How cloud and AI are changing the future of network architectures

The intersection of cloud computing and artificial intelligence is reshaping telecommunications network architecture, with Nokia leading efforts to create seamless network-cloud integration across multiple infrastructure layers.

The big picture: Nokia’s vision centers on developing a network-cloud continuum that integrates devices, edge infrastructure, and network core components, incorporating both traditional AI and generative AI technologies throughout the system.

• Nokia’s Cloud and Network Services CTO Jitin Bhandari estimates that Communication Service Providers (CSPs) could access an additional $1.1 trillion market across multiple vertical sectors
• The integration of cloud and AI technologies presents significant implementation challenges that require careful consideration and planning
• CSPs possess unique capabilities in building and managing large-scale networks, positioning them well for this transformation

Current AI implementation status: Nokia has already demonstrated significant progress in AI integration within telecommunications networks, showcasing practical applications and real-world results.

• The company has successfully implemented over 80 use cases with more than 300 operators using machine learning-based techniques
• Conversational AI and agentic AI are identified as the next major developments, enabling natural language interaction with networks
• The combination of structured and unstructured data in multi-modal generative AI models is expected to enhance network capabilities

Future timeline and predictions: Nokia projects significant changes in network architecture and functionality within the next three to five years.

• Networks are expected to become conversational, allowing natural language interaction
• Multi-modal generative AI models will process both structured and unstructured data
• The evolution will require significant changes in how networks are operated and maintained

Implementation roadmap: Nokia outlines three critical areas that CSPs must focus on to achieve this transformation.

• Use case development must prioritize understanding human personas and their interaction with networks
• Data management requires robust governance, observability, and privacy frameworks
• Model selection should balance between closed and open source Large Language Models (LLMs), with consideration for domain-specific smaller models that may provide faster results

Technical considerations: The successful implementation of AI in network architecture requires a sophisticated approach to model development and management.

• Organizations need to develop “model farms” or “gardens” to manage various AI techniques
• Essential skills include knowledge graphs, low-rank adaptation, and retrieval augmented generation
• The focus should be on operational techniques rather than just foundation models
• Data science expertise is crucial for implementing these various AI techniques effectively

Looking ahead: While Nokia’s vision presents a compelling future for network architecture, success will largely depend on how well CSPs can navigate the technical complexity while maintaining security and reliability. The integration of conversational AI capabilities could fundamentally change how networks are managed and operated, but this transformation will require careful balance between innovation and practical implementation.