Laoyu Interview (IV) | On Data Sharing and Security Isolation for China-Developed Databases

Introduction

Recently, Laoyu Notes, a well-known independent media outlet, interviewed Cui Zhiwei, General Manager of the GBase 8s Product Division at General Data Technology, focusing on the core technology of the GBase 8s database. The discussion highlighted the application scenarios of ER data sharing and exchange replication technology, the balancing strategies between data sharing and security isolation, while also interpreting the advantages of migration tools such as MTK and RTSink over similar products on the market. Citing real-world cases in government, retail, and telecommunications, the interview clearly demonstrated the flexibility and security of ER technology, analyzed the core challenges and solutions in data replication, and detailed the optimization and upgrade achievements of migration tools, providing a comprehensive reference for understanding the core value of GBase 8s data processing technologies.

Laoyu: I remember that a core selling point of GBase 8s is ER data sharing and exchange replication. From your official perspective, can you tell us about 8s's ER data sharing and exchange replication—what application scenarios is it suitable for? In core business operations, how do you balance data sharing and data isolation security?

Cui Zhiwei, GBase 8s: The core essence of ER is a table-level real-time data replication technology with extremely high flexibility. It can realize various modes such as full-table synchronization, bidirectional synchronization, and unidirectional synchronization based on customers' actual needs, while also enabling data distribution and aggregation. The application scenarios are extensive and closely aligned with real business. For example, in the government sector, data exchange and sharing between central ministries and provincial agencies—whether the ministry distributes policy-related data downward to provinces or provinces aggregate local business data upward—can be efficiently completed using this technology. In the retail sector, supermarket front-end databases store every transaction in real time, and this data needs to be timely aggregated and synchronized with the headquarters central database for overall management and precise inventory control; ER technology perfectly meets this need. Similarly, at highway ETC toll gates, each station's vehicle passage and fee collection information also needs to be synchronized and exchanged in real time with the central system via ER technology to ensure traffic efficiency and data accuracy.

In terms of data security isolation, ER technology's precision is particularly advantageous. It can accurately distribute specific rows or columns—whether distributing a whole table or distributing only selected rows and columns. For instance, if a table has ten columns, you can choose to synchronize only the last four columns without any issues. We also perform very well in synchronization isolation. For example, when a central ministry holds data for all provinces nationwide, it distributes only Hebei provincial data to Hebei's relevant departments and Henan's data to Henan's departments, rather than pushing all provinces' data in bulk, thereby avoiding unauthorized data access or leakage and strictly adhering to permission control requirements. The synchronization mode is also flexibly adjustable, supporting real-time, delayed, and scheduled synchronization. For example, during peak business hours during the day, it can be set to delayed synchronization at night to avoid occupying core business bandwidth. It also features a breakpoint resume function: if a server failure or data center power outage occurs, data synchronization pauses, and once the equipment is restored, the system automatically continues pushing data from where it left off, ensuring no interruption or loss of data.

Furthermore, ER technology is also widely used in the microservice architectures of application developers and the intelligent network services of telecom operators. For example, some application developers equip each microservice module with a dedicated small database storing user information, authentication data, and permission data for that module. With ER technology, when permissions or user information change in one node, it can be synchronized in real time to all other microservice databases, ensuring users can log in and operate normally from any node. In the intelligent network services of telecom operators, when a user tops up or adjusts a plan, the relevant authentication information must be instantly synchronized across all nodes, and number information processing during calls also relies on synchronized permission data for authentication—all supported by ER technology, effectively enhancing business processing efficiency. Of course, there are two major challenges in data replication: real-time transmission efficiency and handling of exceptions. We continuously optimize solutions based on the unique requirements of different customers to make the technology more aligned with practical application needs.

Laoyu: The scenarios you described are all backed by actual cases, and some were even discovered by customers in their practical use, right?

Cui Zhiwei: Yes, many scenarios were accumulated by customers through real-world applications.

Laoyu: What advantages do GBase 8s migration tools such as MTK and RTSync offer over similar tools on the market in terms of migration efficiency and data integrity?

Cui Zhiwei: MTK and RTSync have been our migration tools for quite some time, but honestly, the iteration speed wasn't fast enough, mainly because server compatibility hadn't kept up. Over the past two years, we have focused on optimizing server compatibility and completely resolved this pain point, greatly enhancing the stability and adaptability of the tools. For example, in a recently undertaken ultra-large-scale data migration project involving 128TB of data, our tool was able to fully utilize bandwidth resources and maximize migration speed, demonstrating outstanding efficiency advantages. At the same time, we addressed two critical issues: data comparison and migration assessment. Data comparison not only checks data volume but also employs mature validation strategies for precise alignment, ensuring data consistency before and after migration. Migration assessment predicts the success rate in advance and adapts to full, incremental, and reverse synchronization scenarios, comprehensively safeguarding the integrity and security of data migration. These are our core advantages over similar tools on the market.