Last night, Mr. Chen from vn.py shared an article titled "Understanding Various Systems of Algorithmic Trading with One Image!", which provides an overview of algorithmic trading service systems and is well organized.
(Image source: VNPY Official Account)
Technology vendors should also include three major quantitative platforms - JoinQuant, RiceQuant, and Youyu.
High-speed trading platforms of various securities firms:#
First Securities is currently considered a characteristic securities firm. In 2018, it adopted JoinQuant's product and officially launched the "Yichuang Jukuan" quantitative platform, which integrates real-time trading and provides customers with a programmatic trading platform similar to CITIC XTP.
In 2020, at least 11 securities firms have updated or launched high-speed trading systems, platforms, or services.
In April 2020, Huachuang Securities launched the TradeX high-speed trading system. TradeX is a one-stop intelligent trading solution independently developed by Huachuang Securities. It adopts an All-In-One design concept, providing convenient interface services, high-speed market services, intelligent algorithm services, smooth trading experience, and customizable personalized services. TradeX supports various trading varieties such as stocks, ETFs, bonds, futures, Hong Kong stock connect, margin trading, and options, covering all trading types. It supports multiple workflows such as independent trader mode, investment manager trader instruction mode, and investment advisor mode. It supports single-product and multi-product joint risk control. Risk control checks are processed in parallel with ultra-low latency. It provides a 24/7 simulation testing environment with professional support available online.
In May 2020, Guolian Securities' new generation high-speed trading system (UST) passed the certification test in the production environment conducted by China Software Evaluation Center. The core trading penetration time of UST is as low as 5 microseconds, and the end-to-end order submission time is as low as 9 microseconds.
On July 1, 2020, Dongwu Securities and Vertex Software jointly held a press conference to officially announce the launch of the new generation core trading system A5. It is worth noting that A5 is a centralized trading counter, not a specialized institutional trading counter. It is claimed to be the first fully in-memory, fully business centralized trading system in the industry, breaking the dependence of brokerage core trading systems on foreign commercial databases and specific commercial hardware platforms for the past 20 years, and declaring that China's brokerage IT core technology has entered a new stage of comprehensive self-control. The system was fully launched on June 29. Compared with the previous generation trading system, the average latency of the new system has been improved from the original 10 milliseconds to within 1 millisecond. The single-node business concurrency has been increased from 30,000 transactions per second to at least 100,000 transactions per second, supporting tens of millions of transaction orders.
In July 2020, Western Securities launched the HTS high-speed trading system, claiming that the system adopts a new architecture, communication services, and business processes, and the internal consumption efficiency is 300% higher than the previous high-speed trading system, reaching a level of less than 100 microseconds.
In September 2020, Yuekai Securities officially released the high-speed trading system, claiming that the upstream full-chain latency of its trading is as low as 6.5 microseconds, and the average delay of market decoding is 1.12 microseconds. Yuekai Securities stated that the high-speed trading system has three highlights: first, it has created the industry's most extreme speed through software and hardware optimization and deployment adjacent to the exchange, and has been certified by China Software Center; second, Yuekai Securities has a professional business and technical team to serve customers and meet their needs; third, Yuekai Securities has established the Smart Finance Joint Laboratory and will continue to invest in trading technology to maintain its leading position in this field.
On October 18, 2020, Shenwan Hongyuan Securities released the SWHY Matrix high-speed trading platform. With the advantages of multi-node deployment, multi-counter complementarity, and multi-functional customization, SWHY Matrix provides professional trading solutions for private fund managers at different growth stages, featuring "extreme speed, personalized customization, and full business support." It provides trading services for various business varieties such as stocks, bonds, fund ETFs, Hong Kong stock connect, block trades, margin trading, options, financial futures, and commodity futures.
In October 2020, Wanhe Securities adopted a new generation distributed core trading platform and launched the E-ONE high-speed trading system, leading high-frequency quantitative trading into a new stage of microseconds. Under the high-speed trading mode, the measured upstream full-path latency can be as low as 13 microseconds for 20,000 transactions per second, and the end-to-end latency of market data is 1.2 microseconds. The advantages of E-ONE also lie in its security and reliability. It adopts a distributed architecture, can be quickly expanded, and can handle 10 million orders per day without any pressure, with zero data loss.
In November 2020, Huatai Securities released the "FPGA Securities Trading Counter System" to solve the problems of high latency, low concurrency, and unstable performance in traditional counters for high-frequency investors. It has supported trading of stocks, funds, ETFs, bonds, and other assets. Prior to this, Huatai Securities had launched FPGA market data. This time, it released the FPGA securities trading counter system, leading securities trading from market data to offering into a comprehensive high-speed era.
