The infrastructure requirements of the next generation immersive Internet (aka the metaverse)

The digital infrastructure that powers today’s Internet is the largest infrastructure project in human history. However, this will not be enough to power the advanced digital applications of the currently evolving immersive Internet, what others call the “metaverse”. As I laid out in my last article, the immersive Internet will be much more latency sensitive than today’s digital applications, as it needs to match the speed of human perception in order to offer a truly immersive digital environment. That means data needs to be able to flow seamlessly, efficiently, and very fast – faster than the human brain – not only for a few high-performance applications but for everyone, everywhere.

Moving data to and from the edge

The immersive Internet will demand multiple real-time video streams alone just to create a real-time 360-degree virtual replica of an existing location – and then further data streams for other human senses. This means huge amounts of data need to be shoveled onto the end-user’s device, with extremely fast – in some critical use cases as short as half a millisecond – round-trip times to avoid delays in perception. But how can we achieve these high speeds and low latencies? Well, one of the biggest challenges here is pure physics: Even under the best circumstances and with modern fiber cables, data cannot travel faster than the speed of light. That means that interactions dependent on a latency of around half a millisecond will only be possible with people and objects within the near vicinity. For these use cases the data pathway cannot exceed 35-50 kms for data to travel between the user, the device, and the server for processing. The good news is: There might not be too many of these extremely latency-sensitive use cases, e.g. a surgeon performing remote life-saving surgery at the site of an accident, before the patient is transported in a high-tech ambulance to the nearest hospital. For other use cases for commerce, entertainment, and education, such as virtual business meetings, gaming, or virtual shopping, it is fair to assume that a latency of between 5 and 15 milliseconds will suffice.

But even for these use cases, our current digital infrastructure is not up to par, with the exception of a few major hubs. High-performance computing must get as close to the users and smart devices as possible, and do so everywhere. The immersive Internet will need data centers that house AI applications and digital twins and cache virtual environments much closer to the user. Therefore, it will continue to be necessary to build more data centers of all shapes and sizes in more and more regions around the globe: centralized and decentralized, large hyperscalers, colocation data centers, container edge solutions, and shoe boxes in the basements of buildings, each with their own relevance for a variety of use cases. To move the huge amounts of data needed for the immersive Internet to and from these datacenters and to and from the end-user, all available network technologies – from fixed line FTTH/FTTB and fiber backbones to LEO satellite networks, Wi-Fi 6, 5G, 4G, and potentially even 3G (at least to begin with, in some regions) – must work in concert. This is the only way the required network density can be achieved.

Interconnecting closer to the end user

But the data streams for visual, auditory, and tactile information are not the only ones that must flow to keep digital applications in the immersive Internet running. They are basically just the front-end for these new applications and use cases. To enable these business cases, a growing number of different companies need to exchange data to provide back-end services. This might include sourcing details of virtual twins of objects and people hosted in different clouds, harmonizing customer preferences across different platforms to personalize the virtual experience, or processing payments for purchases in the metaverse. This means it will never be just a single route from a data center to the end user that needs to comply with the latency requirements of the immersive Internet, but countless other connections between different companies working in the background to make the immersive experience possible. All these need to deliver high-quality and low-latency data transfer as well. Here, at the very latest, interconnection platforms must also be a part of the infrastructure mix. Not only data centers, but also interconnection platforms (or Internet Exchanges) must move closer to the edge. High-performance interconnection platforms will be needed, in the long run, every 50-80 kms, to keep local traffic local and offer suitable latencies for increasing performance requirements. For the bandwidth-intense immersive applications of the future, interconnection platforms will need to not only offer 400 GE connections, but also 800GE and beyond. Ideally, these platforms will be data center and carrier neutral, to ensure the best possible reach by enabling all relevant players to connect to them.

Building the infrastructure for the immersive Internet of the future is an enormous undertaking. In my next article I will share my thoughts on how and why this undertaking can only be addressed with the spirit of open standards, neutrality, and cooperation.

Interconnection is the glue that can bring infrastructure providers together, not only networks, but also data centers. If you want your data center to participate in serving the needs for the Internet of the future, get in touch – let’s talk about how we can work together.