Case Study: CERN achieves scalable, fault-tolerant data processing with Akka

Akka helps CERN to keep groundbreaking physics experiments running smoothly

CERN, through its Accelerator Controls Group, needed a way to keep the NXCALS data acquisition system running smoothly while handling massive, fast-growing subscription loads. Using the Akka framework and its Master-Worker model, CERN built a scalable, fault-tolerant processing setup for its Java-based time-series acquisition software.

Akka enabled CERN to distribute workloads dynamically, rebalance workers, and isolate problematic subscriptions to protect system stability. The result was a fully horizontally scalable system processing about 90,000 subscriptions and roughly 100,000 messages per second, with additional reliability achieved by adapting Akka for at-least-once message delivery.

CERN

Marcin Sobieszek

Senior Computing Engineer