HIT Shenzhen-GBase Cloud Database Research Center Holds Achievement Exhibition to Support the Growth of China's Homegrown Databases
On November 4, 2021, the "Domestic Database Industry-Academia-Research Cooperation Exchange and HIT Shenzhen-GBase Cloud Database Research Center Achievement Showcase," hosted by Harbin Institute of Technology, Shenzhen (HIT Shenzhen), Tianjin General Data Technology Co., Ltd. (GBase), and Shenzhen Computer Society, and organized by the HIT Shenzhen-GBase Cloud Database Research Center, was successfully held at the Sheraton Shenzhen Bolin Tianrui Hotel.
This achievement showcase summarized the work of the HIT Shenzhen-GBase Cloud Database Research Center over the past year, presenting the center's current research status, achieved results, and future expectations. The conference also invited several database experts from academia and industry to deliver technical reports on topics related to China's homegrown databases.
More than 50 experts and representatives attended the conference. Key guests included:
Zhao Yijie, Director of the Science and Technology Department, Harbin Institute of Technology, Shenzhen
Professor Li Jianzhong, Recipient of the National Science Fund for Distinguished Young Scholars, Winner of the CCF Wang Xuan Award, Chief Scientist of the National 973 Program Project, and Professor at the School of Computer Science and Technology, Harbin Institute of Technology
Professor Wang Xuan, Vice Chair of the Guangdong Computer Federation, Chair of the Shenzhen Computer Society, Deputy Director of the HIT Faculty of Computing, Director of the HIT Shenzhen-GBase Cloud Database Research Center, and Dean of the School of Computer Science and Technology, Harbin Institute of Technology, Shenzhen
Zhao Wei, Senior Vice President and CTO of General Data Technology Co., Ltd. (GBase)
Professor Qin Jianbin, Director of the Data Science and Engineering Committee of the Shenzhen Computer Society, Distinguished Professor at Shenzhen University, and Research Scientist at the Shenzhen Institute of Computing Sciences
Tang Bo, Deputy Director of the Data Science and Engineering Committee of the Shenzhen Computer Society and Assistant Professor at the Department of Computer Science and Engineering, Southern University of Science and Technology
Xia Wen, Member of the CCF Technical Committee on Information Storage, Member of the CCF Technical Committee on System Software, Member of the HIT Shenzhen-GBase Cloud Database Research Center, and Associate Professor at the School of Computer Science and Technology, Harbin Institute of Technology, Shenzhen
Zhang Shaoyong, General Manager of the Data Intelligence Product Division, General Data Technology Co., Ltd. (GBase)
Li Shihui, Product Manager of the Data Management Product Division, General Data Technology Co., Ltd. (GBase)
Su Yuanchang, Technical Manager for South China, General Data Technology Co., Ltd. (GBase)
Zhang Yuzhi, General Manager for South China, General Data Technology Co., Ltd. (GBase)
Yang Weiwei, Deputy Director of the HIT Shenzhen-GBase Cloud Database Research Center and Director of the General Engineering Office, GBase
Liu Yang, Deputy Director of the HIT Shenzhen-GBase Cloud Database Research Center, Member of the Data Science and Engineering Committee of the Shenzhen Computer Society, and Assistant Professor at the School of Computer Science and Technology, Harbin Institute of Technology, Shenzhen
Qi Shuhan, Member of the HIT Shenzhen-GBase Cloud Database Research Center and Assistant Professor at the School of Computer Science and Technology, Harbin Institute of Technology, Shenzhen
At the beginning of the conference, Zhao Yijie, Director of the Science and Technology Department at HIT Shenzhen, delivered sincere congratulations on behalf of the university leadership for the phased achievements of the university-industry collaboration. Director Zhao noted that HIT Shenzhen's research funding had already reached RMB 820 million this year, with a forecast of RMB 1 billion for the full year 2021, demonstrating the university's continued leadership in industry-academia-research application in engineering. In the current era, close collaboration between industry and academia is essential. The HIT Shenzhen-GBase Joint Research Center has long been deeply engaged in the database field, jointly committed to building China's homegrown databases as a pillar of national strength. With joint efforts, the center will continue to make new contributions to cultivating talent through university-industry cooperation.
