OceanBase Completed the HyBench Benchmark Test
HyBench is a benchmark for HTAP databases with a hybrid workload of OLTP, OLAP, and OLXP, supporting throughput evaluation and freshness evaluation simultaneously. It aims to guide the development of database technology, improve database product performance, and provide important references for enterprises' database selection.
Recently, OceanBase has successfully completed the HyBench benchmark and performed outstandingly in the test. In the scenario with 1000X data volume, OceanBase V4.3 achieved an H-Score of 384.39.
As an essential capability in modern data architecture, HTAP enables databases to simultaneously handle online transactional processing (OLTP) and online analytical processing (OLAP) workloads. This allows enterprises to efficiently handle multiple data requirements on a single platform, thereby improving operational efficiency and decision-making speed. The completion of the HyBench benchmark test further demonstrates the leading capabilities of OceanBase V4.3 in HTAP technology and modern data architecture.
In the context of digital transformation, HTAP is a key requirement of enterprise core systems. Based on its native distributed architecture, OceanBase can handle both transactional processing scenarios and analytical and batch processing scenarios, with one engine supporting OLAP + OLTP workloads to achieve two sets of system functionalities, significantly reducing the costs. HTAP requires high-performance OLTP. And the ability of real-time analysis is built on top of OLTP. OceanBase provides transactional and analytical processing with a single system, uses a single set of data for different workloads, fundamentally maintains data consistency, and minimizes data redundancy, thus helping enterprises significantly reduce costs.
To ensure that online transactional processing is not affected by sudden batch analytical processing, OceanBase supports resource isolation in scenarios of mixed workloads. This includes physical isolation through multiple zones and replicas, as well as using CPU resource groups and IOPS resource groups to control the usage of resources for different workloads. The system also automatically identifies and isolates slow queries to avoid affecting the overall transactional business response latency.
OceanBase V4.3 introduced a brand new columnar engine. The hybrid row-column storage solution based on the LSM-Tree architecture achieves a good balance between OLTP and OLAP workloads. Resource isolation capability ensures that online transactional processing is not affected by sudden batch analytical processing. Furthermore, the optimizer, parallel execution, vectorized engine, and other features have been enhanced in OceanBase V4.3 to improve the performance of complex queries and large data volume queries.
HyBench:
