Bruno Cornaglia, Vodafone and SDN/NFV Work Area Co-Director and Mauro Tilocca, FiberCop and member of Board of Directors at Broadband Forum
It might be tempting to dismiss phrases like ‘artificial intelligence’ and ‘machine learning’ as buzzwords. After all, articles claiming that AI will transform industry x or replace technology y seem to pop up every day with promises of a radically different future thanks to automation.
Yet there is an area for which the impact of AI/ML is already delivering tangible benefits and can continue to expand its influence. Broadband networks and, more specifically, the improvement of the customer broadband experience, have much to gain from the implementation of performance and maintenance tools powered by automation.
Maintaining customer loyalty and providing an enhanced Quality of Experience (QoE) while avoiding rising costs is a key concern for any operator, a goal made more difficult by the complexity and density of data traversing the broadband network.
Broadband Forum’s Automated Intelligence Management (AIM) framework specifications are helping operators leverage the intelligence of AI and ML in the cloud to digitalize and monitor vital data in their physical networks, identifying the most impactful areas to improve performance and QoE.
Letting the AI do the work
For example, performance monitoring of the network can be used to empower proactive maintenance and fault detection. ML can support this with data profiling, rapid response, and taking predictive measures before a problem arises. One use case is the monitoring and troubleshooting of the home Wi-Fi network. Automatically detecting when service levels are not met, it results in the generation of recommendations to fix this and reconfigurations to reestablish the expected levels.
A key principle of the broadband network becoming more self-driving, it needs help from the intelligence of closed-loop automation with AI and ML embedded on all layers of the network in each network domain. Closed-loop AI is a predictive analytics software that uses ML to predict the network’s likely evolution. This can provide the likes of resource management provisioning without manual intervention and has the potential to deliver greater Return on Investment (ROI) by automating network operations, intelligent fault diagnosis, and anomaly detection.
The monitoring of network performance can also be used to improve Quality of Service (QoS) by prioritizing certain types of traffic based on user demands. Network probes that recognize network performance between the subscriber premises and any location in the network. Closed-loop service assurance, implemented by the AIM framework, identifies the best network path for service flow and preserves user experience in case of network congestion. It allocates network resources based on demand and usage patterns, dynamically adjusting bandwidth to accommodate peak periods and preventing network congestion.
The AI and ML impact needn’t stop at improving network performance, though. Its impact on network performance can even open the door to operators and providers offering new services to customers.
Differentiation with new services
From AI-powered home assistance to network personalization, operators and providers can introduce new services to differentiate their offerings from competitors and further reduce customer churn. For example, AI can analyze user behavior and preferences to personalize content delivery. By understanding individual preferences, AI can recommend content, advertisements, and services that are more likely to resonate with specific users.
AI can also assist in planning and expanding broadband networks by analyzing demographic and usage data. This can help service providers determine where to invest in network infrastructure, ensuring that resources are allocated efficiently.
As more home network users’ traffic is diverted towards bandwidth-heavy applications like VR & AR, AI can help to tailor services to meet customer needs. Looking outside of the home, cloud robotics, Industry 4.0, smart city, and smart mobility can also be impacted.
What tools do operators and service providers need?
For operators to truly make the most of what AI/ML can offer, the first port of call should be assessing how it can benefit other industry specifications. Broadband Forum’s AIM framework was designed to help service providers further improve the quality of network services delivered to customers thanks to AI, as well as solve or anticipate network faults.
The AIM specifications, of which one use case is the monitoring and troubleshooting of the home Wi-Fi network, have already been tested. Demonstrations have shown how it can automatically detect unmet service levels and generate recommendations and reconfigurations to fix them.
Returning to the concept of closed-loop service assurance, the AIM framework has showed how it can be used to find the best network path for service flow. Building on the Broadband Forum’s QED and Subscriber Session Steering work, it was demonstrated at a Broadband Forum CloudCO Demo at Network X last year and Innovation Demonstrations this year.
Further progress is being made by the AIM framework group on Issue 2 which will improve both existing TR-436 and TR-486 by closing existing gaps, by adding security considerations and by also introducing the data collection capability defined in new TR-508.
For more information and to get involved in the AIM framework project, please contact: [email protected].