According to: The current popular multi-relational database, we should also understand their development history while learning technology, which can play a key guidance on the direction of our future technologies, perhaps You can also find something more important than technology from this article. Before the start of the article, we must first introduce some people who have developed a decisive role in the relational database. Although I am a historical materialist, I can't change my feelings of these predecessors. -Edgar frank Codd: The father of the relational database. - Don Chamberlin: The father of SQL. One of the creators of the SQL language is also one of the creators of the XQuery language. - Donald J. Haderle: DB2 father, IBM CTO, vice president. --Larry Ellison: Oracle founders, the world's second largest software company CEO. - Mark Hoffman: Sybase founder. --Bob Epstein: Sybase founder. - M & I: PowerSoft Corporation, after Sybase. In 1970 IBM researcher Edgar Frank Codd published the industry's first papers on the relationship database theory "a relational model of data for larding Data Banks", the concept of relationship model is proposed for the first time. Later, CODD has been published in many articles, laid the foundation of the relational database. The relationship model has a strict mathematical foundation, the abstract level is relatively high, and it is simple and clear, it is easy to understand and use. However, some people think that the relationship model is an idealized data model. It is used to implement DBMS is unrealistic. In particular, it is difficult to accept the performance of relational databases, more people depend on the serious threat of standardized work in progress. . In order to promote the understanding of the problem, 1974 ACM led a seminar, and a debate between CODD and BachMan-based support and opposition relationship databases were launched. This famous debate promoted the development of the relational database, which eventually became the mainstream of modern database products. In 1974, IBM's Don Chamberlin and Ray Boyce showed the mathematical definition of the 12 guidelines of the CODD relational database, and the milestone proposed a SQL (Structup Query Language) language. The SQL language features include query, manipulation, definition, and control. It is a comprehensive, universal relational database language, and it is a highly non-processed language that does not need to do anything else. SQL integration implements all the operations in the database life cycle. From the date of the self-production, the SQL language has become a test stone of the inspection relational database, and each change in SQL language standards guides the development direction of relational database products. While making progress in SQL language, IBM Research Center began to start SYSTEMR project in 1973. Its goal is to demonstrate a full-featured relationship DBMS feasibility. The project ended in 1979 and completed the first DBMS that implemented SQL. In 1986, ANSI put SQL as a US standard for relational database languages, and the same year announced standard SQL text. There are 3 versions of the SQL standard. Basic SQL definition is ANSIX3135-89, "Database Language - SQL with Integrity Enhancement" [ANS89], is generally called SQL-89. SQL-89 defines mode definitions, data operations, and transaction processing. SQL-89 and subsequent ANSIX3168-1989, "DatabaseLanguage-EmbeddedSQL" constitutes the first generation SQL standard.
ANSIX3135-1992 [ANS92] describes an enhanced SQL, which is now called SQL-92 standard. SQL-92 includes mode operation, dynamic creation, and SQL statements, dynamic execution, network environment support, and so on. After completing the SQL-92 standard, ANSI and ISO have started working with SQL3 standards. The main feature of SQL3 is support for abstract data types, providing standards for the new generation of object relational databases. ◆ After IBM DB2: SYSTEM R is completed, the group led by Donald J. Haderle (DB2 Father) completed the prototype on the DB2 for VSE / VM between 1982 and 1983. At the end of 1983, DB2 FOR MVS V1 was officially released, which marked the beginning of DB2 product, and marked the founding of DB2 brands. Frankly, DB2 is just like a baby in a born - the development is not complete, delicate, and can't get wind and blow. To solve this problem, Haderle has established DBTI (DB2 Technology Institution, DB2 Technology Research Institute) in 1984. This is a department that has integrated research, development, market, is committed to developing DB2. The performance, reliability, function and throughput of DB2 have greatly improved, and it is already able to adapt to the needs of various critical services. In 1988, the excellent processing capacity and stability of DB2 for MVS were widely recognized by the industry. DB2 finally became a swan from ugly duckles. 1987: IBM released an OS / 2 V1.0 extension with relational database capabilities, which extends the relational database processing capability to the microcomputer system for the first time. This is also the prototype of DB2 for OS / 2, UNIX and window. 1988: IBM releases SQL / 400, providing SQL support for the AS / 400 server integrated with relational database management systems. 