I recently had the pleasure of attending this year’s FŸUZ conference, held in the charming city of Dublin for the first time. For those unfamiliar, FŸUZ brings together the TIP (Telecom Infra Project) community, an annual gathering of telecom professionals, visionaries, and thought leaders. As a long-time observer but first-time attendee, I was curious about what would be discussed.
The concept has always intrigued me – my preconception being an innovative movement, a fusion of pioneering work led by theoretical idealists and ambitious visionaries. It seemed to be supported by major operator groups that are eager to shift away from the rigid market dominance of the two main vendors, limiting choice.
However, what I discovered was an earnest team of devoted professionals, who, over the years, have been collaborating with a shared belief in their collective potential. These individuals have the confidence that they can achieve remarkable feats and truly unlock the potential of technology when working together. They aspire to nurture a thriving, dynamic ecosystem.
However, it must be noted that the take-up from operators who weren’t initially involved and committed to this initiative has been less than enthusiastic.
I was initially surprised that the event was over three days but it seems that most of the community has something to say and add to the discussion as to how these initiatives can move forward. I could categorise the event into three main themes:
- O-RAN, disaggregation, RIC, API’s- deployments, proof points and customer success stories
- TIP Open Transport, disaggregation, T-API’s MUST project, being the optical and IP together
- Network Automation and AI (enabler for the above). Surprised how much was about generative AI.
Amid the intellectual discourse, I also relished the opportunity to explore Dublin’s Docklands area, which is being used for trials in collaboration with Trinity College and the Connect Centre on a Sunday tour led by Luke Kehoe.
It was encouraging to hear from the representative of Telus that they have had successfully swapped out Huawei RAN, and have replaced it with an O-RAN compliant stack working with Samsung. Key to the success of this project being that they were able to meet and match the KPI’s of Huawei. This is a victory for sceptics who doubted O-RAN’s maturity.
Kudos also to DISH who rightly claimed that they are the flag bearer for the industry and delivering a fully cloud-placed disaggregated network. It is time for other operators to step up to the mark and fully embrace this technology.
However, the journey to full disaggregation still faces minor setbacks, with key components being provided by single vendors, despite claims of O-RAN compliance. This pattern is evident in deployments like Telus- Samsung and Ericsson- AT&T. There are other installations such as with Vodafone which follow a similar pattern.
The ultimate aspiration is to achieve a fully disaggregated RAN deployment over time, with continuous evolution in the performance of standard chipsets CPU GPU.
Another assurance for the future was Ericsson’s declaration of officially joining TIP. Without mentioning names, one of the leading researchers for one of the tier one Operators said to me that this would make their life easier as they would no longer need to explain to management whether they would be committing to ORAN or not based on this announcement.
However, going forward a note of caution must be sounded for operators not to regress and show favouritism towards single vendors, such as Ericsson, claiming O-RAN compliance. Such a move would be a setback to the TIP ecosystem’s collective effort.
You will not be surprised to hear that naturally, AI was a keynote topic. NVidia mentioned the AI RAN initiative in North America where T-Mobile US, Ericsson and Nokia have agreed to invest in an AI-RAN innovation centre, in line with the GPU hype curve. I was surprised that the main part of the NVidia presentation was related to Generative AI, without mentioning LLM and not mentioning more traditional preventative and route cause analysis operational telecom use cases. The pitch was to persuade Operators in the value of investing in GPUs for AI RAN and providing GPUaaS back to the market to increase the utilisation creating a positive business case overall. The potential achilles heel I see in this approach is making sure the backhaul connectivity can support this type of distributed workload at the edge.
I would also like to take the opportunity to show appreciation for the Indonesian government’s initiative to establish a digital strategy and ecosystem surrounding O-RAN, encouraging collaboration between academia, universities and Operators.
I extend my thanks to James Crawshaw from Omdia and Robert Curran from Appledore Research for the great discussion over dinner and commend Eugina Jordan, Kristian Toivo and the entire FŸUZ team for conducting such an inclusive event bringing together the best our industry has to offer.
