Financial IT Application Innovation Frontier: GBase's Research Projects Accepted by Innovation Application Committee

Recently, the Innovation Application Committee of the Beijing Fintech Industry Alliance (hereinafter referred to as the Innovation Application Committee) officially announced the acceptance results of the 2022 research projects. The multiple innovative achievements led by GBase, including "Support for Distributed Databases on China-Developed Heterogeneous Platform Servers", "Research on Application of Compression/Decompression Hardware in IT Application Innovation Distributed Databases", and "Research on Multi-Level Caching in IT Application Innovation Distributed Databases", have passed the committee's acceptance review.

To implement the requirements of the People's Bank of China's fintech development plan and promote the scientific and sustainable development of financial IT application innovation, the Innovation Application Committee was established. Close collaboration among industry, academia, research, and application entities within the committee has focused on research around technological innovation and business scenarios, effectively solidifying the theoretical foundation for advancing financial IT application innovation.

The approval of GBase's series of research achievements in IT application innovation distributed database applications provides effective support for the Innovation Application Committee's promotion of financial IT application innovation and distributed theory research.

Research Topics

(I) Support for Distributed Databases on China-Developed Heterogeneous Platform Servers:

This project optimized the database for in-depth adaptation to China-developed CPUs, masking differences in CPU architectures (ARM, x86, SW), reducing branches in business application code so that it can be deployed simultaneously on different CPU architecture platforms.

Optimization for the NUMA multi-core architecture of China-developed CPUs enabled the distributed database to fully leverage the characteristics of the domestic NUMA architecture and unleash multi-core performance advantages.

Moreover, to address the memory misalignment issues of China-developed CPU chips, the distributed database optimized memory usage to enhance efficiency.

(II) Research on Application of Compression/Decompression Hardware in IT Application Innovation Distributed Databases:

This project leveraged the acceleration features implemented in CPUs of new-generation servers. Through optimization and transformation, the distributed database fully utilized hardware acceleration features to improve computing service efficiency, enhance server performance, and reduce hardware costs.

(III) Research on Multi-Level Caching in IT Application Innovation Distributed Databases:

In distributed databases, as computing nodes manage increasing data volumes and individual node failures cause their replica nodes to handle more computation, the performance of computing nodes reading data files from distributed storage services declines under a storage-compute separation architecture. This requires users to configure larger memory capacities to cache more data. This project studied multi-level caching technology for distributed databases, using local SSDs on computing nodes and new types of hardware such as non-volatile memory as part of the data cache to improve database performance.

Facing new database trends such as open source, cloud computing, and distributed systems, GBase breaks free from convention. It has launched a new generation cloud-native data warehouse, GCDW, while closely engaging with the Chinese open-source root community openGauss and commercially promoting the distributed database GBase 8c. GBase actively embraces new industry directions, building on the past and innovating for the future.