Normally database means deatailed data, and Volatile data and which contains summaraised data ,Non Volatile and subject oriented and time varient systems called as datawaer house
As per my knowledge : actually clusters will be using to not to have down time in the source system. mpp means massivly parlell processing (i.e parlelly processing from the source to target massively by placing multiple processors) but coming to cluster point, first of all when ever we are fetching data from one 'A'system to 'B' system,'A' system should be down then only we can fetch data from 'A' to 'B'. so now here by using cluster we do not want to down the 'A' system
Data mart is subset of dataware house In Brief: Assume that there is an company called vijayJDK, and it is havaing 4 dept 1:petrolium 2:deisel 3:kirosine 4:crudoil now if you develop dataware house for any of them above 4 dept's we call them as datamarts,and then if you consolidate 4 dept's then ,that will be called as dataware house