GBase CTO Zhao Wei, representing the company's leadership, greatly affirmed the joint research center's work from research achievements to engineering implementation. He stated that database relational theory and new data management technologies are inseparable from academic research, and the company will continue to support the advancement of the joint research center's projects. At the same time, he put forward higher requirements for the center's research commercialization, hoping that the industry-academia-research cooperation between HIT Shenzhen and GBase could further contribute to the high-quality development of the fundamental information industry and accelerate breakthroughs in core database technologies.
Professor Qin Jianbin, Director of the Data Science and Engineering Committee of the Shenzhen Computer Society, delivered a speech on behalf of the committee. He introduced the committee's research in areas such as relational databases, data governance, privacy computing, and AI+DB. He stated that he would continue to promote strengthened collaboration between universities and enterprises, enhance talent cultivation within industry-academia-research application, and contribute to the overall industry in these areas.
Subsequently, Professor Li Jianzhong, Chief Scientist of the National 973 Program Project and Professor at the School of Computer Science and Technology of Harbin Institute of Technology, delivered a keynote report titled "Computational Theory and Efficient Algorithms for Data Usability."
In the report, Professor Li introduced the basic concepts of data usability from five measurement dimensions: "data consistency," "data accuracy," "data completeness," "data timeliness," and "data uniformity." He mentioned that low data usability can have a huge impact on the physical world. For example, over 25% of critical data among the Global Fortune 1000 companies contains errors, and 4.5 million out of over 36 million social security records in a Chinese city had errors. Furthermore, low data usability renders the reliability of data mining results using AI technology insufficiently guaranteed.
Professor Li stated that no matter how large the data is or how perfect the theories and technologies of big data computing are, if data usability cannot be ensured, big data could produce erroneous or even catastrophic results. Against this backdrop, Professor Li further introduced several research questions on data usability, including the representation mechanisms of data usability, the decidability theory of data usability, data error detection and repair, approximate computing on weakly usable data, and high-quality data acquisition theory and technology.
Addressing these questions, Professor Li's team has published dozens of academic papers in top-tier international journals and conferences, built a data usability management system, and established a prototype system with functions such as data usability representation, usability determination, error detection and repair, approximate computing on weakly usable data, and high-quality data acquisition. This system was applied to 800 million social security records in a Chinese city, improving data accuracy from 75% to 90%, data completeness from 71% to 95%, data consistency from 83% to 100%, and data timeliness from 78% to 87%. In the future, Professor Li will continue exploring new methodologies to reduce computational complexity and expand the theory of approximate computing on weakly usable data.
Assistant Professor Tang Bo, Deputy Director of the Data Science and Engineering Committee of the Shenzhen Computer Society, delivered a keynote report titled "Big Data Query Processing and Acceleration Technologies for Heterogeneous Computing Hardware." In the report, combining the challenges and opportunities that new hardware brings to database systems, he shared his team's research work on database system observability, query optimization, and hardware-conscious execution engines, introduced the specific functions and testing results of system prototypes, and finally, by analyzing the shortcomings of existing technologies, envisioned the challenges and difficulties of architecting a heterogeneous hardware-conscious data management system.
After the coffee break, Associate Professor Xia Wen, a member of the HIT Shenzhen-GBase Cloud Database Research Center, delivered a research outcome report titled "Research on Efficient Delta Compression Technology for GBase Logs."
