How Fastweb + Vodafone reimagined data workflows with Spanner & BigQuery

We were covered on the warehouse side: BigQuery was already in place and handling core analytical workloads. What we lacked was a serving layer that could match BigQuery’s speed and flexibility while simplifying our architecture. We also wanted tighter integration across the stack so we could more easily surface insights and begin incorporating AI. 

Google Cloud brought all these capabilities together in a unified way. With native integrations between products, it was the right choice. We could spend less time connecting systems and more time building the features our business actually needs.

Spanner accelerates serving without the overhead

Giving every channel real-time access to accurate customer data meant finding a serving layer that kept pace with telecom traffic. Spanner was the answer. Spanner delivers low-latency reads, horizontal scalability, high availability, and a fully managed environment with zero ops overhead.

Spanner’s native integration with BigQuery was the real game-changer. Before Spanner, we relied on several custom layers built solely to move warehouse data into downstream systems. But the direct interoperability between Spanner and BigQuery has made most of that complexity disappear. With Spanner as the front-end serving layer, we were able to migrate ten applications in two weeks.

Spanner also transformed how we monitor and maintain our workflows. Four separate monitoring processes collapsed into a single view, making troubleshooting and optimization far faster. And with Apigee exposing Spanner data to the call center, digital channels, and partner systems, the improvements were felt across the entire organization — not just within engineering.

Making sense of thousands of data flows 

With Spanner in place as our serving layer, the next challenge was making sense of the thousands of batch and real-time ingestion flows that power Fastweb + Vodafone Data platform. Understanding how data moves across these pipelines is essential for quality, compliance, and day-to-day engineering. We needed something faster and far more intuitive than our previous solution. 

Spanner’s multi model capabilities gave us that foundation. 

Spanner Graph allowed us to map lineage in a way that reflects how our platform actually works: which tables drive specific jobs, how transformations cascade, and where dependencies sit. From there, vector search and full-text search added a richer discovery layer. This made it easier to surface results quickly, even in a large and complex ecosystem. 

Together, these features gave us the governance experience we wanted — not only more accurate, but much easier for engineers and analysts to work with.