1988: IDUG (International DB2 User Organization) organization was established. This marks that the user of DB2 has formed considerable scale. 1989: IBM defines the CommON SQL and IBM Distributed Relational Database Architecture (DRDA) and implemented on all IBM's relational database management systems. 1993: IBM released DB2 for OS / 2 V1 (DB2 for OS / 2 can be disconnected to DB2 / 2) and DB2 for RS / 6000 V1 (DB2 for RS / 6000 can be short-written as DB2 / 6000), this is DB2 The first time I appeared on the Intel and UNIX platforms. 1994: IBM released the DB2 parallel version of RS / 6000 SP2, DB2 has been an extensible architecture that adapts to large data warehouses and complex query tasks. 1994: IBM extends DB2 Common Server to HP-UX and Sun Solaris. DB2 began supporting UNIX platforms developed by other companies. ◆ Oracle: In June 1977, Larry Ellison has funded $ 2,000 in $ 2,000. It has established software Development. RelationalSoftware, Ellison has 60% of the shares, ie later Oracle. In 1976, Dr. IBM Researchers published a milestone "R System: Database Relationship theory", introduced the relational database theory and query language SQL, Ellison read this article very carefully, and was shocked by its content, this is The first time someone manages data information with a comprehensive and consistent program. The author CODD has published a relational database theory almost ten years ago, and developing prototypes in IBM research institutions. This project is the R system. The language of accessing data tables is SQL. The article describes the research results of his decade and how After the implementation, Ellison reads, keen realizes that the commercial software system can be developed on this research.
At that time, most people think that the relationship database will not have business value, because the speed is too slow, it is impossible to meet the processing of large-scale data or large user access data, the relationship database is very beautiful and easy to use, but insufficient is too simple to achieve speed Too slow. Ellison thinks this is their chance: they decided to develop universal commercial database system Oracle, which comes from the names they have done to the Central Intelligence Agency. However, it is not only their exclusive in action, Berkeley University has also begun to develop relational database system Ingres. The following year completed oracle1, developed with assembly language, based on the RSX operating system, running on the PDP-11 mini machine of 128KB memory. But this product is not officially released. In 1980, Oracle officially released Oracle2 based on the VAX / VMS system. Two years later, they released oracle3, mainly developed with C language, with transactional functions. In 1983, RelationalSoftware was renamed Oracle Company. The following year, they launched Oracle4, expanded data consistency support, and began supporting a wider platform. Oracle's main opponent is Ingres. The sales of Oracle in 1984 was $ 12.7 million. Ingres was 9 million US dollars. In 1985, the sales of two were more than doubled, but INGRES grew faster, if this is developed, INGRES It will surpass the opponent, but at this time, the blue giant IBM helped Ellison. In 1985 IBM released the relational database DB2, using the different data query language SQL, INGRES is quel, Ellison seizes the market Opportunity, to promote Oracle and IBM compatibility, from 1985 to 1990, although Ingres sales increased above 50% per year, Oracle ran in front, more than 100% annual growth rate, SQL in 1986 Become an official standard, INGRES's boss can't believe Ellison's luck, but this is the savvy of Ellison, "following the blue giant, never failed." IBM established SQL standard is the key to Oracle success, but The current war in the database is far. Sybase and Informix have also joined the ranks of competition. IBM did not plan development, why the blue giants gave up this value of hundreds of billions of products, there are many reasons: IBM researchers are academic origin, Their most interesting is theory, rather than the product of the market, from academic, research results should be open, published papers and speeches make them famous, why not? There is also a very important reason is that IBM has a market-selling hierarchical database product IMS, launching a competitive product will affect IMS sales and personnel work until 1985 IBM released relational database DB2, Ellision There is already a millionaire. Ellison used IBM to select Microsoft MS-DOS as the IBM-PC's operating system ratio: "The most serious mistake in the history of world business, worth more than 100 billion US dollars.", IBM published R system papers, and no The error that quickly launched a relational database product may only be, Oracle's market value reached 28 billion in 1996. The 1986 Oracle5 realizes the true Client / Server structure, starting to support the VAX platform cluster and become the first database product with distributed features. In 1988, Oracle issued Oracle 6.1992. ◆ SYBASE: 1984 Sybase founded by Mark Hoffman and Bob Epstein. The company name "Sybase" is taken from "System" and "Database".
