Edge computing is more than just a hot topic. It’s poised to significantly reshape telecom networks as we know them, driven by 5G and the internet of things (IoT), but also gathering momentum of its own.
Consider this: Mattias Fridstrom, chief evangelist at Telia Carrier, said recently in an interview that he doesn’t expect his company’s backbone network traffic volume to be dramatically impacted by 5G and IoT because “the vast majority of that traffic is likely to stay local.” Telia Carrier is an international network operator that connects some of the largest networks, public clouds and content providers, but its top visionary sees most of the content and data that 5G and IoT generate as being stored or handled locally.
For that to happen, there needs to be significant edge-computing infrastructure put in place, and we’re seeing that first waves of that.
One example is the recent burst of open source activity at the edge. The Open Networking Foundation launched a major new initiative, driven by eight leading communications service providers (CSPs), which is intended to jump-start commercialization of open source SDN activity at the edge. That effort has already produced four reference designs that are intended as blueprints for vendors, to pull them into the open source process and get them producing what operators need – and are committing to deploy – at the edge.
Add to that the Akraino Edge Stack project, launched this summer with seed code from AT&T, with the goal of creating key blueprints for telco-based edge computing. Its early work is focusing on the operator’s own edge, which can be at the customer’s premises or other site, as well as the traditional edge, including base stations, central offices and points of aggregation.
Vendors such as Juniper Networks are rolling out new smaller footprint versions of their technology designed for the space and power constraints edge will require. And the network operators themselves are talking openly about where and how they plan to take advantage of their existing base stations, central offices and access facilities to create a distributed cloud environment that can support low-latency requirements for some applications and potentially house content as needed.
Where is the edge?
Despite all this activity, however, there is still a lot we don’t know about edge computing and how it will evolve, starting with how to define what the ‘edge’ really is – or where it is. Industry experts offer plenty of ideas and perspectives on that.
For example, Joe Reele, VP of datacenter solution architects at Schneider Electric, believes there is a quiet real estate land grab currently under way. And companies such as EdgeConneX are among the players looking for the logical locations at which edge traffic might be aggregated or analyzed.
It’s also not yet entirely clear whether network operators will be sitting pretty at the edge, given the physical infrastructure they own, if they aren’t able to roll out edge infrastructure as economically as companies like Amazon, Facebook and Google have scaled their infrastructure.
The ability to do this economically at scale is critical for CSPs, which is why they’re driving open source groups to move faster and help them deploy without locking them into a single vendor’s approach.
There is one safe bet, however. Edge computing will not be done on traditional vertically integrated systems. Rather it will exploit, to the greatest extent possible, SDN and network functions virtualization, using smart software and commodity hardware, while counting on evolving open source standards to drive both.
Written by Carol Wilson, a freelance writer and analyst who has covered the telecom industry for more than 30 years | Original story posted at TM Forum Inform.
The TM Forum’s Digital Transformation Asia event is coming to Kuala Lumpur on November 13-15. Disruptive.Asia is an official media partner with the event. All the details you need are right here.