Next-gen Edge: What We Can Learn from the Evolution of Cellular Architectures

Jack Pouchet • July 10, 2018

Today’s avalanche of activity around the edge obscures the real story of edge computing – one of incremental evolution with mileposts marking significant technological advances. The earliest iterations of the edge were simple network closets, housing just enough IT to support on-site needs. They connected to the outside world via low-speed, low-bandwidth modems and POTS (plain old telephone service – think War Games and that dial-up modem), and had little if any real-time connectivity through the internet. Calling them the edge of the network would be a stretch, but those modest IT closets clearly represent the genesis of today’s more robust edge.

The impact of 3G and 4G LTE

The introduction of 3G cellular technologies and wi-fi changed the profile of those early network closets, introducing routers and wireless capabilities and opening the door to consumer-focused applications. Soon, Starbucks and McDonald’s leveraged wi-fi service as a competitive advantage to attract consumers seeking connectivity, and airports used the tech to make waiting for flights considerably less miserable.

Eventually, 4G LTE brought mobile connectivity and computing to the masses – first in major population centers, but the reach expanded quickly. And while the scope and nature of 4G and mobile computing varies around the world, it is not at all limited to the U.S. and similarly developed countries. In Africa, for example, the entire IT infrastructure is wi-fi enabled on a 4G backbone. As demand for high-speed computing and video increased, so too did the need for more robust edge resources.

The future state: 5G

Today we are seeing a host of emerging edge use cases that will strain current edge sites and the cellular network. As we’ve analyzed these use cases at Vertiv, we identified four major edge archetypes:

  • Data Intensive
  • Human-Latency Sensitive
  • Machine-to-Machine-Latency Sensitive
  • Life Critical

The cellular evolution increasingly has converged with this evolving IT edge, and it isn’t hard to see some parallels – not to mention progress toward a shared objective. The 5G cellular networks currently in development will rely more than ever on next-gen IT edge technologies to enable edge-dependent applications such as the Internet of Things, autonomous vehicles, and smart cities.

Eventually we’ll see traffic lights coordinated with real-time data on traffic flow and weather, smoothly adjusting for emergency vehicles to create an ultra-intelligent transportation system – all made possible by edge and 5G technologies. Virtual reality will allow us to experience sports and entertainment in whole new ways and eventually may fundamentally change the way we teach and learn.

It’s not as far off as you might think; apps like Waze are delivering similar benefits already. That’s an important point: There is a tendency to think these types of advances will be government-driven, but it’s actually more likely to originate in the private sector. Development of consumer-friendly apps and services will push increased investment in infrastructure and influence the edge and 5G architectures that emerge. Government and civil infrastructure systems are likely to be playing catch-up.

An important note on 5G cellular architectures: We don’t know what they will look like. We know they will enable more consumer applications, and likely others we haven’t imagined yet. The network architecture, however, remains largely speculative.

It’s a safe bet that we’ll see increased densification of the wireless network – after all, more sites will almost certainly be required to meet increased demand. With that in mind, providers are more likely to rely on overlapping coverage from dense networks to cover power outages at a given site, changing the equation on backup power. As batteries move out, additional computing equipment – with different thermal requirements than traditional telecom gear – will move in, likely increasing infrastructure needs at the tower site. Again, it’s a more converged edge/5G model.

Getting the next-gen edge 5G-ready.

What does all of this mean for companies planning edge deployments today? Three things:

  1. The next-gen edge must be 5G-ready. How, you might ask, is that possible when 5G isn’t defined yet? We may not know the details, but we know this: When we say the edge must be 5G-ready, what that means is it must be communication-ready. That’s the key. And already there are solutions available – Vertiv is a leader here – that deliver fully integrated, scalable edge systems with state-of-the-art communications capabilities.

  2. Existing telco/cloud/colo architectures will evolve to meet edge needs. Central offices will morph into urban data hubs. Cloud and colo companies will continue to expand their footprint to provide services close to users. Combined with the continued deployment of edge-ready local infrastructure, matched to archetypes and use cases, this will create an edge ecosystem that extends far beyond traditional small-space edge deployments.

  3. The staffing model for data centers and the edge is going to change. Today’s enterprise and hyperscale data centers are more intelligent than ever before and are moving toward lights-out operation requiring virtually no full-time staff. Edge facilities do not need full-time employees to run them, but they do need support staff and service technicians. Keep in mind, we’re often talking about thousands of edge sites in a single network. That’s going to require a massive increase in service personnel. It’s the telecom model, and there’s a good chance telcos and companies with edge IT assets will develop shared service arrangements. Of course, this creates additional issues, not the least of which is data security, but that’s a subject for another blog.

If we can learn anything from the evolution of cellular architectures as it applies to the IT edge, it’s that progress isn’t predictable and doesn’t happen in a straight line. Smart companies maintain flexibility to adapt to what’s ahead. What’s clear is the future of computing lies at the edge. To move into that future as efficiently as possible we need to fundamentally change the way we think about edge computing and distributed networks, to accept new models for traditional enterprise data centers and telco central offices, and invest in technologies engineered specifically for a future at the edge.

Related Articles

Language & Location