Case Study: Krux achieves always-on, low-latency, scalable user data at 50,000 requests per second with DataStax Enterprise

Krux - Customer Case Study

Krux is a data management platform (founded in 2010) that helps publishers and brands protect, manage and monetize audience data across screens for more than 700 million users worldwide. Faced with a growing customer base, more than 1 TB of client data and peaks of roughly 50,000 requests per second, Krux outgrew a single‑machine data store and needed a horizontally scalable database that could deliver continuous availability and very low latency as mobile and device traffic expanded.

Krux chose DataStax Enterprise (Apache Cassandra) on Amazon EC2 for its continuous availability, tunable consistency and Hadoop integration, powering three core use cases: the real‑time user data service, ETL/analytics and dynamic audience generation. The deployment now processes ≈50,000 rps, turns user actions into segments in under 75 ms, handles concurrent updates from millions to hundreds of millions of rows, serves low‑latency reads from cache, improved page performance by more than 30%, and benefits from responsive DataStax support and simpler operations.

Krux

Paul Lathrop

Senior Infrastructure Engineer, Krux