On November 18, 2020, Hongta Securities issued a tender notice for the construction of the company's high-speed trading investment platform software project, aiming to optimize the company's high-speed trading investment platform construction through bidding. According to industry sources, the bid was eventually won by Huarui Financial Technology.
As early as April 3, 2020, Guosen Securities also issued a tender notice to build the company's new generation low-latency high-speed trading system. According to public information, the bid was eventually won by Huarui Financial Technology for 2.6 million yuan. According to public information, Guosen Securities' trading system construction has been quite effective. Guosen Securities' new generation low-latency trading solution includes four modules: FPGA market data system, new generation distributed counter system, high-speed network environment, and dual-center offering for Shanghai and Shenzhen. It provides end-to-end integrated low-latency trading services for high-speed trading investors. Guosen Securities' previous high-speed trading system, Guosen TradeStation platform, was jointly created by Guosen Securities and TradeStation from the United States. It was designed specifically for active trading clients in China and supports stock, futures, and options trading, with built-in powerful order tools, comprehensive strategy backtesting, and automated trading functions to meet the trading needs of multiple markets, varieties, and cycles.
In June 2020, Guosen TradeStation platform released the new version NeXT U17, upgrading and optimizing the trading functions.
In December 2020, Guotai Junan Securities officially released the STS intelligent trading service at its first Financial Technology Conference, creating four featured services: high-speed trading, customized strategies, exclusive margin trading, and exclusive investment research. In terms of high-speed trading services, it meets customers' ultimate pursuit of speed with ultra-low latency, striving to build an industry-leading quantitative trading infrastructure with an overall network latency of less than 5 microseconds. Based on supercomputing power and tick-level data, it provides trading strategy production services with over 3,000 intelligent strategy factors, and launches intelligent order services.
On January 4, 2021, GF Securities released the self-developed new generation ultra-high-speed quantitative strategy platform GFQuantGenius 1.2 version (referred to as "GFQG"), which is the first securities firm in the industry to use the latest version of AMI to create an ultra-high-speed quantitative strategy platform. As a unified platform, GFQG adopts a loosely coupled design pattern. GFQG divides the system into three layers: the underlying layer, the business logic layer, and the strategy layer according to functionality. The underlying layer can be subdivided into the core layer, the driver layer, and the access layer, which include functions such as shared memory, message bus, database, interface, and programming entry. The business logic layer encapsulates the underlying functions into various microservices for the strategy layer to call according to business varieties. The strategy layer is the core business asset, providing various business component functions, trading tools, and trading strategies. By transforming and improving the core layer, the driver layer, and the access layer with cutting-edge technology, adapting the business logic layer, and fine-tuning the strategy layer, the GFQG platform can achieve rapid updates and upgrades, and meet higher requirements for quantitative development. GF Securities stated that the internal strategy latency of GFQG has been reduced to within 100 microseconds, compared with the previous 1,700 microseconds, reducing the latency to 1/17 of the original level. The performance of volatility surface calculation has been improved by 187 times, supporting 40,000 transactions per second and a total of 400 million transactions. At the same time, in order to improve the applicability of the platform, GFQG has designed and implemented a new interface similar to the CTP futures counter, which is easy to learn, efficient to run, and can cover all-business quantitative platforms for Shanghai and Shenzhen exchanges, financial futures, commodity futures, interbank markets, and overseas derivative investments.
Glossary:#
- Counter System: A server system used by securities firms to provide services for customer order submission, transaction records, and fund position settlement.
- High-speed Counter: Specifically optimized for algorithmic trading, providing faster trading speed, but supporting fewer types of businesses, mainly focusing on buying and selling stocks (spot and margin trading).
- Centralized Counter: Targeting ordinary investors (mainly retail investors), providing the most complete support for various types of businesses (including stock trading, fund subscription and redemption, IPO, etc.), but relatively slower speed.
- Market Data: A server system that provides real-time market data push (including L1 order book, L2 order book, tick-level transaction orders, composite order book) and can be divided into TCP subscription, TCP full push, UDP multicast software decoding, and UDP multicast FPGA decoding in terms of data reception mode.
- Trading Gateway: A gateway system that interfaces with various types of backend counter systems and provides standardized API interfaces to external clients. It can reduce client development costs and provide additional trading behavior risk control.
- Algorithmic Bus: Provides intelligent algorithmic trading services (the two largest being Kafang and Jinnar) in the form of API similar to trading API (similar to various built-in algorithms provided by overseas IB API), providing more powerful intelligent trading algorithm orders (such as VWAP, TWAP, Iceberg, etc.) than regular limit and market orders.
- Quantitative Platform: Interfaces with all the backend systems mentioned above (via API) and is a front-end software used directly by traders every day. Most of them provide UI graphical interfaces, and users can develop strategies, run historical backtests, and execute automated trading on them.