Http://www.cnblogs.com/xxxtech/archive/2011/12/30/2307859.html
by Euan Garden
I have heard of many versions of this rumor over the years, the most interesting being that "Microsoft is too late to confuse the purchase of Sybase code, so SQL Server is actually a Sybase product".
First, let's review history. In 1987, Microsoft and Sybase worked together to build and sell a database management system based on Sybase DataServer. Sybase enjoys the rights of SQL Server on the unix/minicomputer platform, while Microsoft enjoys the rights of SQL Server on the then nascent OS/2 platform and any Microsoft-developed platform.
At that time, Ashton Tate's dBASE was in the lead position in the PC database system (non-mainframe/minicomputer/unix) (My first database application was the dbaseiii+ used on the DOS 3.3 platform). To make better use of existing large dBASE and enable Ashton Tate to provide a true client server, Microsoft and Ashton Tate released Ashton Tate/microsoft SQL with the help of Sybase Server (Sybase later renamed their products on the UNIX and VMS platforms to Sybase SQL Server), which later became the name of Sybase DataServer on the OS/2 platform.
The first version of Ashton Tate/microsoft SQL Server was released in 1989. Initially, dBASE IV planned to provide front-end development tools for SQL Server (based on previous collaborations), but the late release of dBASE IV in the current year resulted in a miscarriage of the plan. As a result, Microsoft and Ashton Tate removed the partnership, and the product was named Microsoft SQL Server (some of the developers at the time still had a very proud package with the Ashton-tate/microsoft SQL Server box).
Since then, Microsoft has continued to improve SQL Server by adding support for the Windows platform in SQL Server 1.1, which was released in 1990. However, much of this work is still done by Sybase, and Microsoft is only responsible for testing, project management, and some minor development. By the year 1991, Microsoft's team was able to read and change the code for SQL Server in order to fix flaws in the product. The first truly cooperative version was the 4.2 release of 1992, with the OS/2 version of the code and the Sybase 4.0 code in the same step. It is important to note that Microsoft's code was used for this version for the first time, like Sybase's code.
Then the situation began to heat up, Microsoft started in the 32-bit version of SQL Server,sybase to expand on the system 10 platform development. At this time, the OS/2 platform showed a lack of competitiveness, while Microsoft had completed the beta version of Windows NT and had only a 32-bit version. So the SQL Server team decided to build a port on the NT platform based on a stable version of the 4.2 code, and Sybase continued to develop for the system 10 platform. Now the word "port" is no longer appropriate. The NT platform provides many opportunities for the product to use the operating system features, and the key is that the operating system provides the ability to support running on an SMP system rather than the code that must be written to the database. What this team has done has resulted in a big difference between the 4.2 version on the OS/2 platform and the SQL Server on NT (Side note: SQLNT was once an internal mail account for developing communications, and the account is still present and has a team of development teams, support teams, documentation and courseware members).
At this point, Microsoft and Sybase's cooperation began to lose its value. Sybase wants to remain platform neutral, and Microsoft wants to be fully committed to the Windows NT platform. In addition, Microsoft must obtain Sybase approval to add new features to SQL Server, subject to the agreement. Their cooperation collapsed in 1994, and Sybase was first allowed to release its products on the OS/2 and Windows platforms, and Microsoft could take the product to the direction it wanted. In just 18 months, Microsoft published SQL Server 6.0 and 6.5 on the basis of the original NT platform SQL Server code. These versions rewrite existing code and add new code. It was at this point that Microsoft successfully owned the rights and code of SQL Server and made it significantly different from the 4.2 version of Sybase code on the original OS/2 platform.
In this transition, Microsoft decided to make the data management frontier bigger. That means they have a lot of work to do, including building a more powerful team. Microsoft employs some of the best and most experienced experts in the database industry (from Dec Hal Berenson, Peter Spiro, David Campbell, IBM James Hamilton, Lubor Kollar and others, Oracle's bill Baker, Pedro Celis, and Tandem's Pat Hellan) and equip them with the best researchers (Jim Gray, Phil Bernstein and others) as well as masters and PhDs from the world's smartest database majors. They formed a great and focused team within a short period of time with the original team of soul figures (such as Ron Soukup) and some within Microsoft. Some people in the team are committed to SQL Server 6.0/6.5, while most people do the research and development of SQL Server 7.0, codenamed Sphinx.
Sphinx's goals are clear and a set of new standards is set up to streamline the complete data management. This means building a scalable platform for the next few years and learning from previous versions of SQL Server and other database platforms. To do this, you must rewrite the database engine, the new query processor, the new storage engine, and a new set of data access APIs (OLE DB and ADO, which eventually place Dblib).
Although SQL Server is known for its relational database management systems, Microsoft wants to provide a complete data solution in SQL Server 7.0. It adds support for online analytical processing through OLAP Services (this part of the code comes from an acquisition of the Israeli "Plato" Company), which integrates data extraction, transformation, and loading through DTS (this part of the code is developed by the internal Starfighter/tools team).
The SQL Server 7.0 RTM edition was released in November 1998 (I just received a visit in Redmond that week, remembering that the main park's 1th and 3rd buildings were very lively) and released its official version in January 1999. Although I have some prejudice, I believe that Microsoft does more work than I have listed above. So can you also say that Microsoft SQL Server is a product of Sybase? The main components are not newly added or rewritten, so Microsoft SQL Server does not have any relationship with Sybase SQL Server.
Apparently the SQL Server journey did not end in version 7.0. followed by the August 2000 release of SQL Server 2000. Its main work includes multi-instance and cluster implementations (all of which are Microsoft-produced), data mining (from research projects carried out in the early Microsoft Studies), and SQLXML (from the MDAC and network and interface team's ancillary projects).
Before SQL Server 2005, the most influential new feature that SQL Server 2000 provided was Reporting Services. The code for Reporting Services in SQL Server 2000 is entirely Microsoft-produced. The team that developed the Reporting Services consisted of the original Repository/meta Data services team, some members of the English Query team, and a group of new recruits.
The main new features and code in SQL Server 2005 are worth a few blogs, but I would like to highlight SQLCLR (from integration with a long-term strategic project), XML support (we've investigated the best way to store XML, The latter is selected using a dedicated XML database or integrating XML into SQL Server. This requires a large number of two planning and language changes to support XQuery, and Service Broker (after Dave Campbell and the Israeli MSMQ team surveyed Messaging/queuing/soa, we decided to integrate this ancillary project), Integration Services (from some fragmented code, from the original DTS team and some people from C + +, hardware, as, and COM + teams to build a new team, plus a small-scale acquisition to create an environment for design and execution), the report Builder (based on the concept and architecture of an acquisition, but the code is brand new).
Dave Campbell a day in his 10 SQL Server career, and some members are still in the team longer than he is, but the SQL Server team (MDAC, System.Data, System.Xml, and WinFS) Almost no one in all 1000 people has ever seen a line of code written by Microsoft employees.
Conclusion: The rumor was ended!
See the original: http://blogs.msdn.com/b/euanga/archive/2006/01/19/514479.aspx
Follow up: http://blogs.msdn.com/b/euanga/archive/2006/05/23/597677.aspx
Rumor Terminator--sql Server is a product of Sybase rather than Microsoft's