Distributed Database Management Systems: Lecture 17

Distributed Database Management Systems: Lecture 17. The main topics covered in this chapter include: continue with VF to information requirement; attribute affinities; replication of Key attributes does not violate the disjoint ness condition;...

Distributed Database Management Systems

In this Lecture

• Continue with VF

– Attribute affinities

Replication of Key attributes does not violate the disjoint ness

condition

Vertical Fragmentation Information Requirements

• Basic idea of VF is access efficiency

• Information Requirement is application

based

• Attribute affinities: obtained from more

primitive usage data

• (80-20 Rule)

• Attribute usage values: Given a set of

queries Q = {q 1 , q 2 ,…, q q } that will

run on the relation R[A 1 , A 2 ,…, A n ]

PROJ(jNo, jName, budget, loc)

JNO=Value

q2: SELEC JNAME, BUDGET FROM PRO J

q3: SELECT JNAME FROM PRO J

• AUM does not represent the query frequency at

aff(A i , A j ) =

∑ ∑ refl(qk)accl(qk)

where refl(qk) is number of accesses to attributes (Ai,

Aj) for each execution of qk at site Sl, and…

acc 1 (q 1 ) = 15, acc 2 (q 1 ) = 20,  acc 3 (q 1 ) = 10 acc 1 (q 2 ) = 5,   acc 2 (q 2 ) = 0,    acc 3 (q 2 ) = 0 acc 1 (q 3 ) = 25, acc 2 (q 3 ) = 25,   acc 3 (q 3 ) = 25

Clustering Algorithm

• VF is based on identifying groups of

attributes based on AA

Energy Algorithm (BEA); it uses AA; identifies groups of similar items

• Large affinity attributes are combined

together and lower together

• BEA takes as input the AA and

generates the cluster affinity matrix CA

Global Affinity Measure (AM)

• Affinity Measure is a single value

that is calculated on the basis of

positions of elements in AA and their surrounding elements

AM =  ∑ n

i = 1

n

j = 1

aff(A 0 , A j )= aff(A i , A 0 )=

aff(A n+1 , A j )= aff(A i , A n+1 )=0