Copy
ri
ght
©
Q
uali
ty
Wize
™
,
A
lon
L
in
et
z
ki
alo
n
l@
q
u
alit
y
-
w
iz
e.c
om
The need for Speed
…
P
er
for
mance
&
Lo
ad
T
esting W
orkshop
Alon
Lin
e
tzki
alo
nl@q
uali
ty
-
wiz
e.c
om
Copy
ri
ght
©
Q
uali
ty
Wize
™
,
A
lon
L
in
et
z
ki
alo
n
l@
q
u
alit
y
-
w
iz
e.c
om
Alon
Linetzk
i
CEO
and
Found
er
o
f
Qua
li
ty
Wi
ze
™
35
+ y
ear
s
in
SW engin
eer
in
g,
agil
e
tr
ansition,
t
est
ing,
qua
li
ty
assur
ance a
nd
Co
-
authored
IS
T
QB® Agile
T
ester
C
er
ti
fication,
c
o
-
authoring
the a
dv
an
c
ed
lev
el I
ST
QB
®
T
ec
hn
ica
l
Ag
ile
T
este
r
IS
T
QB
® M
ark
eting
Chair
Co
-
founder
o
f
the
Israeli
T
esting
Cer
tifi
cation Bo
ard
,
Vi
ce
Pr
esident,
Ma
rk
et
ing
Dir
ect
or
Found
er
o
f
the
SI
GiST
Israel
(t
est
ing
for
um)
in
Isra
el
•
C
e
rtified
Sc
ru
m
Mas
t
e
r
•
C
ertified
S
y
s
t
em
Analy
s
t
•
I
S
T
QB
Adv
an
c
ed
Le
v
el
–
T
es
t
Man
ag
er
an
d
T
es
t
Analy
s
t
•
I
SO
-
9000
Le
ad
As
sess
or
•
TMMi
As
sess
or
•
C
ertified
Mobile
T
es
t
er
(
C
MAP
)
•
C
ertified
C
omm
u
n
ic
a
tion
Skills
c
oach
•
MB
A
•
Ms.
c
St
a
tis
tic
s &
C
rimin
ol
ogy
Copy
ri
ght
©
Q
uali
ty
Wize
™
,
A
lon
L
in
et
z
ki
alo
n
l@
q
u
alit
y
-
w
iz
e.c
om
Agenda
Intr
oduction
What
is
P
er
formance T
esting?
What
do w
e focus
on
?
Prime
di
r
ec
tiv
e
o
f
pe
r
forma
nce
t
esting
Benefi
ts
per
formance
t
esting
T
ype o
f pe
r
formance te
st
s?
Common
pe
r
formance pr
oble
ms
What
t
o
mea
sur
e?
Baseli
ne
and
Bench
marking
Sett
ing
th
e
r
eq
uir
ements
Common
Mistak
es
Su
mmar
y
[op
tional]
Lo
ad
and
pe
r
forma
nce Simula
tion Mode
ls
[opti
on
al]
Ho
w
earl
y
sh
ou
ld
we
pe
r
formance te
st
?
[o
pt
iona
l]
Lo
ad
T
est
R
eso
ur
ce
T
op
olo
gies
Copy
ri
ght
©
Q
uali
ty
Wize
™
,
A
lon
L
in
et
z
ki
alo
n
l@
q
u
alit
y
-
w
iz
e.c
om
Introductio
n
T
h
e
i
m
por
tan
ce
o
f
p
e
r
fo
r
manc
e
t
e
s
tin
g
W
h
a
t
i
s
p
e
r
for
man
ce
t
e
s
t
ing
W
h
a
t
d
o
w
e
f
o
c
us
o
n
?
Pr
i
me
d
i
r
ect
ive
o
f
p
e
r
for
man
ce
t
e
s
t
ing
Be
n
efi
ts
o
f
p
e
r
fo
r
man
c
e
t
e
s
ti
ng
Copy
ri
ght
©
Q
uali
ty
Wize
™
,
A
lon
L
in
et
z
ki
alo
n
l@
q
u
alit
y
-
w
iz
e.c
om
S
yst
ems
ar
e
getting m
or
e a
nd mo
r
e
comp
lex
New tech
nol
ogies
ar
e arriving
at o
ur
door
st
ep
–
AI, Io
T
, Io
E, mobile, c
loud, big
-
data,
Non
SQL
DB, mor
e
…
Netw
orks
and
t
op
ol
og
ies
ar
e getting
mo
r
e
div
er
sified and co
mplex, and pr
ov
ided
as
co
mm
odity
(cloud P
aa
S,
et
c.
