The TM Forum #DTW24 Ignite conference was held in Copenhagen, Denmark, from June 18−20. Three things stood out at the conference:

• Autonomous networks and autonomous operations hit the tipping point and are now widely adopted as the direction for the network and overall operations.
• genAI is solving real problems, especially in providing human interfaces to systems and information. It is also more useful as a behind-the-scene agent than I realized.
• API interfaces to provide NaaS are widely touted, but most of us are skeptical about their ability to provide much new revenue from enterprises.

Autonomous Networks and Autonomous Operations Hit the Tipping Point

Huawei, China Mobile, and several others presented at the AN Symposium at the conference, discussing their success in achieving Level 3 operations autonomy and moving to Level 4 in selected areas in the last few years. This continued the long-standing tradition of them reporting on their progress. Nokia and Ericsson also co-presented with their customers on how they are helping the CSPs move through the various levels of autonomous network operations.

MyTake

I have discussed in the past how the early majority of CSPs have signed onto the AN and AO visions of the future, with many making statements about reaching Level 3 and Level 4 autonomy in selected areas within the next few years. The industry has gotten much more real in its thinking with 100% AN/AO as not a goal and only in selected areas.

It is highly important that Ericsson and Nokia both are publicly stating how they can help their customers achieve L3 or L4 operations autonomy using the AN/AO terminology and concepts.

genAI is Solving Real Problems

Demonstrations and discussions of gen AI (and other AI techniques) were everywhere. The most impressive demo I saw was from Amdocs and AWS, which used AWS AI cloud technology and Amdocs systems (also running on AWS). The demo detected macro-cell overloads, suggested mitigation by turning on several micro-cells and adjusting the parameters of the macro-cells, got permission from a person, and did it. It also patterned over time that this congestion happened every Sunday at the same time and asked if these mitigation measures should be done every week.

MyTake

genAI is real. It obviously is good for intent-based interfaces and general knowledge management, but has many more applications, including an example as above, patterning and reporting on network fault, performance, and service performance information in human terms, not requiring the technicians to learn the ins and outs of the various systems.  Its use as an agent, working behind the scenes, is more promising than I realized before the conference. Other AI technologies will also be brought to bear, such as video processing.

We are Attempting NaaS Once Again

NaaS (sometimes also called CaaS) was discussed quite a bit as a new source of revenue for CSPs as well as a way to increase cross-CSP transport offering efficiency. CSPs and vendors touted the number of developers they have signed up to use the interfaces being opened. But not much commercial is happening yet.

MyTake

We are trying this, once again. I hope for the best, although it feels like a “must do” for CSPs to meet the minimum needs of their enterprise customers, not a great revenue-generating offering. I hope I am wrong.