Frost & Sullivan releases 'The Development and Industry Applications of Relational Databases in China 2023'

In the Chinese relational database market, local brands have risen to prominence, with the combined share of the top five domestic vendors increasing from 27.1% in 2018 to 55.4% in 2022; conversely, the combined share of the top five international vendors decreased from 57.3% in 2018 to 27.3% in 2022. This includes not only the rapid development of local public cloud database services but also the continuous efforts and high-speed business growth of local traditional database vendors in recent years.

Domestic database products are increasingly being used to replace other existing databases in key industries and enterprise core systems. In the markets for incremental databases such as distributed, cloud-native, and HTAP, the advantages of domestic database brands are becoming more apparent. The market structure of relational databases in China is being reshaped. With the opening up of epidemic control measures, the business operations of database vendors have returned to normal, and enterprise database construction projects are being carried out extensively. 2023 will be a very important year for establishing a new market pattern. After more than 20 years of development, domestic database products' technology has gradually matured, and the domestic database industry is presenting a scene of flourishing diversity and competition.

Domestic database vendors are accelerating their pursuit of international leaders. In terms of basic research, innovative theories are continuously emerging. In terms of technological breakthroughs, significant advancements have been made in several key technology areas. The industry has basically acquired the capability to develop key technologies and is forming a positive competitive edge. The entire database industry is presenting a scene of flourishing diversity and contention: overall, relational databases are dominant, supplemented by non-relational databases, with diverse development paths. Database products such as distributed and cloud-based ones continue to gain popularity. At the same time, domestic database procurement volumes have increased significantly, market scale has grown steadily, and the industry already has the foundation for 'commercial use' in core scenarios of various industries. In 2022, the market share of domestic databases increased from 35% in 2019 to 57%.

Source: Compiled by Frost & Sullivan

The compute-storage separation architecture fully decouples computing and storage resources and achieves resource pooling. Computing nodes become stateless nodes, supporting on-demand allocation; external shared storage can use centralized storage clusters to enhance storage reliability and scalability. Compute-storage separation splits storage and computing resources into independent modules for construction, offering significant advantages in resource utilization, efficient shared storage, flexible deployment across multiple scenarios, etc. Moreover, with the emergence of new technologies such as high-performance disks, NVMe protocols, and RDMA/NoF networks, which provide better technical support for compute-storage separation architecture, making it a choice in the trend of technological evolution. Currently, AWS Aurora, Alibaba Cloud PolarDB, openGauss, and others have all adopted the compute-storage separation architecture to enhance the overall capabilities of their databases.

In scenarios that require simultaneous support for OLTP and OLAP, based on an innovative computing storage framework, HTAP databases are built to ensure real-time analysis while supporting transactions on the same data without the time-consuming ETL process. HTAP databases use a variety of technologies such as transaction processing, analysis processing, data synchronization, query optimization, resource scheduling, etc., to regularly merge incremental data into column storage based on transactional row storage to meet analytical loads. They also combine distributed scheduling technology to achieve parallelization and further accelerate processing. Currently, major vendors are deploying HTAP, such as Alibaba Cloud PolarDB, Tencent Cloud TDSQL, openGauss, etc.

(3) Multimodal databases will become popular

The next mode of the database will develop towards multi-modeling, supporting relational, KV, document, graph, temporal, and other models. Multi-model databases support flexible data storage types, centralizing the storage, querying, and processing of various types of data, which can meet the unified management needs of applications for structured, semi-structured, and unstructured data. Currently, there are several native multi-model databases, such as ArangoDB and OrientDB, as well as traditional relational databases like PostgreSQL and openGauss, which support multiple modalities such as temporal and spatiotemporal through plugin extensions.

The application of domestic databases in several key industries in China, including finance, telecommunications, government, and manufacturing, is significantly increasing, with their market share expanding year by year. As the largest user of the database market in China, the financial industry has seen domestic databases occupy 48.5% of its general system applications. At the same time, the telecommunications industry is also actively adopting advanced database platforms, such as database all-in-one machines, with domestic databases accounting for 45.3% of general system applications.

The government industry ranks second in sales scale in the database market, accounting for 20.7% of the market share. Among general systems, domestic databases account for as high as 53.3%. Manufacturing is at a critical stage of digital transformation, and databases play a central role in this process. In general systems of this industry, the proportion of domestic databases has reached 42.3%.

Domestic customers in 14 key industries are actively innovating and transforming application systems. Among them, the trend is that database applications are gradually penetrating from general systems into core systems, becoming a development focus for these industries.

In terms of database architecture, although centralized databases remain the mainstream, distributed databases are gradually emerging as a new force in the field due to their ability to meet high data processing requirements.

Meanwhile, enterprises are facing a major challenge of how to effectively manage the data flowing through various applications, which has become one of the urgent problems that enterprises need to solve.

Finally, there has been an increasing number of database migration project cases, and the corresponding migration planning theories have become increasingly mature. This reflects that enterprises are gradually recognizing the necessity of migration in terms of databases and have accumulated rich experience in practice. Overall, the database field is facing multi-level changes and development.

With the deepening of industry applications, domestic databases are facing a series of severe challenges. Firstly, the migration and replacement of industry applications have become exceptionally difficult due to their complexity. At the same time, the layout of forces within the industrial ecosystem is relatively scattered, lacking synergies, which has hindered overall development. In addition, industrial risk issues urgently need to be addressed, and the shortage of professional database R&D talents has also become a constraint on the development of domestic databases.

(5) Strengthen the cultivation of R&D talents:To strengthen the cultivation of database talents, it is necessary to combine industry with education and establish an internal training mechanism within enterprises.