The ScaleDB Storage Engine: Enabling High Performance and Scalability Using Materialized Views and a Shared-Disk Clustering Architecture

ScaleDB is a general-purpose transactional storage engine for MySQL providing High Performance and Scalability for large disk based data sets.

The ScaleDB architecture is based on a shared disk approach. It allows multiple MySQL servers to share the same physical database while guaranteeing the consistency of concurrent updates and reads operations. This approach supports growth in the data size and the number of users by adding additional servers to the cluster with no disruption to existing processes. It eliminates the need for sharding to balance data distribution and workloads.

ScaleDB indexes are unique and based on innovative Multi-Table Indexing. These indexes are highly compressed, independent of the key size, and minimize the number of IOs to update and search data. In addition, the indexes provide views over the data by retrieving data dynamically from multiple tables, in a highly efficient manner, and without the need to build the views in advance.

This session will demonstrate and explain the technology behind the Multi-Table Indexing, the architecture of the shared disk clustering, some of the challenges in integrating these technologies to MySQL and the impact on partitioning, load balancing, data consistency, high-availability, scalability and adaptability to evolving usage patterns.