Overview

Shared-nothing partitioning is a divide-and-conquer strategy for solving hard computationally complex problems over large data sets. The idea of breaking large, difficult problems into smaller ones is an ancient one that crosses many cultures. The expression divide and conquer is common is several languages: French diviser pour r gner, German teile und herrsche, Latin divide et impera, and Spanish divide y vencer s. The term is most famously associated with Julius Caesar for his use of divide-and-conquer strategies in his military campaign against Gaul. More important, divide and conquer is a well-known technique in computer science. The technique is named divide and conquer because a problem is solved efficiently by dividing it into several smaller problems.

So when you have a truly large data set and a nasty looking query to resolve, why not divide up the data into little bits and exploit a team of computers to each solve the problem on just a fragment of the data, and on just a fragment of the data? In this way queries can be answered in a fraction of the time. That s the concept of shared-nothing partitioning, and we ll see...