Why the metaverse will depend on advances in edge computing

Web 2.0 brought us user-generated content and interactivity – think Twitter, Facebook, Slack and Zoom. He enabled online startup Dollar Shave Club to gain global recognition in 2012 with a YouTube video that cost just $4,500 (£3,300) to produce. The company shook up its previously bombproof market incumbents to the point that Unilever reportedly paid $1 billion to acquire it four years later.

Web 3.0 has brought us even more disruption, in the form of technologies like big data, machine learning, and blockchain. Businesses are understandably interested, therefore, in getting ahead of Web 4.0. Definitions vary, but this iteration promises highly personalized and immersive online services, blurring the physical and the digital.

How will that look? the metaverse it is both the greatest promise and the greatest threat of Web 4.0. It’s a threat to incumbents like Facebook (now Meta), largely due to its decentralized nature. It holds promise for startups because they can harness technology in ways that give them a competitive advantage.

All sectors are eager to get to the forefront. Microsoft recently spent $68.7 billion on gaming company Activision Blizzard. JPMorgan, for its part, has created his own metaverse room – in which visitors are greeted, of all things, by a virtual tiger.

A less headline-grabbing issue is what CIOs need to do to make such technology work well.

Companies offering augmented reality (AR) and virtual reality (VR) in the metaverse will almost certainly have to reconfigure their computing power to make experiences feel properly immersive. Centralized provision of cloud computing, as impressive as it is, will not be enough. The reason for this is latency: the delay time on the network.

Swedish telecommunications company Ericsson has pointed out that time-critical video games, such as first-person shooters, do not need end-to-end network latency greater than 30 milliseconds to ensure a high-quality experience. . The further the data center is from the end device, the higher the latency. Even on the fastest fiber links, there is a latency of 5 microseconds (0.005 milliseconds) for every 1 km of cable that data travels, according to Infinera research.

That’s why serious gamers use expensive hardware that can do the processing right then and there. The problem for businesses is that consumers are unlikely to want to spend a lot on special hardware to access metaverse services. To do?

Gain a competitive advantage

edge computingwhere the processing muscle is placed closer to the data being processed, it’s the next step up from the cloud.

“People have been talking about the edge for two decades, but it’s been limited to niche use cases,” says Ishu Verma, emerging technology evangelist at Red Hat, a provider of open source enterprise software. “Now, the idea of ​​placing compute and storage closer to data sources is being more widely adopted in consumer and industrial applications.”

A major reason for this is that data systems have become much more capable, cost-effective, and energy efficient, making deploying them at the edge on a large scale much more feasible than it once was.

“In the cloud, you scale capacity. On the edge, you scale it outside to millions of sites,” says Verma, adding that there is demand from all sectors that need low-latency services or simply want to avoid batch processing.

Some industries are already way ahead of the rest. Manufacturing companies, for example, use edge computing to build digital twins: intelligent virtual replicas of physical infrastructure.

Rolls-Royce, for example, can offer customers a virtual aircraft engine in flight that responds as the physical machine does. Sensors on the engine send data back via a satellite link, although most of the information is collected after the plane has landed. Machine learning models inform the digital twin, and by extension the end goal, which is better design and maintenance of the physical engine.

Rolls-Royce chief information and digital officer Stuart Hughes says this simulation facility is important for the company, because some of the things it does are “really on the verge of physics”. Having a virtual replica of an engine allows testers to put it into many more scenarios than they could physically.

The company’s long-term goal is to have an engine that is “increasingly connected, context-aware and understanding.”

As it is, most companies have their core computing capacity in the cloud, although there are day-to-day instances of edge computing. Each ATM is a small data center, for example. And the user interface for Amazon’s voice-activated virtual assistant, Alexa, doesn’t rely on sending all data to the cloud. Some of it is analyzed in the machine.

In the cloud, capacity increases. At the edge, it scales to millions of sites

The offerings provided by Amazon Web Services (AWS) are a continuum, right down to Internet of Things sensors on user premises. So says AWS Product Development Director George Elissaios, who adds: “Edge computing is cloud computing.”

Companies that are not cloud hyperscalers tend to make a clear distinction between the cloud and the edge. The general rule of thumb is that the cloud offers economies of scale, more processing control, and greater computing power.

“Training and machine learning model development happens in the cloud, while inference with real-time data happens at the edge,” explains Verma.

Enterprises also have to consider more than just latency when looking at the edge. Data sovereignty, for example, is a thorny issue. As Facebook is finding out in the EU, some jurisdictions don’t like data being transferred to servers across their borders.

Security is another key consideration. “The approach to take is ‘trust no one,'” advises Verma. “Any data from a remote device is suspicious.”

This is in part because remote devices can be easily manipulated, unlike those in the cloud or on premises.

Will Edge Usage Displace Cloud Usage? The consensus among experts is that it won’t.

“I don’t think the edge will replace the cloud. i don’t think the edge compete with the cloud,” says Matt George, director of segment marketing and business transformation for Equinix in EMEA. “As you go down the road [towards more real-time services]what you want is the most agile and flexible IT setup you can have.”


Leave a Reply

Your email address will not be published. Required fields are marked *