Shenzhen’s Jiuyou Database, in joint research with Peking University, has passed a national-level sci-tech achievement evaluation, a marker of progress in China’s basic-software layer. The expert panel was chaired by CAS academician Zhang Jingzhong, a leading mathematician and computer scientist whose work spans machine proof, educational mathematics and intelligent-reasoning engines.
What the evaluation found
The panel concluded the project fills a core gap in China’s domestic multi-modal AI databases and offers a new paradigm for modernising basic-software architecture and fusing AI with the data layer. Targeting the strategic need for local substitution in critical fields, it tackles entrenched industry problems: rigid sharding in distributed databases, low access efficiency across single-modal databases, and coarse security granularity. It independently broke through dual distributed architecture with heterogeneous-replica intelligent routing, multimodal hyper-convergence with cross-modal strong consistency, and an AI-native, security-fused kernel, forming a fully self-controlled next-generation hyper-converged database solution.
Already in production
The achievement holds 21 authorised invention patents, 24 software copyrights and 10 industry and group standards. It has been deployed at scale in finance, government, energy, ocean and transport, and fully adapts to the Kunpeng, Hygon, Kylin and UnionTech Xinchuang stack, pushing forward coordinated upgrade of the localisation ecosystem. The experts rated the work domestically leading overall, with the dual distributed architecture and heterogeneous-replica technology internationally leading.
Read the original report (Shenzhen News Network)
Translated and adapted from Shenzhen News Network (sznews.com).