Verizon engineers used Multi-Access Edge Compute (MEC) equipment and MEC platform software on a live 5G network that slashed network latency in half. Latency remains one of the most challenging aspects of 5G network deployments that are looking to power latency-sensitive use cases like autonomous vehicles and telemedicine.
The test took place at Verizon’s 5G test bed in Houston. Company engineers installed the equipment and software at a network facility closer to the network edge, which reduced the physical distance data needed to travel between a wireless device and the compute infrastructure. It also relied on the carrier’s Intelligent Edge Network architecture.
Adam Koeppe, senior vice president of network planning for Verizon, explained that the MEC equipment and software was internally derived and not specific to any of the dozens of “MEC”-related edge computing platforms in the market.
“This is an element of our Intelligent Edge Network and not just software defined or virtual network functions out to the cloud,” Koeppe said. “It’s really a capability that allows us to go after 5G use cases but also use cases outside of 5G.”
The test itself used an automated intelligence-enabled facial recognition application to identify people. Equipment was installed at a cell site hub location that the carrier uses as part of its cloud radio access network architecture (cRAN). The MEC equipment was able to analyze the information at the network edge where the application was being used, which resulted in the lower latency result.
Koeppe said that the testing turned in round-trip latency results of between 10 milliseconds and 15 milliseconds. This include both the backend latency, which the MEC hardware and software were targeted at, and the air interface latency that involves the specific 5G technology and millimeter wave (mmWave) spectrum used in the test.
Verizon is looking at edge networking components as part of its broader Verizon Intelligent Edge Network initiative. That initiative is targeted at changing how the service provider is running the network by making software the control point for the network, which means it’s easier to automate services and share different network assets. Company executives have also mentioned that the initiative will significantly lower the cost of running a network.
Koeppe said that the timing of commercial deployments of its MEC architecture would be tied to specific use cases that require high network speeds and very low latency. “We have our cloud platform in place so we can feel how the market is evolving and then see what types of solutions require this architecture,” he said.