Bench Talk

Diverse workloads require a smart edge to manage and secure a wide variety of IoT devices

(Source: photon_photo- stock.adobe.com)

The edge is getting smarter. After years of consolidating data in centralized cloud computing instances, processing is being pushed back out to the edge as the maturing internet of things (IoT) becomes more intelligent.

The early days of the IoT laid the groundwork for edge computing that went beyond distributed data centers and devices that do more than just collect data with sensors and relay it to the cloud. Advancements in IoT technology have increased capabilities and performance at the edge; edge devices can now do so much more than just send infrequent, small data packets while rarely receiving.

According to Grand View Research, the global edge data center market size¹ was $9.36 billion in 2022 and has an expected 18.4 percent compound annual growth rate (CAGR) from 2023 to 2030. The demand is being driven by broad industry adoption of IoT, AI, big data, cloud, streaming services, and 5G, which are all driving up network data traffic.

The edge is now comprised of a broad mix of IoT devices, distributed data centers, and wireless base stations. Communication standards continue to be updated to enable more essential computing and compression tasks to be conducted at the edge, including artificial intelligence (AI) inference and machine learning (ML).

The evolution of IoT beyond a mere collection of wireless sensor nodes pouring out data streams hasn’t made the deployment or operation of IoT applications any easier. The sheer number of devices and rapidly evolving capabilities create significant logistical challenges.

However, today’s smarter edge devices and distributed data centers are better equipped to pull their weight out in the field. These edge centers act as a local host for the initial commissioning of large numbers of networked IoT devices and manage subsequent over-the-air updates. Edge devices are now also able to provide local versions of cloud-based services to maintain operations when connectivity to the cloud is interrupted or lagging, as well as continue to perform local processing essential to support short-latency feedback loops.

The proliferation of edge devices has broadened the attack surface available to threat actors looking for avenues to access the central cloud data centers. At the same time, the smarts of today’s edge computing systems can be used to actually bolster IoT security.

Intermediate edge devices tasked with deployment, maintenance, and ongoing operations can handle the burden of sophisticated security for IoT devices that are designed to be efficient and low-cost. Edge devices also allow developers to address privacy requirements more effectively by enabling concepts like privacy-by-default and privacy techniques like data minimization.

As data scientists encounter more constraints on pushing detailed data to the cloud, IoT solutions must pull data-intensive algorithms—such as anonymized machine learning and advanced pattern matching—into the edge.

Early IoT devices were essentially sensors with connectivity with little onboard processing, memory, or storage.

These simple devices are not going anywhere, but the emergence of the smart edge means there’s a lot more communications and computing power in compact form factors, replete with application processors and microcontrollers. These devices offer greater performance and specialized processing capabilities necessary to run sophisticated algorithms that can improve security and privacy while running more complex analytics and AI workloads.

When acting as an intermediary to cloud IoT infrastructure, more powerful computing resources like microcontrollers or central processing units (CPUs) are required, while image- and video-based machine learning systems rely on graphics processing units (GPUs).

Even when more is done by the edge device, it still requires secure connectivity to a central cloud data center as well as the many IoT devices it manages. The latest and greatest wireless standards, including 5G, Wi-Fi 6e, and eventually 6G, are all part of the communications backbone of the smarter edge, with 5G and Wi-Fi^® now the de facto interconnectivity high-performance IoT deployments.

Today’s smart factories leveraging automation are using AI and ML to analyze data, improve processes, and proactively predict equipment failure with end-to-end computer vision. An IoT edge device running an ML model can calculate, infer, and send actionable output based on collected images to a central cloud data center for further processing. Learning gathered in one location can be shared with other factories.

Warehouse systems are not only using more advanced IoT for inventory tracking purposes, but edge computing plays a critical role in managing the distributed sensors that support fully autonomous mobile robotic (AMR) systems aiding in order fulfillment.

Healthcare is also benefiting from more advanced edge computing—not only has IoT enabled remote patient monitoring using devices like blood pressure cuffs and glucose monitors, but also real-time monitoring and management of connected medical equipment distributed throughout hospitals.

IoT deployments are growing every year and the edge must get smarter to safely support the diversity and functionality in a broad range of use cases across many industries. Device capability and performance must continue to advance and evolve as IoT deployments leverage faster processors, AI, and ML to do more at the edge independently from cloud data centers without compromising security or privacy.

• 1. Edge Data Center Market Size, Share & Trends Analysis Report By Component (Solution, Service), By Facility Size (Small & Medium, Large), By End-use (BFSI, Government), By Region, And Segment Forecasts, 2023 - 2030. Accessed August 11, 2023. https://www.grandviewresearch.com/industry-analysis/edge-data-center-market-report.