How edge computing is accelerating innovation in hardware, software and service provider domains

More and more companies are placing more emphasis on edge computing. According to a report by AT&T Cybersecurity, 75% of security leaders are planning, deploying, or fully deploying an edge use case. This is largely attributed to the technology’s ability to save bandwidth, speed up response times and enable data processing with fewer constraints. In fact, the State of the Edge study of the Linux Foundation predicts that by 2028 companies will use edge computing more widely.

During VentureBeat’s Transform 2022 Virtual ConferenceDavid Shacochis, vice president of enterprise product strategy at Lumenmoderated a panel discussion to talk about how edge computing is transforming use cases and strategies for some of the industry’s true giants, in hardware, software and service provider domains.

The discussion also centered on Shacochis’ colleague Chip Swisher, who leads the internet of things (IoT) practice for Lumen; Rick Lievano, CTO for the Global Telecommunications Industry at Microsoft; and Dan O’Brien, general manager for HTC Vive.

Technology evolutionary cycles

Shacochis said computing power has gone through evolutionary cycles that oscillate back and forth between centralized and distributed models. Shacochis looked at periods of technological achievement and said steam power enabled mass production industries, while electrical distribution fueled the modern industrial economy that sparked the dawn of computing power in microprocessing. This, he said, has now led to the present day and what is called the Fourth Industrial Revolution.

He further noted that the dawn of mainframe is the original dawn of compute, which is distributed to client-server models, then consolidates back to the cloud, bringing all business logic into more centralized postures.

“Now we see this explosion of all the different data sources, the different ways of processing that data, the different types of sensor-actuator cycles that can really add a lot of value to customer experiences and industrial efficiencies,” Shacochis said. “All these different types of business results from the many different ways to reap the benefit. So those industrial cycles that happen over decades, the computer cycles that happen over even smaller periods of years, have really led us into this exciting time in the industry.”

The fourth industrial revolution

O’Brien looked at the era of the fourth industrial revolution from a hardware perspective and said HTC started out as an original design manufacturer (ODM) company. He said HTC made motherboards and chipsets for other companies and other products and PCs, using immersive silicon. He added that the company moved very quickly towards application-specific integrated circuit (ASIC) chips and GPUs that evolved into smartphone technology.

O’Brien noted that “a lot of people don’t realize this was the beginning of what we’re seeing today in augmented reality [XR] world, building these new kinds of immersive products. It actually evolved from so many chipsets and evolved so much from the smartphones. What’s in Modern Virtual Reality [VR] headsets and displays is a smartphone panel that was driven by the need to have higher visual quality and reliability in a smartphone.”

“Now we see where we need even more processing power,” he continued, “We need even more visual quality and performance in VR headsets for an XR headset and an augmented reality [AR] kind of solution. We see this increase in terms of demand and overall performance needs. The additional products require large PCs and GPUs to make all of this work. Now we are moving all of this to a cloud environment.”

He added that there are now also artificial intelligence (AI) and machine learning (ML) that will optimize processes for all virtual content and interactions.

In addition, Lievano said that the cloud has really changed everything and that the edge is an extension of the cloud. He noted that at Microsoft they talk quite a bit about this notion of the intelligent cloud and intelligent edge, which he believes is a way to deliver applications across the computer canvas where they are needed.

“As a developer, you like to think that you can build an app once,” says Lievano. “You want to use the latest technologies, which are currently cloud-native principles, and you want to be able to deploy it anywhere, whether it’s a public cloud in the air or an edge location. So this vision that we have for intelligent cloud and intelligent edge is largely dependent on our telco partners, because ultimately they provide that connectivity – the connective tissue needed to make this vision a reality. But the cloud must connect to the edge. And without telcos like Lumen, there’s no intelligent advantage.”

According to Lievano, this is different from the switch from mainframe to client server, where each mainframe and cloud server had its own uniquely developed models with its own governance. The cloud-native capabilities are the same whether they’re available in Azure or in the cloud, he said.

He also noted that you may have a subset of those cloud capabilities at the edge because of scale, but the programming model, devops model, management portals, management interfaces, and APIs are all the same. He also said the ad will be another cloud region for a developer to deploy their applications in, which is a huge difference between a mainframe and a client server.

“Again, as a developer, I’m amazed at the progress and tooling, especially in recent years,” said Lievano. “AI, for example, has had an incredible impact not only on the applications we create as developers, but also on the applications we write and how we develop those applications. The cloud therefore offers you unlimited computing options [that are] really within reach. Again, scale isn’t an issue, but features like serverless computing, for example, allow you to take your applications to the next level. In science, you can create and deploy complex applications using microservices.”

Evolution of IoT

From a solutions and service provider perspective, Shacochis said the cloud and some of its tools make some things easier, but customer opportunities and expectations make things more complex. However, Swisher, speaking from his specialty around IoT, said that while some say IoT is a new concept, in reality it has been around for more than 20 years. It’s a concept that explains the ability to grab data from machines and devices and perform certain operations with it, Swisher said.

“I’ve been through the wave of what I call IoT 2.0 where you might be held on a factory floor, a localized production line control machine doing the processing locally there,” Swisher noted. “Then we saw the rise of the move to the cloud, and several stovepipe cloud providers offering centralized end-to-end solutions in that space. Now we’re really seeing the need for integration on the IoT 2.0 where we’re starting to see cross-stovepipe use cases where data is coming from multiple different IoT infrastructures and IoT paradigms and being able to bring all that data together in a single view.”

Swisher added that machine learning is the next evolution of complete visibility into everything that happens in the city, factory, warehouse and distribution to bring data together.

He noted that IoT 2.0 “creates new challenges both from a compute point of view and from a network integration and services point of view, where it’s necessary to calculate even closer to those aspects, because we’re building all those things together, we really have the needing ability to make that happen even more in real time to be able to adjust if we need it The concept of using compute in a location, you know, compute in a metro edge or a near edge capability , as well as the cloud, and being able to have all those parts to put all those pieces together and being able to move, calculate around those different locations has really become critical.”

Miss the . not full discussion of how edge computing is transforming use cases and strategies for some of the industry’s true giants, in hardware, software and service provider domains.