|Next Page||Prev Page||Main Index||Home|
Why does MPbase speed up the whole system? The answer has
two main parts. Part one is the environment external to the DBMS.
In this part, the application's Graphical User Interface (GUI)
first asks a question. This question is translated into a number
of database calls. Each call to the database may result in one
or more disk accesses. Once the application-data-handling logic
has assembled the answer, it is presented back to the GUI.
With MPbase the question is passed directly from the GUI
to the database engine. MPbase then makes a number of disk
accesses to answer the question. The answer is then passed back
to the GUI.
The resulting reduction in communication overhead produces a very
significant savings. This combines with the ability of MPbase
to resolve the query in a more straightforward manner. A manner
not possible under the restraints of the more traditional interface.
The second part of the answer is internal to the DBMS itself.
For each query to the DBMS the above pattern repeats internally
inside the database. In the case of MPbase the above pattern
is again repeated internal to the database.
This results in even more reduction of communication overhead
combining with even more intelligent ways of resolving the query.
The reason MPbase is so much faster should be obvious from the above diagrams. MPbase can answer the same questions/queries with much less effort. Two reasons this is possible are data naturalization and content addressable memory. The classic relational model does not allow for this type of optimization. It is therefore not possible under a conventional RDBMS.
|Next Page||Prev Page||Main Index||Home||© 1998-2004 NPSI|