OT: To me it seems that SQL-based database systems lose more and more database market share in favor of old-fashioned API based databases, based on SOAP, MIDAS, CORBA and other technologies, where SQL is not involved, and where data processing and storage easily distributes across several computers without the need of clustering software.
Examples:
- MySQL: No transaction capabilities in this SQL server.
- SOAP, CORBA etc: Retrieve data using HTTP. Stateless.
- MIDAS: From Inprise, makes it possible to have extreme numbers of clients to a database server.
- DBISAM, B-Trees, db-libraries, FlashFiler, dBase files, Paradox etc. Extremely fast and works. This is what makes Microsoft Outlook Express fast.
- Distributed databases like search Engines (Google, Altavista). Might be based on SQL, but that's not the point.
- Microsoft wants people to use ADO instead of ODBC. ODBC is SQL-based, whereas ADO gives direct access to recordsets with no SQL involved, at a much better speed.
I believe that the reasons for the return to non-SQL databases connects with the fact, that SQL is simply not good enough for specialized databases or databases with extreme performance requirements.
We use less and less SQL in our projects with great success. |
