is compared with the experimenta
ranges of various workload types. D
the closeness of the extracted va
stored range, the workload is classifi
improve the accuracy of identifying
type database size in conjunction w
has been used.
IV. FUZZY-CONTROLLED SELF-TUNING &
Figure 2 shows the architecture o
fuzzy-controlled self-tuning database
system. The objective of this system
the DBMS, by proactively mo
performance indicators like bu
number of active processes and the
DBMS
N
BHR
DBS
Per
Ind
Ext
Mo
Figure 2
A. Fuzzy Control Rules
Fuzzy control rules are based
logical reasoning and based on
language constructs namely the IF T
construct. Linguistic terms are use
the extent of tuning required. These
carefully framed and most of the
modifications till desired results are
instance, to find the new buffer ca
fuzzy rules could be formed us
variables like Poor, Low, Mediu
Very High. In all, thirteen rules ha
and are as shown in Table I. T
formulated by dividing the input sp
to four zones depending on the size
al established
Depending on
alue with the
ied. To further
the workloadwith
the BHR
MODERATION
of closed loop
e management
m is to analyze
onitoring the
uffer-hit-ratio,
database size