Nokia and Databricks completed a proof of concept (PoC) demonstrating a unified, cloud-agnostic data platform designed to support AI-driven autonomous networks. The project validated an architecture that enables telecom operators to deploy real-time analytics, AI agents, and data services across multiple cloud and on-premises environments without rewriting code, addressing one of the industry’s longstanding challenges of fragmented operational and business support system data silos.
The PoC focused on a real-time performance management use case and demonstrated the ability to ingest and process network telemetry at scales required by Tier-1 service providers. Nokia and Databricks showed that the same data pipelines could run across both the Databricks platform and open-source environments built on Apache Flink, Apache Kafka, and Apache Iceberg. Nokia engineers also developed a platform-independent data transformation layer in Python that separates business logic from infrastructure-specific connectors, allowing workflows to operate consistently across diverse environments.
The companies also validated an automated deployment framework capable of translating abstract workflows into native execution formats such as Delta Live Tables or Flink SQL. In addition, the project demonstrated AI-assisted creation of new data products using natural-language prompts. The architecture includes query-time data products, zero-copy data sharing, and selective movement of data into higher-level cloud environments where AI agents can perform tasks such as root-cause analysis and cross-domain correlation. Nokia and Databricks said they plan to continue collaborating on technologies that enable increasingly autonomous, AI-driven network operations.
• Demonstrated cloud-agnostic deployment of telecom data pipelines without code changes.
• Validated operation across Databricks and open-source platforms using Apache Flink, Kafka, and Iceberg.
• Introduced platform-independent data transformation logic written in Python.
• Demonstrated automated workflow compilation into native execution environments.
• Showcased AI-driven creation and deployment of new data products through natural-language prompts.
• Implemented query-time data products, reducing data duplication while enabling real-time analytics.
• Supported zero-copy data sharing across operational domains.
• Designed architecture to support multi-agent AI systems performing network automation and root-cause analysis.
“Teaming up with Databricks represents a big step as we work toward building the types of data foundations required for next-generation autonomous networks. By enabling a common, flexible data platform across cloud environments, we can help operators accelerate the adoption of AI and create more efficient, resilient and sustainable networks,” said Oguz Sunay, CTO AI and Autonomous Networks at Nokia.
🌐 Analysis: The announcement highlights a key challenge facing autonomous networking initiatives: AI models and agents are only as effective as the accessibility and consistency of the underlying data. Many telecom operators continue to operate hundreds of OSS and BSS platforms with incompatible data models, creating barriers to large-scale AI deployment.
🌐 Analysis: Nokia has increasingly focused its Autonomous Networks portfolio on creating a common data and automation framework spanning RAN, core, transport, and cloud domains. The collaboration with Databricks aligns with a broader industry trend in which telecom vendors, hyperscalers, and AI platform providers are building “data fabrics” and agentic AI frameworks capable of supporting real-time operational decision-making across increasingly complex network environments.
|
Autonomous Network Data Fabric Architecture
Nokia + Databricks PoC • Updated June 24, 2026
|
|
| Core Goal | Unified AI-ready data platform across cloud and on-premises environments |
| Primary Challenge | Hundreds of telecom OSS/BSS data silos limiting AI adoption |
| Supported Platforms | Databricks, Apache Flink, Apache Kafka, Apache Iceberg |
| Data Processing | Real-time streaming, batch analytics, query-time products |
| Deployment Model | Build once, deploy anywhere without rewriting workflows |
| Abstraction Layer | Platform-independent Python transformation logic |
| Compiler Function | Automatically generates Delta Live Tables or Flink SQL workflows |
| AI Automation | Natural-language generation of new data products and pipelines |
| Agentic AI Support | Agent-to-agent creation and consumption of data products |
| Data Sharing | Zero-copy cross-domain data access |
| Advanced Analytics | Root-cause analysis, correlation, performance optimization |
| Target Outcome | Level 4/5 autonomous network operations with AI-driven decision making |