Since the 1990s, the computer industry has developed rapidly, shifting from focusing on the development of computer hardware equipment to focusing on data processing and data management. my country's demand for building databases, data warehouses and data mining applications continues to deepen.
In order to ensure data security and reliability, Peking University and Renmin University of China are responsible for developing the core layer of the database management system, and China Software and Service Corporation and Huazhong University of Science and Technology are responsible for research and application development tools. Developed COBASE, a practical relational database management system with independent copyright.
Originated from PBASE
The overall technical level has reached the international advanced level
Under the leadership of Professor Wang Shan, with COBASE as the core and various database architectures as the basis, Renmin University of China has successively developed the teaching database Easybase, the embedded mobile database system "Xiaojinling", and the highly adaptable parallel database system PBASE .
1996 Parallel Database Management System PBASE Appraisal Meeting
As the first prototype system of parallel database management system in China, PBASE has realized a complete high-performance parallel processing scheme with distinct originality, and the overall technical level has reached the international advanced level. It has reached the international leading level in key technologies such as thread-based parallel operation algorithms and parallel packaging.
However, as far as the market is concerned, breaking through the core technology does not mean it can be commercialized. PBASE is still not a product. There is a huge gap in database programming interfaces, including SQL, PLSQL, JDBC, ODBC and other fields.
Birth of KingbaseES
Domestic large-scale general-purpose relational database products
In order to better transform scientific research results into the market, in 1999, under the leadership of Professor Wang Shan, a group of scholars engaged in database research and development at Renmin University founded Renmin University Jincang Company.
Renda Jincang integrates the scientific research achievements of Easybase and "Xiaojinling", inherits the gene of independent innovation of PBASE, and successfully develops a large-scale general-purpose relation The first-generation product of KingbaseES, a large-scale database, also laid a solid foundation for the road of independent innovation of products in the future.
KingbaseES has benchmarked against the top international database products since its birth, and hopes that one day it can surpass its opponents and become a world-class great product!
Great products are not born great
It is "evolved" in actual combat
In the field of basic software, no product designer can predict all scenarios and requirements that may be encountered, and consider all problems in advance.
The real situation is-always encounter problems, and then design reasonable solutions to the problems. Both short-term cost and difficulty must be considered, as well as long-term development. Even a great product like Oracle took years of hard work to get to what it is today.
Therefore, making database products is a slow and continuous process. It is necessary to accumulate product capabilities through continuous iteration in the practice of thousands of industries. At the same time, in the process of iteration, the product must be stable and easy to maintain. Only in this way can the product continue to evolve positively in a "positive cycle" system.
Easier said than done, the truth is simple, but the practice is full of challenges. In order to achieve this goal, Renda Jincang has built a "positive cycle" system of products through the following three aspects during the development of more than 20 years, so that the products of Renda Jincang can continuously evolve and strengthen following changes in market demand.
Product architecture capability: the key to achieving sustainable iteration and avoiding low-level duplication
For database software, with the increase of requirements and the accumulation of functions, the complexity of the code is also increasing. What follows is that the difficulty of code maintenance continues to increase until one day it exceeds the control of human beings. If the code is extremely difficult to maintain, it will be difficult for the product to continue to iterate and develop at this time, and it can only be reinvented.
If the product is to achieve sustainable iteration, the complexity must be controlled within the human controllable range. To achieve this, it is necessary to design an architecture that conforms to the principle of "high cohesion and low coupling", reasonably divide the modules, and at the same time clarify the boundaries of each module to avoid crossing each other and avoid "one hair that affects the whole body" ", control the problem within each module.
To design such an architecture, product architects need to have a deep understanding of upper-level application scenarios and basic theoretical knowledge of databases and rich practical experience, and constantly refactor and optimize the architecture during the iterative process.
Only by designing an excellent architecture that satisfies the principle of "high cohesion and low coupling" can the complexity of the product be controlled within the controllable range of manpower, and the product can be easily maintained. Only when it is easy to maintain, can the product capability be continuously improved.
This is similar to building a 100-story high-rise building. If the structural design is not good, it will start to shake when the building reaches the third floor, so you can only tear it down and start over. If a good structure cannot be designed, it can only be repeated at a low level on three floors. Under such circumstances, the 100-story building will never be built.
Since 1999, the core architects of Renmin University of Jincang have continued to research and invest in product architecture, to learn and absorb the design ideas and architecture design skills of foreign excellent products, and to constantly compare the architecture designs of various foreign database products At the same time, in the actual project combat, we will improve the architecture design according to the needs of different scenarios. In more than 20 years, the product architecture design ability of Renda Jincang has been greatly improved, and it has the architecture design ability to create a world-class product. This also ensures that KingbaseES can continuously improve its capabilities in the market demand and achieve sustainable development!
R&D organizational management capabilities: specialized division of labor, quickly catching up with the top international
When the National People's Congress Jincang was established, foreign database companies had developed for more than 20 years. To some extent, "starting is already behind", so "catch-up" has become Jincang's top priority. To catch up with rivals, large-scale research and development capabilities are required. Only by running faster and producing more than the opponent can it be possible to catch up with the opponent and get the chance to surpass!
Along the way, the R&D system of Renminda Jincang has gradually developed from the initial manual workshop mode of masters and apprentices into a process-based, automated, tool-based R&D management that is oriented to large-scale R&D teams working in parallel based on specialized division of labor. system.
Just as Tesla can produce cars on a large scale and at low cost, database research and development also needs to build a similar system and assist various tools to improve research and development efficiency and surpass rivals in research and development output.
From nothing to something, from zero to one. On the basis of absorbing advanced external experience, Renda Jincang has built a R&D management system with Jincang's characteristics and surrounding supporting tools, which has greatly improved the team's R&D efficiency and output, and ensured that in recent years, in the face of big When the demand for scale hits, the R&D team can still complete tasks with quality and quantity, and solve user problems in the shortest time and at the lowest cost.
