Verizon’s utilizing engineers and IT equipment removal and relocation as they work on slashing 5G network latency by half. The latest tests, with Multi-Access Edge computing equipment and software, are part of their efforts to deal with one of the biggest issues of 5G networking, which many foresee will power latency-sensitive uses like telemedicine and autonomous vehicles.
The test was conducted at Verizon’s 5G test bed, in Houston, Texas. Company engineers went in and installed the equipment and software at a network facility closer to the company’s 5G network edge. This installation resulted in less physical distance data needed to facilitate communication between a wireless device and the computer infrastructure.
According to Verizon Senior Vice President of Network Planning, Adam Koeppe, the MEC equipment and software was something Verizon had been working on for its own operations, and isn’t related to the “MEC’-related edge computing platforms that are currently available in the market.
He elaborates, saying that it’s an element of Verizon’s Intelligent Edge Network and not just software defined or virtual network functions out to their cloud. This, he states, is a capability that allows them to operate for 5G use cases and outside of it as well, should the need arise.
The test used an AI-enable facial recognition software to identify people. The equipment was installed by IT equipment removal and installation experts at a cell site hub location that the carrier uses as a testing site in their cloud radio access network architecture (cRAN). The MEC equipment was able to analyse the information at the network’s edge, where the application was being used, which resulted in lower latency in the edge, the desired result.
Koeppe said that the testing turned in round-trip latency results between 10 and 15 milliseconds, which covers both backend latency and air interface latency.
The carrier is studying edge networking components as part of their Verizon Intelligent Edge Network Initiative, which is their work on finding out what changes they need to make in how they run their network, primarily by making their software the control point for the network, which would result in making the automation of services easier.