Professor Xia primarily introduced the center's research foundations and accumulations in areas such as data deduplication and compression, along with the delta compression algorithm developed for GBase log storage. Multiple research outcomes have been published at top-tier international conferences and journals in the data storage and compression field, and several patents have been applied for. He mentioned that during the summer of 2021, two students from the center went to Tianjin to participate in field testing of GBase business scenarios, repeatedly verifying and testing with relevant technical staff. The delta compression algorithm validated in business scenarios achieved a compression rate exceeding 60% for GBase logical logs, and by reducing log write operations, the execution and synchronization time was shortened by approximately 1%. This shows that the center's delta compression algorithm significantly improves compression rates without increasing log storage throughput and latency overhead, still meeting the high-performance business requirements for GBase log storage, ultimately achieving the goal of storing massive GBase data both quickly and efficiently.
This work has now entered the GBase engineering version phase as a first-stage result of the research center. Professor Xia stated that the next phase of work will focus on the massive data storage demands in database backup scenarios, applying the center's data deduplication and compression achievements to further enhance GBase's storage efficiency in this scenario, achieving a higher-quality win-win collaboration.
Subsequently, GBase General Manager Zhang Shaoyong, Manager Li Shihui, and Manager Su Yuanchang respectively reported on "Market Position, Product Advantages, Key Technologies, and Typical Cases of GBase 8a MPP," "GBase 8s V8.8 Safeguarding Core Transactions," and "GBase's Practice in Distributed Transactional Databases."
The distributed analytical data management system GBase 8a is a high-performance, new-generation domestic database product tailored for big data analytics applications, designed to meet the ever-growing demands for data querying, statistics, analysis, mining, and backup in data-intensive industries. It can serve as the foundational database for data warehouse systems, BI systems, and decision support systems. GBase 8a MPP has ranked on the TPC-DS global benchmark with the fewest nodes, the highest single-node computing power, and optimal loading performance. It also passed the largest-scale MPP cluster test in China (4,096 nodes). The big data platform built for the Agricultural Bank of China headquarters was awarded the first prize of the 2017 Bank Technology Development Award by the People's Bank of China. It has deployed over 2,000 nodes, manages more than 20PB of data, and has been running stably for over 2,000 days.
The transactional data management system GBase 8s is an enterprise-grade distributed transactional database independently developed by Tianjin General Data Technology Co., Ltd., featuring maturity, stability, and independent intellectual property rights. It offers high reliability and high availability: ushering in the third generation to safeguard every transaction with high reliability; a two-city, three-center architecture for continuous service assurance; an automated migration tool enabling direct use of PL/SQL to reduce migration costs; an appliance model with a unified management platform to reduce O&M costs; and a full-stack domestic ecosystem that breaks foreign monopolies to reduce procurement costs. In the core system of a regional commercial bank built with this product, backup time was reduced from over two hours to under 20 minutes; the unavailability of table recovery was resolved, and the inability to insert data into large tables was addressed; separate data storage and sharding of large table data across different table spaces improved performance; and configuration optimizations, such as table lock granularity, indexes, and parameters, enhanced concurrency performance.
The distributed transactional data management system GBase 8c is a shared-nothing architecture distributed transactional database cluster featuring high performance, high availability, elastic scalability, and high security. It can be deployed on physical machines, virtual machines, containers, private clouds, and public clouds, providing secure, stable, and reliable data storage and management services for core systems in critical industries, internet business systems, and government and enterprise business systems. It has already achieved a perfect score in the distributed transactional database evaluation by the CAICT (China Academy of Information and Communications Technology).
After the successful completion of expert presentations and technical reports, Professor Wang Xuan, Chair of the Shenzhen Computer Society, Director of the HIT Shenzhen-GBase Cloud Database Research Center, and Dean of the School of Computer Science and Technology at Harbin Institute of Technology, Shenzhen, delivered the closing remarks. Professor Wang expressed that in the over one year since the cloud database research center was established, both the university and the company have leveraged their respective strengths, fully utilizing the center as a window for cooperation and as a pilot and industrialization base for new technology achievements, truly achieving a seamless integration of industry, academia, and research, thereby improving the efficiency of core technology research and industrialization. The joint research center should focus on overcoming bottleneck technologies to build China's homegrown databases as a pillar of national strength, continuing to contribute to the national IT application innovation industry.
