PINO G. DICORATO AI & NETWORK MANAGEMENT
THREE WAYS OF MAXIMISING AI IMPACT FOR OPTICAL NETWORK MANAGEMENT The days in which telecommunications providers (telecom providers) delivered simple connectivity between endpoints are long gone. They now must cater to a variety of services and requirements to remain relevant to their customers. The most recent challenge is meeting the need for high-speed, ultra-low-latency connections required by data centres that host compute clusters for artificial intelligence and machine learning, writes Nokia Director of Automation Solution & Technology Pino G. Dicorato .
Modern networks developed by these telecom providers are optically interconnected in many configurations, including: point-to-point, multipoint (for resiliency purposes), across the same- vendor open line system networks, and across multi-vendor optical line system networks. This is a massive, diverse and dynamic infrastructure that challenges the limits of current network operational practices. Networks must support a variety of services and service-level agreement (SLA) policies while taking advantage of the programmable optical performance enabled by coherent pluggables. At the same time, demand is no longer growing linearly at historic levels of 30% per year, but accelerating much faster, creating new challenges for deployment and service delivery. This new rate of change requires a different approach to designing and managing optical networks. It demands AI-enabled automation. USING AI TO ENHANCE NETWORK AUTOMATION AI-enabled automation initiatives are evolving current workflow-driven automation, which uses rule-based policies. These initiatives are moving toward more sophisticated intelligent systems that use trend insights to identify prematurely aging equipment in the optical network. They also enhance the user experience while addressing the growing shortage of experienced network-operator personnel. In addition, these systems increase operators’ ability to understand network behaviour. Ultimately, they
enable proactive closed-loop operations and maintenance, allowing service-level issues to be remediated before SLA disruptions impact customer experience. AI-based automation capabilities also provide further information from the network to both help prevent security threats triggered by anomalous events, and to mitigate service disruptions that impact SLAs due to external mechanical forces on outside-plant fibre caused by environmental conditions. AI-enabled automation can also incorporate deep reasoning to correlate network state or historical logs across multiple domain layers to identify potential causes behind issues with degrees of certainty. The foundation for AI-driven automation is KPI data from the optical network. This means network operators must continue
intent and fine-tune system parameters to sustain business outcomes. AI will play a key role in the shift from reactive, closed-loop operations triggered by deterministic thresholds to fully autonomous networks. Several standards forums (e.g., Mplify, TM Forum) are defining frameworks that describe the functions into which AI technologies can be integrated to enhance network resilience. These include predictive analysis, intelligent decision making, self-healing, resource optimisation, solution reporting and issue resolution. Using AI to migrate towards autonomous networks will build on
to partner with trusted optical system vendors with domain expertise in optical networks to develop the AI training algorithms so they can build operational use cases for high-value outcomes. Beyond simple robotic actions that use scripts or templates, AI-powered automation, together with automated closed-loop operations, can incorporate key requirements in the form of intents. These intents are abstractions of SLAs, route- specific policies or geographical service endpoints. They allow the underlying AI-enabled system to orchestrate the operations required to achieve the goals for a specific operational use case. In addition, AI ensures the underlying network can continuously monitor
Figure 1 - User experience enhancements with AI- enabled automation
24
www.opticalconnectionsnews.com
| ISSUE 44 | Q3 2026
Made with FlippingBook Annual report maker