The creation of this capability has allowed Dajincang to realize the construction of new product capabilities at a faster speed and at a lower cost. In the process of accumulating over time, the gap with foreign leaders has been continuously narrowed.
Quality Assurance Capabilities: Test Automation Factory is the Strongest Blue Army
For basic software such as databases, stability is the first priority, but because domestic database products are in the process of "catching up", they need to quickly modify the code for iteration, which conflicts with the goal of "stability"—— Because as long as the code is modified, bugs will occur, which is unavoidable.
How to kill bugs and bring stability while iterating quickly has become a problem that must be solved.
After the code of the product is modified, a full regression test is performed on all functions of the product, which becomes the means of final quality control. However, the full regression test brings a huge workload. If it is done manually, a full regression may take several months. The time cost is very high, and it cannot meet the customer's time requirements.
Under such circumstances, building an automated factory for full regression testing becomes the key to solving the problem. The so-called automated factory is that test engineers simulate various operations of accessing database product functions by developing codes, and at the same time simulate various extreme scenarios that may cause failures, and the server automatically runs these test cases to realize all functions of database products in all scenarios. All regressions below.
In more than 20 years, Renminda Jincang has built a strong automated testing team, they are the blue army of Jincang. They test the Red Army of KingbaseES by continuously creating various failure scenarios. Through the hard work of the Blue Army, the products are fully tested before leaving the factory to avoid repeated mistakes. The creation of this core capability ensures that KingbaseES products can maintain good stability while rapidly iterating, and further enhances users' confidence in using KingbaseES.
Go deep into the industry and grow in actual combat
Control every line of code
Sitting at home can never design a good product. Only by going deep into the industry and users can we truly understand the real scenarios and needs of users. Whether the product code is well written or not, and whether the quality is high or not, ultimately needs to be verified and improved in practice.
Starting from the core power dispatching system D5000 of State Grid in 2008, KingbaseES started this journey of penance. Jincang engineers spent a year migrating D5000 from Oracle database to Jincang database. With the migration of more than 5,000 stored procedures in the system, Oracle language compatibility has been greatly improved!
In 2016, the Jincang database began to be applied in Everbright Bank's corporate liabilities and intermediary business systems. In this process, the problem to be solved is not only migration, but also the problem that users can fall back when encountering risks. For this reason, we began to design a dual-track parallel strategy-the Kingbase FlySync product of Jincang was born!
In 2020, China Mobile's Internet settlement system will be launched. As Jincang's first core application in a giant operator, this process made the team feel what is the operator-level requirement and what is "cutover"? How can Jincang's product capabilities be enhanced in terms of details?
In 2022, the train monitoring system of Line 14 in a certain city will be launched. For this purpose, Jincang has designed a dual-network redundancy strategy and a cross-network switching mode, which has further improved Jincang's product capabilities.
Medical care, finance, transportation, energy, national defense... With the deepening of all walks of life, various demands and scenarios have also brought influx, so that the product capabilities of Jincang can be fully verified, and then the continuous optimization of products has been formed. Positive feedback mechanism. The so-called "every line of code can be controlled" is the result of Jincang's real experience in actual combat! As long as the gold warehouse industry is given the opportunity to practice, there will be no problems that cannot be solved.
Follow up, run side by side, and lead the product strategy
last come first come
Become a standard setter in the digital age
"First-class enterprises make standards", in the database industry, the de facto standards of the industry are formulated by foreign advanced enterprises, and domestic database products mostly follow the standards established by foreign databases? Why not deviate from foreign standards and build a new standard system?
What exactly are de facto standards in the database world? De facto standards are essentially solutions formed by foreign database companies when solving customer problems. After long-term verification and polishing, these solutions have gradually become best practices, and then formed a consensus on the cognition of software developers and users, becoming The optimal path with the lowest cost and the best effect.
Therefore, the standard is the consensus on the optimal path formed after various trials and errors and summaries. If you want to surpass your opponent, you need to learn from your opponent first, take a path of following, running together, and finally leading!
Follow along, be compatible in product functions, learn the best practices that have been formed in various foreign databases, reduce the time and cost of trial and error, and quickly shorten the gap with foreign products.
Run side by side, explore best practices for solving problems in new digital scenarios and fields that are not touched by foreign databases, and then form product capabilities and de facto standards.
Leading the way, in China's digital no-man's land, explore and build new capabilities that meet the requirements of digitalization, and finally form a "new standard"!
History of domestic database industry development
Gold warehouse imprint
With more than 40 years of deep cultivation of technology and more than 20 years of industrial development, Renda Jincang has always adhered to the product development strategy of combining original innovation and inclusiveness, and built a positive cycle system that supports the sustainable development of products. "The world-class Chinese database products have laid a solid foundation!
In the current era of major changes unseen in a century, Renminda Jincang has never stopped as a national team in the field of database. No matter how many trials and tribulations he experiences, Jincang will never give up his original intention and mission! KingbaseES will also take advantage of the momentum, move forward courageously, and eventually surpass its opponents, put Chinese database products on the top of the world, and become the world's outstanding database product and service provider!
postscript:
From the transformation of scientific research results to the continuous positive evolution in a "positive cycle" system, Jincang Database has passed the test of the market, industry, and users, faced the gap, tolerated learning, and stood out in the process of practice and evolution for more than 20 years. Riding on the momentum, we have explored a road to the rise of domestic databases. The "Three Independents" series of articles of Renda Jincang will give you an in-depth understanding of the development history of Renda Jincang. Please continue to pay attention!
Contributed by: Product R&D Center
Editor: Wang Jin
Audit: Ri Yao
