No industry needs long-termism more than databases, and on the road to practicing long-termism, the domestic database manufacturer Zhonghua cannot be ignored.
Recently, the Gartner Peer Insights "Voice of the Customer for Cloud Database Management Systems, 2023" report was released, showing that HUAWEI CLOUD has become the only cloud vendor in the world to be "Customer's Choice" for cloud database management systems, with a customer satisfaction and recommendation rate as high as 98%.
Prior to this, at the Huawei Global Smart Finance Summit in 2023, Huawei Cloud officially released a new generation of distributed database GaussDB, with software and hardware collaboration and full-stack independence, and handed over the answer sheet of the company's research and development database for 20 years.
Throughout the history of database development at home and abroad, it is indeed not easy to do a good job in database. It not only requires strong technical strength, but also needs to build a sound business ecology. And compared to the former, the latter may be the key to the success of related manufacturers.
From the perspective of commercial substitution, compared with Oracle and other companies that have accumulated more than 45 years in the database field, although domestic database manufacturers have paid huge time costs and capital investment, they still seem to be in a catching-up stage.
And HUAWEI CLOUD continues to deepen the database field and has become the "customer's choice". Does it mean that the real localization replacement has come?
Image source: Internet
60 years of database, born to business prosperity
Chips, operating systems, and databases are collectively known as the three major infrastructures in the technology field.
Compared with the former two, the database is not the "protagonist" in the daily news. But in fact, almost all business systems require databases for organization, storage, management and calling. Therefore, the database is also called "the jewel in the crown of basic software".
In 1964, General Electric Company developed the world's first database system - IDS. In the 60 years since then, the United States has started round after round of waves of database technology. To this day, the United States still occupies a leading position in the database industry.
In order to enhance competitiveness and get rid of dependence on foreign technology and products, domestic database manufacturers are also continuing to promote database research and development, so as to realize localization replacement as soon as possible. In 2009, Alibaba took the lead in proposing the concept of "going to IOE". The so-called "de-IOE" refers to excluding IBM's minicomputer, Oracle database and EMC high-end storage in the IT infrastructure, and replacing it with a system developed on the basis of open source software.
However, more than ten years have passed, and domestic databases still do not occupy a dominant position. According to public data, foreign manufacturers still occupy more than 50% of China's database market share, and the status of giants such as Microsoft, Amazon, and Oracle is difficult to shake.
Insufficient technical strength may be one reason. According to the data released by the Institute of Information and Communications Technology, as of June 2022, there are 363 global database companies, including 145 in the United States and 116 in China. Among these enterprises in China, only three companies have a scale of more than 1,000 employees, while Oracle and other American database manufacturers have reached tens of thousands of people, reflecting a large talent pool gap.
According to public information, the proportion of open source and commercial databases in the United States is 48% and 52%, and that in my country is 83.4% and 16.6%. Great room for improvement.
Compared with technical advantages, the commercial advantages of database ecology of foreign manufacturers may be more difficult to surpass. For example, Oracle, which occupies a major market share, completed its own database ecological construction early. In 2009, Oracle launched the Partner Program (OPN), and the number of partners worldwide has already exceeded 25,000.
Image source: Oracle, OPN official website
This also means that third-party products and services can basically solve the abnormal problems of most Oracle databases. In contrast, at present, many domestic database software still rely on the technical support of the original manufacturers, and it is difficult to form a complete business ecosystem, which has also become a major obstacle to the realization of localization replacement.
But domestic databases are not completely without opportunities. When the benefits of migration are far greater than the cost of migration, product substitution will occur naturally.
In 2012, Amazon launched DynamoDB, a distributed storage system, based on the actual needs of its own business. After that, distributed cloud databases began to develop rapidly, constantly seizing important positions of traditional databases. The seemingly impregnable "Maginot Line" of traditional databases is encountering the "blitzkrieg" of distributed cloud databases.
Source: Gartner Magic Quadrant for Cloud Database Management Systems
Why did Amazon's DynamoDB succeed and quickly become the leader in the database industry? This is mainly due to the new business scenario of e-commerce. In this business scenario, Amazon's DynamoDB shows strong migration benefits, so more and more companies choose the "Amazon solution".
It can be understood that higher migration benefits actually come from the evolution of business scenarios. Throughout the history of the database, companies such as General Electric, IBM, Honeywell, Amazon, and Alibaba have further developed database technology because of new business scenarios.
For database manufacturers such as Huawei, if they want their own technical routes to be recognized, they must find new business scenarios for the future. This new business landscape includes more cutting-edge technology concepts, better performance processing, and a more complete business ecology.
20 years of GaussDB, from technical feasibility to ecological construction
In 2001, in order to support its own telecommunications business, Huawei began to develop the data storage component DopraDB. Ten years later, Huawei established the Gauss laboratory, and DopraDB also became the GMDB V1 series of GaussDB. Then, after another ten years or so, GaussDB has now completed the comprehensive replacement of more than 600 sets of databases in Huawei, and has built more than 6,000 distributed database nodes, with a data volume of up to 6PB. , successfully walked from behind the scenes to the front of the stage.