In the future, the cloud database research center will continue to rely on the scientific research strength of the School of Computer Science and Technology at HIT Shenzhen to drive the business development of General Data Technology Co., Ltd. (GBase). At the same time, guided by GBase's demands and fully leveraging HIT's talent and technological advantages alongside GBase's leading technology and experience in the database field, it will support and promote the center's research and accumulation in relevant core technologies, using China's homegrown databases as a pillar of national strength to ensure the intrinsic security of China's data assets in the digital era, and continue to delve into research and achieve new successes.
About the HIT Shenzhen-GBase Cloud Database Research Center
The School of Computer Science and Technology at Harbin Institute of Technology, Shenzhen, and Tianjin General Data Technology Co., Ltd. (GBase) established in-depth cooperation and officially launched the "HIT Shenzhen-GBase Cloud Database Research Center" on September 1, 2020. This center helps to foster a partnership between GBase and HIT Shenzhen based on mutual benefit and common development, achieving a close integration of industry, academia, and research. Concurrently, both parties will prioritize long-term collaboration in areas such as cloud database technology, including strategic consulting, technological innovation, new product development, and talent cultivation. Additionally, they will identify and implement joint development of technical projects, promoting the industrialization of project technology.
In its first phase, the research center has conducted research on multiple topics, including intelligent database parameter optimization, columnar data compression for databases, and intelligent database security auditing technology.
(1) Intelligent Database Parameter Optimization Technology
Parameter optimization is a crucial method for database tuning, and the large number of database parameters increases the difficulty of parameter adjustment. In the big data era, faced with constantly expanding data volumes, complex and diverse application scenarios, heterogeneous hardware architectures, and varying levels of user proficiency, traditional manual database tuning struggles to adapt to these new scenarios and changes. Machine learning-based intelligent database configuration technology, with its strong learning capabilities, can effectively enhance the efficiency of database parameter tuning. This direction focuses on research into intelligent database parameter optimization technology, using AI techniques to automatically configure database instances for different business scenarios while balancing performance and cost.
(2) Columnar Data Compression Methods for Databases
In the era of rapidly developing information technology, massive data management technology has become an urgent task for societal informatization. How to effectively store and manage massive data and efficiently support queries on it poses severe challenges to database management systems. The most prominent feature of massive databases is the significant amount of data redundancy, where identical data appears repeatedly in different places. This direction begins with delta compression of GBase database logical logs, compressing before and after images within logs through a lightweight and efficient Ddelta delta compression algorithm to save space occupied by logical logs during large table modifications. This research has now entered the product engineering phase and is expected to save 40% of log storage space.
(3) Intelligent Database Security Auditing Technology
As the core and foundation of information technology for business platforms, databases carry increasingly critical data, making their security ever more important. Database auditing technology can record accessing behaviors on the network in real time and conduct fine-grained audits of database operations. In addition, database auditing can also alert on risky behaviors encountered by databases, such as vulnerability attacks, SQL injection attacks, and high-risk operations. This direction focuses on research into database auditing technology, using a security audit system that meets the requirements of the National Information Security Classified Protection to defend against the growing security threats from insiders. It also employs intelligent and comprehensive data security solutions for vulnerability scanning, preventing data loss, and protecting data privacy.
In the future, the cloud database research center will become a window for cooperation and a pilot and industrialization base for new technology achievements. It will leverage the scientific research strength of the School of Computer Science and Technology at HIT Shenzhen to drive the business development of General Data Technology Co., Ltd. (GBase). At the same time, guided by GBase's demands, it will support and promote the center's research and accumulation in relevant core technologies, using China's homegrown databases as a pillar of national strength to ensure the intrinsic security of China's data assets in the digital era.