AI proves to be a good choice for ensuring network diagnostics due to its ability to learn, adapt, and evolve alongside the network
The ANDIAMO suite in a few words
The ANDIAMO Suite is a software suite developed by Orange Innovation to meet the operational automation needs for optical and microwave transport networks. It includes several modules. First, ANDIAMO Data aims to collect data in order to create a data lake bringing together all information related to Orange's optical network. Next, the ANDIAMO Network Design module facilitates the design and engineering of optical networks. Lastly, the Fiber RootCause AI Ready module, to which I am contributing, uses AI for diagnosing optical and microwave networks.
The choice of AI to develop this solution
The idea is the result of extensive reflection, based on active listening to the needs of Orange subsidiaries and insights from international scientific conferences.
This needs assessment study identified several use cases, including network fault diagnosis. This particular use case was prioritized due to its importance in the daily activities of technicians.
The network is constantly evolving: new services are created, new equipment is installed, and software is updated with new versions. In such a dynamic environment, AI's ability to learn, adapt and evolve with the network makes it a good choice for network diagnostics.
Several challenges encountered
Several challenges have arisen since the start of the project. Data collection proved to be complex and required convincing operational staff to access network data and test the solution in real conditions, which was achieved with Orange France. The integration of various suppliers and the development of data collectors were also essential to ensure comprehensive coverage of the optical transport network. Raw, unlabeled data required the involvement of technicians for labeling, a process that is currently underway.
It is important to recognize the commitment of the technicians who dedicate time to this labeling process in addition to their normal duties. In terms of internal reluctance, technicians have expressed concerns about the impact of AI on their jobs, and the challenge has been to reassure them by demonstrating the benefits that AI brings to their daily work.
Benefits obtained and key lessons learned
The developed solution stands out for its in-depth consideration of the specificities related to optical layers, ensuring optimal adaptation of artificial intelligence. A patent has been filed for the clustering phases, incorporating proven techniques in natural language processing, adapted to the context of alarms and network diagnostics. This patent relies on a multi-step process: topology-based clustering, temporal tracking, and semantic validation of the clusters of alarms.
Another notable feature is the multi-vendor compatibility, allowing the solution to integrate with all Orange France equipment, differentiating it from existing solutions. Ultimately, the ambition is to extend this approach to other areas, including IP and Microwave networks, providing a comprehensive view of the network.
Currently, the solution is in production at Orange France, managing approximately 50 critical incidents per day, and is also accessible on the pre-production network of Orange Poland. Feedback from technicians indicates a processing time gain of about 30 minutes per incident, resulting a saving of 25 hours per day. This saved time will enable technicians to better anticipate diagnostics, while also reducing the number of urgent and reactive interventions.
Future perspectives
Evolutions are being considered, including the automation of certain fault resolutions, such as restarting a card, and optimizing technicians' interventions through incident geolocation. New equipment suppliers will be integrated according to the needs of the subsidiaries, and automatic assistance for technicians will be strengthened.
Other collaborations are underway, with growing interest in the Andiamo Suite, particularly from OINIS. Regarding the group's strategy, Orange aims to achieve an automation level of 4 on the TM Forum scale, currently positioned at 2.5. An RFP has been launched to improve operational efficiency across all network areas. ANDIAMO Suite positions itself as a facilitating module, integrating key data for the chosen solution, thus paving the way for generalization within the subsidiaries.
This solution relies on a synergy between artificial intelligence, optics, and software deployment, illustrating the commitment of Orange France and Poland to innovation. Thanks to close collaboration with their teams, it has been possible to test the solution in the field and achieve the current results.
