Field specialization in a nutshell
The core technology offered by Dataware is field specialization. The benefit of this technology is simple to state: field specialization speeds up DBMSes by substantial amounts: 3X for some queries and approaching a factor of 2X for industry-standard benchmarks. The basic underlying approach of this technology is also easy to state: field specialization removes machine instructions that have been, through prior analyses, determined to be unneeded. Doing so requires a very carefully orchestrated blend of sophisticated mechanisms (which cannot be captured in a nutshell…), but these are hidden to the end user of the DBMS: to that user, the DBMS just runs faster.
Field specialization has three significant advantages.
First and most importantly, it does not require an expensive and time-consuming rewrite of the DBMS source code, but rather can be applied to existing DBMS code.
Second, it does not affect the functionality of the DBMS, so for example, the entire test suite and all applications written against the DBMS continue to work.
Third, field specialization is independent of, and orthogonal to, the other approaches to DBMS performance improvement, which means that it can be applied in conjunction with them. This means that any investments a DBMS vendor may have made in developing other performance improvement technologies are not nullified by field specialization—rather, field specialization further enhances the effects of those technologies to produce multiplicative improvements in performance.