How to Save Latency and Money with ScyllaDB | ScyllaDB Summit 2017

After four years of using Apache Cassandra in production, Allegro Group has 15 clusters running on over 200 nodes (LXC containers). After attending ScyllaDB Summit 2016, they decided to run ScyllaDB tests in their environment expecting to see much better latency and higher throughput that would allow them to reduce cluster size and hardware cost. This presentation shows the result of their tests including performance metrics and a comparison of maintenance operation times (join, decommission, cleanup, repair) and the hardware reduction plan.