GBase 8a: Continuous Innovation, Driving Digital Business

Published on 2026-07-01

At the 2026 GBASE Technology Cloud Sharing Conference, Guan Lianpo, General Manager of the GBase 8a Product Business Department at General Data Technology, delivered a keynote titled “GBase 8a: Continuous Innovation, Driving Digital Business.” This article focuses on Guan Lianpo’s session, tracing GBase 8a’s evolution from standalone to cloud data warehouse, from data lakehouse to DataAgent, and exploring how continuous iteration meets the new demands of the digital era.

GBase 8a Product Family: The Full Landscape from Standalone to Cloud Data Warehouse

Guan Lianpo began by outlining the complete GBase 8a product family, covering a full range of scenarios from small and medium-sized deployments to hyperscale, and from traditional deployment to cloud-native:

GBase 8a Standalone

Ideal for scenarios with total data volumes under 50 TB and demanding high-performance analytics. It delivers columnar storage, high availability, vector search, and more.

GBase 8a Cluster

The most widely deployed form factor, designed for massive-scale data analytics with petabyte-level capability. Supports data lakehouse, real-time analytics, and hybrid AI search.

GBase 8a Cloud Data Warehouse (GCDW)

A next-generation cloud-native data warehouse built on a disaggregated storage and compute architecture. Suitable for migrating data warehouse workloads to the cloud, centralized management, and optimizing costs and efficiency as enterprises modernize their technology stacks.

GBase HD

A data lake product that together with GBase 8a forms a comprehensive data lakehouse solution.

GBase 8a DataAgent

A next-generation Data+AI agent that enables efficient exploration of data value.

Guan Lianpo specifically noted that the existing coupled-compute GBase 8a Cluster can be smoothly upgraded to the GCDW cloud data warehouse using migration tools, offering a clear path for evolving enterprise technology architectures.

Core Technology Iterations: Disaggregated Storage & Compute, Data Lakehouse, and Active-Active Disaster Recovery

Disaggregated Storage & Compute
GBase 8a Cloud Data Warehouse GCDW

Guan Lianpo highlighted GCDW’s core capabilities. Metadata is fully separated from data, and all components are stateless, enabling compute resources to be scaled in or out online in seconds. Multi-tenancy allows independent compute resources to be created for different business units on the same infrastructure, achieving resource isolation and elastic provisioning. The data lakehouse leverages a unified metadata hub through Catalog, allowing direct access to and analysis of open-format data stored in the data lake. With a single copy serving both warehouse and lake, it dramatically reduces data redundancy and ETL costs.

Columnar Transactions
Lightweight Transactional Capabilities for OLAP Scenarios

Transactional capabilities have been added to the columnar storage engine to meet lightweight transaction requirements in OLAP scenarios. The GTM component handles transaction ID generation and state management, ROWINFO implements MVCC multi-version visibility, and row-level locks support concurrency in the thousands. Logged and non-logged tables coexist, balancing analytical performance with transactional capability.

Active-Active Clusters
Disaster Recovery & High Availability

GVR provides a graphical interface for configuring, managing, scheduling, monitoring, and alerting for GBase 8a active-active clusters. It delivers flexible configuration and highly efficient synchronization, supporting scenarios such as intra-city active-active, cross-city active-active, two-site-plus-one-center, data center relocation, read-write splitting, and heterogeneous hardware replacement. It is the first MPP database in China to support cross-data-center high availability.

Continuously Enhanced Security

The HSM module generates, stores, and operates on root keys and data encryption keys. It supports China’s cryptographic standards and TLS 1.3 communication encryption. The performance impact of data encryption is kept below 5%, and it supports custom masking functions and dynamic data masking.

Guan Lianpo concluded that GBase 8a’s technical iteration never stops—a data lakehouse integrates all enterprise data, disaggregated storage and compute enables centralized management, full-lifecycle operations are fully visual, active-active clusters ensure business continuity, and security capabilities continue to strengthen.

Data+AI: GBase 8a DataAgent Intelligent Data Analytics Platform

During the keynote, Guan Lianpo introduced GBase 8a’s latest initiatives in the AI era. The GBase 8a unified data-intelligence platform follows a technology roadmap of “data lakehouse architecture + ontology-based semantic layer + natural language querying + end-to-end intelligent operations.”

The semantic layer unifies business concepts, data metrics, and indicator systems through a three-tier model: business ontology, data ontology, and technology ontology. DataAgent leverages the semantic layer, metric systems, and knowledge graphs to minimize AI hallucinations. AI+GDS assists with SQL generation, SQL quality management, and query modeling. GDOM+DBClaw uses AI algorithms and skills to automate end-to-end monitoring, operations, remediation, and analysis, significantly reducing manual operational costs.

Guan Lianpo emphasized that GBase 8a DataAgent is driving a data democratization revolution—business users can now converse directly with data using natural language, without needing SQL skills. Through conversational feedback, the system continuously accumulates business semantics, refines query logic, and extracts more data value, delivering a one-stop service from data ingestion to value discovery.

Industry Case Studies: Proven at Scale in Finance and Telecom

Guan Lianpo shared a series of real-world cases.

  • In a cloud data warehouse project at a major state-owned bank, the bank began upgrading to a disaggregated storage and compute architecture using GCDW in 2023. Multiple business modules have already gone live, reducing compute resource requirements by over 50% and saving 90% on disaster recovery cluster resources.

  • In a data warehouse project at another major state-owned bank, since 2013 the bank has pursued a combined MPP and Hadoop technology path, building a cluster with over 7,000 nodes managing more than 100 PB of data. At the time, it was the largest data warehouse and big data platform in China’s financial industry. In 2017, it received the first prize for Bank Technology Development from the People’s Bank of China.

  • In a provincial transport digitalization project, a unified data lakehouse foundation was built using GBase 8a and GBase HD. A data service platform with unified standards was established, forming a stable near-real-time data synchronization link. The data lakehouse integrates provincial-wide transportation data, providing a unified data application foundation for the shift from digitalization to data-intelligence.

  • At a leading telecom operator’s data warehouse, headquarters and provincial branches have deployed over 6,000 nodes cumulatively. Multi-tenant isolation enables resource pool management, supporting self-service reporting and multi-dimensional analysis.

Guan Lianpo’s speech outlined GBase 8a’s continuously evolving technical roadmap. From standalone to cloud data warehouse, and from data lakehouse to DataAgent, GBase 8a meets the new demands of enterprise digitalization through relentless technical iteration. Large-scale deployments in finance and telecom prove its maturity and reliability in hyperscale data scenarios.For the AI era, GBase 8a is embracing “Data+AI” as its core direction, transforming data analytics from a domain of specialists into a democratized service.