Case Study: CERN achieves 6× throughput and scalable, resource-efficient computing with ScyllaDB

CERN Optimizes Computing Resources with Scylla

CERN, the large European laboratory operating the ALICE experiment on the Large Hadron Collider, needed to replace its MySQL-based AliEn Global File Catalogue to handle rapidly growing metadata (projected 5–10x increases in compute and storage). Their requirements included high availability, horizontal scalability, no single point of failure, consistency and transparent sharding, and they evaluated Cassandra and ScyllaDB as candidates.

CERN selected ScyllaDB, which delivered out-of-the-box performance and much better resource utilization than Cassandra—ScyllaDB achieved up to 6× the throughput in CERN’s tests—allowing smaller clusters, simpler tuning, and lower space/budget/personnel pressure. In addition to the measurable throughput gains, CERN cited improved scalability and a supportive ScyllaDB team as key benefits.

CERN

Miguel Martinez Pedreira

Computer Engineer