Shulou(Shulou.com)06/01 Report--

First, the difficulty of database development. Database research and development is a difficult task comparable to the operating system. The reference is very interested in databases and distribution. what is the learning route? This article talks about a lot of learning processes, and its complexity can be seen academically. However, this is only the beginning of the long march, and there is still a huge gap between academia and products. Database is the basic software, just like the engine of the car, it is the key component of the system, and it has the characteristic of "affecting the whole body". This requires a high degree of maturity in the database. This maturity requires considerable accumulation of technology, product, engineering, support and industry experience. Traditional relational databases, such as Oracle, IBM DB2, etc., have taken decades to reach the current version. Any database from R & D to product is a tough process, it usually takes more than 6 years to get out of the stage of 0 to 1, and then constantly polish in the industry and practical application scenarios, and gradually transition to maturity. We can also go through the history of the open source database. Older, it has been 22 years since the first version of Mysql; Postgres began academic research in 1986 and the first version in 1996. Newer, Cassandra is the first initial release in 2008, releasing open source version 1.0 at the end of 11 years; now it is 3.11 version; MongoDB: development started in 2007 and current version 3.4; VoltDB started in 2009 and began to meet people seven years later. In addition, financial-grade database products are faced with large enterprises in leading industries such as banking, securities, insurance and so on, which put forward higher requirements for product complexity and maturity. As a result, its life cycle is longer. Second, the research and development of the difficult products of database production is a process of continuous progress, accumulation and polishing. Or take the automobile industry as an example, it is also the engine, what is the difference between the engine of a sports car and that of a family car? The reason why F1 cars last forever is that manufacturers know that products need to be constantly improved, constantly challenged, and constantly cope with the changing demand scenarios. Take product testing as an example. After 6 years of development, Giant Sequoia SequoiaDB has more than 12000 test cases, with automated test coverage exceeding 99%. The test development team and product development team staff at 1:1, even more, the test development level requirements are higher than most R & D engineers. Each small iteration basically runs for a week, and the test server in the computer room never stops. But so what? it's nothing compared to Oracle and DB2. Then take the demand to the product transformation process. The customer's demand of the application product is easy to understand, and the database is a basic product, and the transformation of the customer's one-sentence demand into the language of the technical product is a huge gap, and the needs of different users are intertwined. How to turn the requirements to define the functions of product development, "this is a problem". All of this cannot be achieved without day and night on-site support. In addition, how to strictly control the core code and control the product development route is also very important, after all, it is a product, can not rely on technical support pile manpower. I think the phrase "win in detail, lose in pattern" applies to every product. Third, the shortcomings and advantages of the development of Chinese database, first of all, the biggest is the inheritance of technology and genes. Take Father Michael Stonebraker as a case, he experienced Ingres,Postgres,Informix,C-Store, H-Store and then VoltDB in his life. The same is true of other databases, all of which have been passed on for decades. China has only begun to develop databases in recent years, mainly from open source products such as Mysql and Postgres; some acquisitions have been introduced into Informix; innovation has never been instantaneous, and there is nothing wrong with standing on the shoulders of giants. It just takes a process of accumulation. We are also derived from DB2 genes in North America, and it is only in China that we have combined Huawei genes to take root. Speaking of advantages, China is a huge market soil, considering the number of users in China, the frequency of mobile phone use, the diversity of application scenarios, and the confidence that the products polished in the Chinese environment can be used anywhere, especially when we discuss the development direction of the database with Gartner. Now is the golden period of database development in China, and I hope all colleagues will work together to promote the development of independent databases.

I heard that there is another dark horse company called Wanli Open Source, which has independently developed databases and operating systems for more than 20 years. GreatDB, an open source distributed database, has supported the storage and processing of massive data by operators. Since China Mobile selected Wanli Open Source distributed Cluster Database for the first time on big data's platform in a certain province of China Mobile in 2013, Wanli Open Source has been applied in many business scenarios in the telecommunications industry. In the future, with the development of 5G and Internet of things technology and the further promotion of localization, Wanli open source distributed database will replace foreign database in more business scenarios!

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.