One of the founders of Sybase is Bob Epstein is the main designer of the Ingres University Edition (relational database model product with SYSTEM / R.). The company's first relational database product is SybaseSQLServer1.0 launched in May 1987. Sybase first proposes the idea of the Client / Server database architecture and takes the lead in implementing in your own SybaseSQLServer. In 1985, Sybase's first accounting year turnover is $ 1.5 million. In 1986, Sybase took the lead in launching a connected database designed for online networking applications and is a strategic alliance with the US Microsoft. In the same year, Sybase achieved profitability. In 1987 Sybase publishes the Sybase Open Client / Open Server application interface, providing a general communication function for different client devices and Sybase Grows and Expands. 1990 Sybase Introduction First, integrate IBM MVS large host to a local network master from the environment. American Wealth (Fortune) is included in the top ten of the fastest growing company in the United States. In 1991, Sybase had more than 100 million US dollars, 1990 accounting year revenue number was $ 4 million, once again listed by wealth magazine as one of the fastest growing companies. In the same year, PowerSoft launched PowerBuilder 1.0. 1992 Sybase Announcement 1991 The revenue of the annual accounting year is $ 60 million, which is 65% greater than the 1990 accounting year. In the same year, Sybase published SYSTEM 10 and took the lead in providing a full-legged work architecture from the architectural operation. In 1993, Sybase launched Omnisql Gateway, enabling the system between the database of different components, can be transparent, cross-platform jobs. In the same year, Sybase published Replication Server, which is the first set of solutions for distributed master from architecture operations, enterprise, high-accessibility solutions. In 1994, Sybase purchased MICRO Decisionware, and quickly published Enterprise Connect. Enterprise Connect is the first architecture that can across platform jobs, providing a function of unprecedented or non-conflict, large host, and master from an unprecedented environment. In the same year, PowerSoft merges Watcom and introduces Infomaker. SYBASE was merged with PowerSoft. It has to be mentioned here is the founder of PowerSoft: MITCHELL E. Kertzman. He has similar experience with Microsoft's Gates, Oracle's Ellison, University dropping out of school, starting a family, operateing a patented technology until it is big, its company's flagship product Power Builder has led an era. In the same year, Sybase is selected by the Time Warner Communications company as a standard system, and the Beijing Commodities Exchange also selected Sybase as a standard system. At the same time, Sybase became the first software company through ISO 9001 highest standards. Moreover, Sybase is selected as the core technology of the US Airlines and Space Agency (NASA) "Database Of The Century". Sybase became the sixth largest software company in the world. ◆ Informix: Informix was established in 1980, the purpose is to provide professional relational database products for open operating systems such as UNIX. The company's name Informix is a combination of Information and UNIX. Informix's first relationship database product that truly supports SQL language is Informixse. Informixse features simple, light, adaptable. Its installed capacity is very large, especially in the microcomputer Unix environment at the time, becoming the main database products.
It is also the first commercial database product that is ported to Linux. In the early 1990s, online transaction processing became an increasingly main application of relational databases, and the Client / Server structure was growing. In order to meet the needs of online transaction processing based on the Client / Server environment, Informix introduces the concept of Client / Server in its database products, and the request to the database with the database, the application is released, and introduces Informix-Online. One feature of Online is a major change in data management, that is, the data table is no longer a single file, but the database space and logic device. Logic devices can not only be built on the file system, but also the partition and bare devices of the hard disk. This improves the security of data. In 1993, in order to overcome the limitations of multi-process system performance, INFORMIX rewrote the core of the database using multi-threaded mechanism. At the beginning of the second year, Informix launched InformixDynamic Server, which is called "dynamic scalable structure" (DSA). In addition to the application thread mechanism, INFORMIX introduces the concept of virtual processing in the database core, each virtual processor is an Informix database server process. In Dynamic Server, multiple threads can be executed in parallel in the virtual processing buffer pool, and each virtual processor is executed in the actual multiprocessor. More importantly: In order to perform high-efficiency and multi-function tuning, Informix classifies virtual processors according to different processing tasks. Each class is optimized to complete a specific function. By the late 1990s, with the rise of Internet, electronic documents, pictures, videos, space information, Internet / web, etc., the data type managed by the relational database is still in the number, string, date Waiting for the level of the sixty-19th year, its handling ability is not strong. In 1992, Professor