Live from the 2026 Trusted Database Development Conference: GBase Database Presents Financial Trusted Lakehouse Integration Practices

On July 9, 2026, the ‘2026 Trusted Database Development Conference,’ hosted by the China Academy of Information and Communications Technology (CAICT) under the guidance of the China Communications Standards Association (CCSA), was held in Beijing. The conference featured a main forum and three sub-forums on finance, telecommunications, and AI integration, accompanied by 14 enterprise-themed exhibition areas.GBASE was invited to participate, showcasing its full-stack database products and financial industry solutions at the booth, and delivering a keynote speech at the financial sub-forum.

Financial Sub-forum: An In-Depth Sharing on “Lakehouse Integration”

At the financial industry database application innovation sub-forum, Zhao Jie, Solution Director of GBase, delivered a speech titled ‘Thoughts on Building Lakehouse Architecture in the Financial Industry.’ He introduced GBase’s experience in building databases for financial enterprises, covering aspects such as how to construct a financial enterprise’s Lakehouse big data architecture from the data layer, the processing methods of different types of data in the enterprise data architecture, the roles and application scenarios of different database products in the Lakehouse architecture, as well as the advantages of GBase database products and financial industry case studies, offering the audience insights from real-world practices.

Starting with a Single Transfer

Zhao Jie used an online bank transfer scenario as an entry point. The moment a user completes a transfer via mobile banking, the bank needs to process data such as time, amount, and location into derived features within an extremely short time, providing them to the anti-fraud system for risk judgment; within minutes, both parties may query the transaction summary for a specific time period, and the system must immediately reflect this transfer in real-time reports; over a longer time horizon, this data will also be used for report statistics, data mining, and user profiling to support business decisions.

He pointed out that from real-time stream processing to real-time on-demand analysis to offline analysis, this entire set of requirements is difficult for a single data warehouse or data lake to support seamlessly.

The Technical Logic Behind Lakehouse Integration

At the technical level, he analyzed the respective characteristics and complementary nature of data warehouses and data lakes. An MPP-based data warehouse excels at relational storage, in-memory optimization, SQL, and analytical functions, offering good workload management for mixed workloads; while a data lake within the Hadoop ecosystem possesses massive I/O throughput, cost-effective storage, and supports streaming complex event processing. The complementarity between the two is precisely the technical foundation for the emergence of ‘Lakehouse Integration.’

Based on this insight, GBASE proposed a Lakehouse integration solution. Its core architecture uses the GBase 8a MPP Cluster data warehouse as the unified management entry point and the GBase HD data lake product as an extended capability, building unified data storage through S3 and HDFS, and providing elastic computing clusters of any scale. The entire process covers the full chain from data ingestion, data lake storage, source-proximate data cleaning, to data warehouse model calculation, and finally to data service presentation, helping financial enterprises uncover data value faster.

Full-Stack Product Portfolio

In terms of products, GBase boasts a comprehensive full-stack database system.

• GBase 8a MPP Cluster adopts a Shared-Nothing + MPP architecture with separated management and compute nodes, eliminating any single point of failure. It features key technologies such as column-level parallel processing, smart indexing, compression, a vectorized execution engine, RBO & CBO optimization, etc. The cluster supports data processing capacity exceeding 30PB, single-node processing capacity exceeding 50TB, and can scale to over 300 nodes. It is also the industry’s first MPP database supporting site-level cluster disaster recovery.

• GBase HD data lake product integrates into the Hadoop ecosystem, supporting both offline and real-time data processing modes, covering full-chain capabilities including data integration, data warehousing, multi-model computing, data marts, and ad-hoc querying.

• GBase 8s centralized transactional database supports shared-storage clusters, and deployment across two sites and three centers. It provides support for hundreds of terabytes of data, over 10,000 concurrent connections, over 10,000 TPS, 95% Oracle syntax compatibility, and 99.999% availability.

• GBase 8c multi-model, multi-state distributed database supports three storage modes—row store, column store, and in-memory—and three deployment forms—standalone, primary-standby, and distributed. It is compatible with Oracle, PostgreSQL, MySQL, and other ecosystems.

It is worth noting that all GBASE core products have passed the national security and reliability assessment. In June 2026, GBase 8s V8.8 also passed the ‘Database Basic Capability Based on Shared Storage Architecture’ product test by the CAICT, further confirming its technical strength in core financial scenarios.

Booth Spotlight: The ‘Financial Confidence’ of China’s Homegrown Database

Among the 14 enterprise-themed exhibition areas at the conference, the GBase booth attracted numerous attendees who stopped by to exchange ideas, with staff discussing topics such as Lakehouse deployment and domestic substitution pathways. As one of the earliest database vendors to enter the financial industry in China, GBase has, over more than two decades of development, provided products to over 240 institutions, covering financial regulatory bodies, policy banks, state-owned large banks, joint-stock banks, city commercial banks, rural commercial banks, rural credit cooperatives, as well as insurance and securities firms, achieving deep validation from ‘usable’ to ‘efficient’ in core industry application scenarios.

In terms of practical cases, at a major state-owned bank, GBase 8a was deployed in a converged architecture with Hadoop, featuring over 7,000 nodes and 100+PB data volume, making it the largest data warehouse project in China’s financial industry at the time, breaking the monopoly of foreign products and reducing construction costs from hundreds of millions to tens of millions of yuan, and was awarded the ‘First Prize for Banking Technology Development’ by the People’s Bank of China. At another major state-owned bank, a deployment of 5,000+ nodes and 30+PB of data saw the first large-scale application of domestic servers + operating systems + databases, achieving a lake-warehouse convergence. At a leading insurer (PICC), 20 clusters with 800+ nodes and 8PB data were deployed, delivering performance improvements of 2–10 times. Additionally, projects such as the external reporting data warehouse for a leading securities firm, the China Securities Quote data middle platform, and the Lakehouse integration platform for Luzhou Bank have all been implemented.

From ultra-large-scale clusters in state-owned major banks to agile deployments in city commercial banks, GBase, with 22 years of technology accumulation and deep financial industry practice, has validated the capability of homegrown databases to meet the most demanding requirements of the financial industry. Moving forward, GBASE will continue to delve deeper into the financial sector, driving database technology toward a smarter and more integrated evolution, and providing a solid data foundation for the digital transformation of the financial industry.