I attended the 2024 Ultra-Broadband Forum 2024 on 30 October through 1 November in Istanbul, Türkiye. I also chaired the Huawei Autonomous Driving Network Summit session. In the sessions, Huawei and several CSPs described their successful efforts to implement a Fixed 5.5G vision, achieve Level 3 Autonomous Network autonomy, and their programs to achieve Level 4 autonomy.

Keynotes and General

With over a dozen keynote speeches by CSPs and Huawei, there was a lot of ground covered. CSPs mostly described recent initiatives in 5.5G and how they fit with the customers’ needs for greater bandwidth, higher satisfaction, and the CSPs’ requirements for cost reduction, increased customer satisfaction, higher ARPU, and greater revenues from emerging more-than-connectivity services.

Highlights

• Artificial Intelligence was everywhere (of course). Many different CoPilot LLMs were introduced that aimed at providing information and recommendations to technicians and engineers. A few AI Agents that provided closed-loop operations were discussed, but they provide only very limited capabilities. There was a lot of discussion about “Networks for AI” (how to support enterprise customers as they build and train AIs) and “AI for Networks” (how AI can be used in network operations and how to build new services around AI). The hope was that AI could achieve three things (in order of importance): cost reduction, increased customer satisfaction, and expanded revenue. Huawei claimed that every $1.00 invested in AI, it is estimated to provide a $4.60 return.
• Nearly all of the over a dozen talks in the keynotes expressed the importance of network automation, but none mentioned the Huawei Autonomous Driving Network by name, nor the TM Forum Autonomous Network program.
• The concept of 5.5G for fixed networks was articulated as important, involving 100G, low-latency networks, and AI technologies, but it was called by many different names, with seemingly somewhat different characteristics, depending upon the talk, including 5.5G, F5GAdvanced, NG5.5, NET5.5G, UBB5.5G, and “Advanced AI-Centric UBBF5.5G Network.”
• Speakers stressed the importance of higher bandwidths for fixed networks (with no discussion of slowing bandwidth usage growth, as there is in mobile) with low latency. I also heard lower “digital jitter” mentioned for the first time as a growing requirement by applications. Talks described the growing need for <1ms latency transmission within an area, 5 ms within a city, and 10 ms within a country.
• There were many examples of offering “Fiber to the Room” with 1GBPS service everywhere, using PON to the premises and transparent short-run fiber within the home to multiple Wi-FI-enabled ONUs.
• All-optical networks were a major topic. This is not surprising as Huawei has been supplying them for the robust all-optical networks within China for some time.
• Swisscom described its recent “100% available network” for enterprises. To offer that, they are building a completely separate network as a (presumably hot or warm standby) parallel network.
• Many CSPs described how they are providing home broadband services, including Wi-Fi connectivity within the home. They see this as a major growth area, potentially including sophisticated home controllers for advanced consumer services.
• Broadband customer diversificationwas a major topic. Companies that focused on home broadband talked about needing to go more into the SMB and enterprise spaces. Companies that focused on enterprise talked about supporting home and SMB markets more.
• SRV6 is happening in the core networks to provide greater engineering control, especially for low-latency services.
• Slicing is still an opportunity “for the near future,” with little commercial application today.
• CSPs are building out their own private cloud computing centers for both internal use and revenue-producing services.
• The TMForum has redefined the characteristics of Autonomous NetworksLevel 3 (Automatic monitor, query, and analysis for predictive troubleshooting) and L4. L4 has been broken into L4 Phase 1 (Autonomous closed-loop within domains by agent) and L4 Phase 2 (E2E closed-loop by multi-agents collaboration). The requirements for self-adapting network automation processes depending upon network conditions were also removed from L3.

MyTake: This is not unexpected. The engineering push for quickly defining and implementing operations flexibility, intent, closed-loop operations, and cross-domain orchestration was too aggressive, as I have been saying for years.

MyTake: The Right Direction, but theTiming Is Optimistic

• The Huawei push for 5.5G is still muddled, internally and externally, and not well adopted in the industry.
• The Huawei all-optical network is ahead of the rest of the world, pushed by the China market.
• The home broadband market represents a major growth opportunity for CSPs offering fixed services.
• Vendors and CSPs hope that network slicing and low-latency services will be future differentiators. The very low latency requirements for the new services are not credible. And slicing is all but moribund.
• I was very impressed by the projects and plans of many CSPs serving smaller and emerging national markets. There is more innovation happening there than in most advanced countries.

Click to read more of Marlk Mortensen's blogs.