The road to the front of the stage is somewhat difficult, and creating technological advantages is only the first step in the long march.
The so-called creation of technical advantages can essentially be disassembled into two things: improving database performance and ensuring database security. As the only self-developed database in China that achieves both software and hardware collaboration and full-stack innovation, GaussDB has achieved new breakthroughs in performance and security.
For example, in terms of performance, in the actual application of China Merchants Bank, the jitter rate of GaussDB is 85.6% lower than that of other manufacturers; in cooperation with Postal Savings Bank, Postal Savings Bank opened 650 million users to Huawei for building distributed For the core system, compared with the original system, the efficiency of GaussDB has increased by 40%. In terms of troubleshooting, in 2023, GaussDB will launch the same-city dual-cluster RPO=0 solution in ICBC’s core credit system. Compared with the original Oracle database dual-cluster solution, which took 15-30 minutes to repair faults, GaussDB only takes 2 minutes.
Source: HUAWEI CLOUD GaussDB conference
However, in the database field, technological leadership does not mean commercial success. It can be said that technological advantage is the lower limit of commercial success, and ecological construction is the upper limit of commercial success, and it is also a shackle that domestic database manufacturers urgently need to break through. In the process of building a business ecosystem, it is very important to determine which industry to start with.
Judging from the current situation, the financial industry has become a major field targeted by domestic database manufacturers.
Compared with other industries, the financial industry has the characteristics of large data volume, high performance requirements, and low fault tolerance. At the same time, the industry's demand for localized database replacement is also extremely strong. Judging from the bidding information of the domestic database in 2022, 46% of the purchasing units are concentrated in the financial industry, followed by the government. Therefore, the financial industry can be said to be a "broad world" for mining commercial value, and database vendors who can gnaw on this "hard bone" will be more comfortable in the face of other commercial applications.
Based on this, like OceanBase and Tencent TDSQL, GaussDB regards financial scenarios as a key commercial application. At present, GaussDB has been widely used in the financial industry to help the core business operation of China Construction Bank, Agricultural Bank of China and other state-owned banks.
But financial scenarios alone are far from enough. To penetrate into more business scenarios, database vendors must establish contact with independent software developers to form a comprehensive industrial collaboration connecting upstream and downstream. If the goal of commercial applications is to solve specific problems in specific industries, then the ultimate goal of industrial collaboration is to establish more universal system solutions.
Faced with the first-mover advantage of leaders such as Oracle, open source may be the most important tool for latecomers to integrate industrial resources. Beginning in 2022, domestic database vendors have begun to launch their own open source products, such as X-Paxos, an open source project of Alibaba Cloud Polar-X, CeresDB 0.2.0, an open source project of CeresDB of Ant Group, and OpenGemini, an open source project of GaussDB of Huawei Cloud.
However, open source is a relatively passive industry collaboration, and a more active approach is to establish cooperative relationships with upstream and downstream. Since the database software is not an isolated system, which includes solutions such as database migration, backup and recovery, disaster recovery, one-stop operation and maintenance, database consulting services, and database training, relevant manufacturers have been seeking more opportunities for cooperation in recent years. To open up all links in the business chain.
For example, OceanBase, which was independent from Ali, has successively established a product adaptation certification system with basic hardware manufacturers and software service manufacturers, and Huawei has also joined hands with 17 software partner companies and 7 service partner companies to jointly launch the ecological development of the GaussDB database financial industry. plan.
But as mentioned above, the sign of successful ecological construction is that the benefits of migration are higher than the cost of migration. In this regard, can Huawei repeat the success of Amazon DynamoDB? Combining product functions and industry needs, there are still certain possibilities.
It is understood that Huawei Cloud GaussDB is easy to deploy and migrate, which can reduce storage costs by 50% and increase overall resource utilization by more than 4 times. At the same time, by providing a one-stop solution for structure + data migration, it can meet the needs of related industries to the greatest extent. needs, and is expected to use application mode replication to move towards more scenarios.
Of course, the advent of Huawei Cloud GaussDB does not mean that the localization of the database has been replaced. It should be noted that what domestic database manufacturers are facing is a "blue ocean". In 2018, "Xinchuang" was included in the national strategy, and a "2+8+N" development system was put forward. Among them, "N" refers to thousands of industries in the domestic market. In addition, the data shows that it is estimated that by 2025, the scale of my country's database market will reach 68.8 billion yuan, with a CAGR of 23.4%.
In this context, the release of Huawei Cloud GaussDB will naturally not be the end of the domestic database, but may be a brand new beginning. In the future, with the deepening of domestic substitution, domestic databases will have the opportunity to play their value in more industries.
By: Tim Ze Tyler
Source: Songuo Finance