IoT Roundup

IoT Roundup: VMware buys, ARM pushes for the edge and IoT in the walls

Virtualization pioneer VMware continues its streak of IoT-related acquisitions, ARM rolls out silicon for the edge, and your walls may be lined with IoT sensors.

IoT > Internet of Things > network of connected devices
Avgust01 / Getty Images

Like all new technologies that cause waves in the world of business IT, the advent of IoT technology provoked a scramble for relevant talent and expertise. Different companies have responded with different approaches, including the acquisition of promising startups in the field.

That appears to be VMware’s strategy, at any rate, as the cloud and virtualization giant last month announced plans to acquire Nyansa, a startup known for its Voyance AI-based networking automation product.

The idea, per VMware’s announcement, is to integrate Voyance’s capabilities into the company’s SD-WAN products, with the aim of making it easier to predict, optimize and manage networks connected to crucial IoT devices. Financial terms of the deal haven’t yet been disclosed

AI-powered network management and configuration capabilities are a good get for VMware. The networking complexity of IoT is one of the central implementation headaches of the technology -- just picture provisioning thousands of devices by hand – and a system that can handle that task not only in an automated manner, but create its own benchmarks and monitor devices for anomalous behavior is powerful tool for IT pros tasked with managing IoT systems.

VMware’s far from the only major player out there building its IoT portfolio through acquisition: Cisco bought up France-based device visibility and security company Sentryo last year in an effort to broaden its appeal for use with industrial control frameworks, and Rockwell Automation acquired Israeli company Avnet last month for similar reasons.

ARM at the edge

Edge computing is regarded as a component technology for IoT because it removes a major stumbling block from latency-sensitive IoT deployments. Instead of data having to flow all the way from a sensor to a data center or cloud instance and then all the way back, an edge module deployed close to the device in question can do basic processing without the need for a lengthy round trip.

ARM’s integration of what it describes as an industry-first neural processing unit into its latest Cortex-M chips, then, could represent a major step forward for businesses that need AI and machine learning capabilities to work with relatively constrained devices. Announced earlier this week, the ARM Ethos-U55 NPU is part of the Cortex-M55 processor line and offers the possibility of AI/ML on a huge number of new platforms.

“With these additions to our AI platform,” said automotive and IoT senior vice president Dipti Vachani, “no device is left behind as on-device ML on the tiniest devices will be the new normal.”

Low latency is important for certain types of physical IoT systems – industrial robotics and machinery may have to be precisely harmonized in order to reap the benefits of IoT tech, for example. Putting even limited AI capabilities right at the edge opens up a wide range of possibilities for smart systems that can partially manage themselves.

IoT in the walls

And you thought solar panels on the neighbor’s roof were a neat trick.  Network World’s own Patrick Nelson wrote this week that a novel RF array design created by MIT researchers could make the very walls of a building into an antenna for IoT or other purposes. RFocus, as the technology is called, works by beamforming, using antenna elements in hanging sheets or even in wallpaper to help guide signals where they need to go.

This could help address two major concerns in IoT architecture: large numbers of devices interfering with each other’s wireless signals and those signals being unable to penetrate thick walls, such as those that might be found in an industrial facility. It’s an innovative, timely idea, according to Nelson.

“Antenna engineering is turning into a vital part of IoT development,” he wrote. “It's one of the principal reasons data throughput and reliability keeps improving in wireless networks.”