Skip to main content

Table 1 Notation used

From: On hierarchical clustering-based approach for RDDBS design

nf

The total number of fragments

m

The number of sites

Q

The number of queries under consideration

pr

The number of predicates

CN

The number of clusters of sites

CQ

The number of clusters of queries

i

Fragment index, I = {1,…, n}

j

Site index, j = {1,…, m}

K

Query index, h = {1,…, k}

cni

Sites’ cluster index Cni = {1, …, cn}

cnq

Query’s cluster index Cnq = {1, …, cq}

Ci

Storage capacity of site Si

V

The attribute size

Size (Fk)

Size of the kth fragment

Sel (Qk)

The percentage rate of rows returned by the kth query

QFMij

The frequency of the ith query over site Sj

CCMij

The communication cost unit between cluster Ci and cluster Cj

CMSij

The communication cost unit between site Si and site Sj, measured in “ms/byte” units, and bandwidths fall in; 64 kbps, 128 kbps and 512 kbps

COMij

CMSij cost, if Si and Sj are at the same cluster; otherwise, CCMij cost

LNF

The lowest number of fragments allowed at each site

UNF

Upper number of fragments allowed at each site

TC

The total cost of data transmission as distributed query processed

XFkj

1, if fragment Fk is placed in Sj; 0, otherwise

ACvalue

ACvalue is a piled-up value that refers to the total access cost produced by each query to access its relative fragment

FPM

Matrix in which each FPMij is 1, Fi contains Pj; 0, else

FSM

Matrix in which each FSMij is ACvalue, Sj requires queries of Fi; 0, otherwise

SPM

Site Predicate Matrix

TSPM

The total cost of SPM which is grouped based on the container clusters of sites

TCCM

Matrix of Total costs of communication between clusters

TCSM

Matrix of Total costs of communication between sites

QUM

Query Usage Matrix; 1, Qi required by Sj; 0, otherwise

QDM

Query Difference Matrix, so each QDMij is the difference value between Qi and Qj