How to NoSQL with Scylla, the Cassandra-Successor with Extreme Performance
In this session, you will learn how to use Scylla, the new open-source NoSQL database that succeeds Cassandra. Scylla applies new systems programming techniques to a horizontally scalable NoSQL column-store design that results in extreme performance improvements; capable of 1 million requests per second per node, with < 1msec P99 latency. Scylla also provides drop-in replacement compatibility with Apache Cassandra, as well as inheriting its scaling properties.
The Scylla design is based on a modern shared-nothing approach that eliminates the performance bottlenecks of existing NoSQL servers by running multiple engines, one per core, each with its own memory, CPU and multi-queue NIC. Scylla storage operates via asynchronous I/O through a fully-owned cache to bypass the OS page cache, and has its own task scheduler and IO scheduler as well.
Scylla leverages the best of Cassandra by mimicking its gossip protocol, flexible replication, multi-datacenter redundancy and Apache Spark integration. To simplify its manageability, Scylla automatically tunes ram, cache and admin tasks such as repair compaction, providing consistently predictable latency. Scylla is an open-source project under the GNU AGPL license.
Avi Kivity, CTO of ScyllaDB, is most-widely known for being the founder of the Kernel-based Virtual Machine (KVM) project, the hypervisor underlying many production clouds. Prior to Avi's role as CTO of ScyllaDB, he has worked for Qumranet and Red Hat as KVM maintainer until December 2012.