant, as they are the elements that make it possible for data to be processed locally, improving speed and efficiency.

What are IoT Edge Devices?

IoT edge devices are physical devices such as sensors, cameras, and other types of hardware, that gather data about a particular environment or system.

IoT edge gateways are intermediary devices between edge devices and the rest of an IoT network. Edge gateways process and filter data from edge devices, as well as transmit the data to a central location or cloud platform.

In certain cases, a network may not have a separate gateway layer, and some degree of data filtering or processing may occur directly on an edge device.

Differences Between Edge Devices and Edge Gateways

Edge Devices	Edge Gateways
Have sensors but may also have some data processing abilities	Exclusively for data processing / transmission; lack sensors for data capture
Usually physically smaller and more application-specific	Larger and more general-purpose

Examples of IoT Edge Devices

Any IoT sensor qualifies as an edge device if it has sufficient storage and computing power to filter or process data locally in a matter of a few milliseconds. For instance, advanced surveillance cameras have the ability to run some degree of video analytics locally on the camera, such as people counting (e.g., footfall counter), crowd density estimation, and vehicle identification based on size, speed, and direction.

Examples of IoT Edge Gateways

The Cisco Catalyst IR1100 router and the Siemens SIMATIC IOT2050 device are examples of IoT edge gateways designed for industrial applications. These IoT edge gateways allow for the:

Gathering and filtering of data from machines and sensors
Ability to run custom programs to analyze data locally

In turn, these IoT edge gateways unlock many benefits, like modulating electrical transformers based on sensor information, to self-healing power grids.

How Can an IoT Edge Device Be Used as a Gateway?

Small, single-board computers (SBCs) and mini-computers, such as the Raspberry Pi, BeagleBone Black, and Intel NUC Mini PCs, are examples of Internet of Things (IoT) edge devices that can also be used as edge gateways.

For instance, the Raspberry Pi series can be paired with a host of sensors, and is often used for automation, such as controlling lights, heating, and motorized gates. However, since it is also programmable, capable of receiving data from other devices, and has the storage and compute power to process this data, it can also easily double as an edge gateway.

How Can IoT Benefit From Edge Computing?

The Internet of Things (IoT) can benefit from edge computing through faster processing, improved network efficiency, increased reliability, cost efficiency, as well as privacy, scalability, and modularity.

Faster Processing

Edge computing enables real-time processing of data collected by edge devices, which is particularly useful in applications where quick response times are important. IoT edge analytics is an example of edge computing enabling faster, timely decisions, thanks to the ability to locally run tasks such as data filtering, aggregation, and transformation. Commonly, latency can be reduced to less than 10 milliseconds with an edge architecture, as compared to latency of hundreds of milliseconds without an edge architecture.

Improved Network Efficiency

Data being processed at or near its source reduces the amount of data that needs to be transmitted to central locations and helps to decrease strain on servers and other resources, resulting in an increase in network efficiency.

Increased Reliability

By processing data at the edge, the risk of data loss or delays – due to network issues – is reduced, increasing the reliability of the IoT system. In the event of a connection loss to a central location, the ability to continue operation is maintained through edge computing.

Cost Efficiency

By transferring some processing tasks to edge devices, the reliance on costly central servers and other resources can be reduced, lowering the overall cost of an IoT system.

Privacy, Scalability and Modularity

The ability to make localized decisions, enhanced privacy, greater scalability, and increased flexibility are further benefits for IoT from edge computing. These benefits can be particularly useful in contexts where:

Specific local needs and conditions must be taken into account
Volume of data or number of IoT devices is expected to grow
Data needs or requirements of the IoT system are likely to change

IoT Edge Platforms

Before the advent of the Internet of Things (IoT), traditional cloud architecture was generally structured to keep most of the computing power centralized, at large ‘hyperscale’ data centers. With the heightened adoption of IoT and emerging use cases that consume ever-increasing amounts of network and computational resources, edge computing has increasingly become a necessity, spawning a host of IoT edge platforms.

The most notable providers of edge computing platforms are Amazon Web Services (AWS), Microsoft Azure, and Google Cloud, through a variety of dedicated hardware and software solutions.

AWS IoT Edge

Amazon Web Services (AWS) offers a range of services for edge computing, including AWS Snowball, a service that employs AWS’ own proprietary, programmable edge devices to transfer data to-and-from the cloud, and AWS IoT Greengrass, which allows cloud-managed edge devices to act locally on data.

Azure IoT Edge

Microsoft Azure’s suite of edge computing tools include Azure IoT Edge, which allows for the extension of cloud intelligence and analytics to edge devices; Azure Stack Edge, which brings Azure compute, storage, and intelligence to the edge through Azure-managed devices; and Azure FXT Edge Filer, which supports hybrid storage optimization. Multiple other Azure services also support edge computing, adding capabilities for real-time insights, blockchain storage, and secure device connectivity.

Google Cloud IoT Edge

Google Cloud’s edge-focused offerings include Edge TPU, and Distributed Cloud Edge Appliance, which are both dedicated hardware appliances designed to run custom data processing applications at the edge, as well as its Distributed Cloud Edge service, which allows cloud-based management of edge devices.