Michaelstonebraker, the founder of the founder of INGRES, proposed the object relationship database model, which found a valid way to solve the problem. In 1995, STONEBRAKER and its R & D organizations have joined Informix, making it a new breakthrough in the development direction of the database: 1996 Informix launched a universal data option. This is a database server for an object relationship model; it is different from the solution of other vendor middleware. From all aspects of the relational database server, the database is subject to object-oriented expansion; abstracting the various mechanisms of the relational database, universalization. UniversalDataOption uses all underground technologies of Dynamic Server, such as DSA structure and parallel processing, while allowing users to create complex data types in the database and user-defined data types, while defining various operations and operations for these data types to implement objects Package. Database procedure language, C language can be used when defining operations and operations, and they become part of the server after registration. In 1999, INFORMIX further optimizes the Universal Data Option to provide a complete tool environment for user-defined data types and operational processes. At the same time, the performance of traditional transactions has exceeded the past Dynamic Server. The new database core is named IdS.2000. Its goal is positioned in the next century Internet-based complex database application. In 2001, it was acquired by IBM. ◆ MS SQL Server: In 1987, Microsoft and IBM cooperated to develop completion OS / 2, IBM binds OS / 2DatabaseManager in its sales OS / 2EXtendedededited system, while Microsoft's product line is missing database products, inadvertent location .
To this end, Microsoft puts its gaze at the time, although the product has not officially launched the product but has been technically emerging Sybase, the cooperation agreement is signed with Sybase, using Sybase technology to develop a relational database based on OS / 2 platform. In 1989, Microsoft released SQL Server1.0. In 1991, Microsoft and IBM announced the development of OS / 2, but Microsoft continued to work with Sybase with Sybase in 1992 to launch OS / 2 SQLServer4.2. At this time, Microsoft has already planned 32-bit versions based on the WindowsNT system. In 1993, shortly after the launch of WindowsNT3.1, Microsoft released the Windows NT version of SQL Server as scheduled and successful. At this time, Microsoft and Sybase have a crisis. On the one hand, SQL Server based on Windows NT has begun to form a competition for Sybase based on UNIX; on the other hand, Microsoft wants to optimize the SQL Server for Windows NT, but cannot obtain Sybase modification code. . After consultation, both parties reached an agreement in 1994, announced that both parties develop database products, Microsoft got free to modify SQL Server code license, after which Windows NT became the only platform for SQL Server running. Following SQL Server 6.0 that issued a code for SQL95 in 1995, Microsoft launched a far-reaching SQL Server6.5. SQL Server6.5 is a performance stable and powerful modern database product. It is worth mentioning that the product is completely used using the API interface of the Windows platform. It does not use unapprounded internal functions, which is completely used as an application, and does not directly use the address space of the operating system. SQLSERVER 6.5 uses multi-threaded models, supports dynamic backup, large-scalely invoking debug objects, providing open interfaces and a set of development, management, monitoring tools, but also supports multi-CPU support. ◆ Interbase: An earlier and more mature RDBMS developed by HP, and then sold to Ashton-Tate, and finally M & A of Borland, mainly integrated in Borland's products. Seeing this, we suddenly found so many relational database products in the same way, each company has a nearly similar history, including the same history, including IBM, Blue Giant In the early 1990s, I also had a risk of bankruptcy. I am beginning to believe in the first and technical second case of the market ... What is the development of the database? How long does it have to have the vitality of the relational database? The database now development direction is a tendency that newly appeared will combine the original relational database with many other features, such as email, personal communication, etc. In the fields of enterprise automation, e-government, etc., people are also integrated with each other, and also integrate with database technology. More popular research areas also have XML database technology, which mainly handles semi-structural data that cannot be processed in the past, such as document data, which is different from the traditional relational database. The data managed by the database has a fundamental change. Its basic logo is from the past, only simply managed by the digital, character, etc. in the past, and today, various types of devices, devices, and calculations must be managed. Complex data. For example, graphics, images, video, audio, e-books, web pages, etc. This change has brought a lot of challenges to database technology, and we need to study many new problems. Relationship databases have not been able to compete for this changed application, which can be said that it faces a new watershed in the development of database technology, we must seek new solutions. First, a basic issue of the database is to find an appropriate data model to express the object it managed.
