Given operator testing trends, expect 2024 to be a breakout year for 5G Standalone deployments
The role of artificial intelligence (AI) in telecoms was undeniably the main theme of Mobile World Congress in Barcelona. But beyond the hype and flash, Spirent is drawing on its extensive work in automated testing and assurance to focus in on the sheer backend datacenter velocity AI workloads require, and how AI will accelerate the deployment of 5G Standalone networks.
Discussing the AI-centric announcements, conversations and demonstrations at MWC, Spirent’s Doug Roberts, vice president and general manager of automated test and assurance, said the company’s focus is on the “unspoken portions” of networking for AI workloads and AI-enabled networking, specifically at backend data centers. As people eagerly use tools like ChatGPT, “What they don’t realize is the backend infrastructure required to process these workloads. And that’s what we’re really focused on…getting Ethernet right.”
This means that the underlying data architecture powering AI-based applications has to focus on lossless transmission and latency “more than ever,” Roberts said. “Everything now has to work together, and it has to work perfectly.” He acknowledged that there was some “give” in 5G and/or Open RAN networks as it relates to lost packets. But, “If I insert an AI workload into that environment, one packet loss could literally render the return of that AI request completely void. The sensitivity and the velocity of these networks has now increased exponentially with the advent of AI.”
Ahead of MWC, Spirent announced a high-density test solution capable of emulating AI workloads over Ethernet and testing tier impact on AI data center networks and interconnects. Read more about that announcement here.
Watch the full video interview with Roberts and his colleague Stephen Douglas, head of market strategy, here.
Douglas, referencing Spirent’s fifth annual report, 5G 2024—Market drivers, insights and considerations, called out the ramp in interest for 5G Standalone testing and what that signals for commercial deployment. Based on Spirent’s 500 engagements with 150 customers during 2023, he sees a ramp in testing focus on 5G Standalone as “really good insight to an acceleration of the commercial launches this year…This transition to 5G Standalone, and the cloud-native environments, is triggering a complete refresh cycle of how [operators] actually monitor their networks for performance and quality. The old passive ways of monitoring seem obsolete to some degree because they now need to be dynamic and proactive.”
So what does this combination of 5G Standalone and strong interest in AI mean when taken together? Roberts added the context that around five years into the 5G cycle, historically the midpoint of a typical cellular generation, there has been little material revenue lift based on new delivery of advanced network-enabled services (although fixed wireless access has been a brightspot with an early consumer focus giving way to an enterprise focus, according to the Spirent report).
“There just hasn’t been that magic trick,” Roberts said. “But now when we see AI being brought into the mix, and you think about that from a backhaul perspective, and the AI workloads in conjunction with all those edge-based services, it really does feel like the tipping point for 5G Standalone now accelerating to the levels that, quite frankly, we expected two years ago.”
