RapidsDB - Distributed, In-Memory, Real-Time Big Data Analysis Engine
RapidsDB accesses and processes data directly in memory. No wait time of loading data from disk to memory is needed. Query processing request is broken up into many smaller tasks, which are distributed intelligently and executed in parallel across nodes for real-time processing and analysis. The distributed architecture of RapidsDB allows a horizontal expansion of hardware. Additional nodes can be added to clusters to meet ever-growing business demand. The unique embedded federated connector system enables users to access various data sources with industry-standard SQL and JDBC interfaces, dispensing with the traditional Extract-Transform-Load (ETL) process and unnecessary data migration.
- Distributed, MPP, shared-nothing memory database
- Unified ANSI SQL query support for multiple data sources
- Adaptive query pushdown and dynamic query optimization
- Multiple table joins across nodes
- Distributed in-memory data storage