Importance

Loading Importance Indicators...

Abstract

A method of declustering data among a number of storage units, e.g., disks, to minimize response time for range queries. Data to be stored is identified in the form of a dataset, wherein the data is divided among uniform data blocks (x, y) and each data block is to be stored in one of a number (M) of storage units each of which has an associated seek time. M pairs (i, key.sub.i) are defined wherein i=0,1, . . . m-1 and key.sub.i is the fractional part of 2i/(1+5). First components of the pairs are ordered according to an ascending order of the key.sub.i values of the pairs, thus defining a permutation array A of size M. Another permutation array B is defined such that for i=0,1, . . . ,M-1 B[a(i)]=i, and data block (x,y) is assigned to storage unit (x-B[y mod M])mod M, wherein mod is the remainder function.

Loading the Abstract Image...

First Claim