As a large national integrated securities company, Guosen Securities has 57 branches and 185 business offices across 118 cities and regions in China. According to data released by Securities Association of China, in recent years, Guosen Securities has ranked among the top in the industry in terms of core indicators such as total assets, net assets, net capital, revenue, and net profit.
Industry data shows that the standard requirement on the performance of a traditional brokerage trading system is to ensure a peak of about 10,000 agency transactions per second. As one of the most data-intensive industries, the securities industry has accumulated a huge amount of high-value data and, as business continuously develops, at least one terabyte of data is generated per day. Meanwhile, the demand for high concurrency and low latency imposes higher requirements on the IT system and database supporting the business.
In recent years, the process of offline services in the securities industry going online has accelerated, involving diversified use of internet media such as online marketing and customer acquisition, AI-based one-way intelligent account opening, online handling of off-site business, live streaming, mini programs, and short videos. Meanwhile, as the securities market has been booming in recent years, the number and concurrency of securities users have increased significantly, thus putting forward higher requirements on the IT system and database supporting the business. As securities companies enter the stage of comprehensive innovation and development, the securities business is increasingly diverse and the business process becomes ever more complex. The centralized database architecture from foreign providers is out of its depth in meeting the needs of new business, new regulatory requirements, and the needs for efficient internal monitoring and management in future years.
As a data-intensive industry, the securities industry has accumulated massive high-value data and currently a huge amount of new data is generated every day. In this context, the shortcomings of foreign traditional centralized databases that are mainly used in the industry gradually become apparent. For example, the centralized database architecture lacks compute-storage separation, auto scaling, and cross-IDC high availability. The traditional centralized databases also face a capacity bottleneck and have to rely on vertical scaling. It is difficult to make the scaling process transparent or imperceptible to the business, which incurs high costs. Moreover, the traditional centralized databases cannot meet the goals of independent control and localization.
However, rapid technological development in the database field in recent years has brought about databases with industry representation, such as cloud-native databases, NewSQL databases, and distributed databases. Based on the understanding of the technological development trend and new technology products in the database field, coupled with the situation of the securities industry and its current IT system, Guosen Securities believes that homegrown distributed databases can solve its current dilemma and meet the needs of the business scenarios in the securities industry.
Tencent Cloud’s database service TDSQL has been successfully implemented in Guosen Securities’s new business system with the cooperation of the strategic partner Shenzhen HuaYun ZhongSheng Technology Co., Ltd. The new system integrates TDSQL, NeoKylin, and Hygon servers and realizes full-stack localization.
The new system carries Guosen Securities’s over-the-counter (OTC) trading counter system, anti-money laundering system, and HR management system. Since its launch and operation, the new system has supported a transaction scale of 100 million yuan per day and can still guarantee smooth data operation during traffic peaks such as market opening and closing, ranking high among brokers. In the future, Guosen Securities will have more business systems running on Tencent Cloud’s database service TDSQL.
Tencent Cloud’s database service TDSQL provides finance-grade high availability of automatic disaster recovery within seconds in case of a traffic surge. This, combined with tested ultra-high concurrency tens of times better than the customer's demand, means that TDSQL can help a broker significantly break through the traditional performance requirements for brokerage IT systems and serve more than 100 million stockholders simultaneously to deal with various extreme trading scenarios with high efficiency.
At the same time, TDSQL’s hybrid transaction/analytical processing (HTAP) capability helps securities systems implement high-performance real-time online analysis in the digital era, providing fundamental technical support for digital transformation of business operations.
In terms of cluster deployment capability, Tencent Cloud’s database service supports command line, container, and GUI-based installation and provides a complete process of installation, testing, acceptance, and training, with experience of serving more than 1,800 customers in installation and deployment. In addition, the database service supports 2-region-3-DC and multi-region-multi-DC deployment.
Guosen Securities found that the anti-money laundering system based on TDSQL for PostgreSQL has improved significantly in terms of performance, scalability, high availability, disaster recovery, Ops, and cost reduction in the context of massive data. The distributed database TDSQL for PostgreSQL can generate application value for securities business scenarios.
Distributed databases can execute compute and storage tasks that cannot be done by a single server, and the distributed architecture can enhance the overall reliability and throughput of the system. In addition, there are many successful cases in financial industries such as banking, insurance, and securities. In one of these cases, for example, the stability and reliability of the core system of a bank have been verified to be able to meet the requirements for finance-grade databases. The bank’s anti-money laundering system has migrated to an HTAP database of TDSQL for PostgreSQL, which proves that homegrown distributed databases represented by TDSQL for PostgreSQL have the ability to supplant existing centralized databases for specific business types of securities systems.
On the whole, TDSQL can replace Guosen Securities’s original business system in certain scenarios in the years to come. With continuous update and optimization of the service and technological development, TDSQL can serve for more and more business scenarios for Guosen Securities.