)
The impor
tan
ce o
f per
forma
nce
t
es
ting
Copy
ri
ght
©
Q
uali
ty
Wize
™
,
A
lon
L
in
et
z
ki
alo
n
l@
q
u
alit
y
-
w
iz
e.c
om
Mass
user
s
ar
e using
the syst
ems
mo
r
e
and
mo
r
e
S
yst
ems ar
e
mor
e
glo
b
al
and w
orking
24
*
7
Bus
iness i
s r
equir
ed
t
o pu
t up
a
fast
suppo
r
t
systems
,
and a
r
oun
d the c
lock
ser
vices
S
ystem
s
ar
e growing
fast
er
than
in the p
ast,
r
equiri
ng
fast
er per
form
anc
e
impr
ovement
over
time
The impor
tance
o
f pe
r
formance
te
sting
Copy
ri
ght
©
Q
uali
ty
Wize
™
,
A
lon
L
in
et
z
ki
alo
n
l@
q
u
alit
y
-
w
iz
e.c
om
What I
s P
er
for
man
ce
T
esting?
P
er
formance
testing
is in gen
era
l,
a
testing
type
per
formed t
o det
ermine
ho
w a
s
ys
te
m
pe
r
forms
in
te
rms
o
f responsiv
ene
ss
and st
abi
lit
y
unde
r
a
p
ar
ticula
r
wor
kload
P
er
forma
nce T
esting is
being
me
asur
ed
(many
times)
in t
erms o
f
r
espo
nse time
,
thr
oughput
an
d
r
eso
ur
ce util
iz
ation
under running
w
orklo
ad
W
e r
egularly t
est
3
ma
jor issues:
Spee
d, sc
alabili
ty
and S
tabili
ty
under
lo
ad
Copy
ri
ght
©
Q
uali
ty
Wize
™
,
A
lon
L
in
et
z
ki
alo
n
l@
q
u
alit
y
-
w
iz
e.c
om
What
do
w
e
foc
us
on?
R
esponse
time
–
the
me
as
ur
ed time
tha
t the
end
-
user
-
r
equest
is waiti
ng
for a
r
espo
nse
fr
om
the appli
cation
Copy
ri
ght
©
Q
uali
ty
Wize
™
,
A
lon
L
in
et
z
ki
alo
n
l@
q
u
alit
y
-
w
iz
e.c
om
What
do
w
e
foc
us
on?
Thro
ughp
ut
-
the amou
nt o
f
transa
ctions
pr
oduc
ed over
time during
a t
est o
r
the a
mo
unt
o
f
cap
acity
the appli
cation c
an
handle
Copy
ri
ght
©
Q
uali
ty
Wize
™
,
A
lon
L
in
et
z
ki
alo
n
l@
q
u
alit
y
-
w
iz
e.c
om
What
do
w
e
foc
us
on?
R
esource
utilization
-
the t
ota
l
am
oun
t o
f
r
eso
ur
ces
ac
tually
co
nsum
ed, c
om
p
ar
ed
against
the amo
unt
o
f
r
esour
ces pl
anned
for a
specific
pr
oc
ess
Copy
ri
ght
©
Q
uali
ty
Wize
™
,
A
lon
L
in
et
z
ki
alo
n
l@
q
u
alit
y
-
w
iz
e.c
om
Prime
dir
ectiv
e o
f P
er
formance
T
esting
The o
bjectiv
e
o
f
per
form
anc
e
t
est is t
o v
alidat
e
the b
ac
k
-
end ar
chit
ectur
e, har
dwar
e
and
applications
sc
alabili
ty:
Det
er
min
e
the
p
er
for
ma
nce, st
ab
il
ity
and
scalab
il
ity
o
f a
n
appl
icat
io
n
under
v
ar
ious lo
ad
conditions
Det
er
min
e
whi
ch
configur
ation s
izi
ng
pr
ovides
the
best
per
for
mance
lev
el
De
t
er
mine
if the
cu
rr
en
t
ar
chit
ec
tu
r
e
ca
n
su
pp
or
t t
he
ap
pl
icatio
n
at peak
us
er
lev
els
ht
t
p://
s
wqu
al
i
t
yas
s
ura
nc
e.co
m
Copy
ri
ght
©
Q
uali
ty
Wize
™
,
A
lon
L
in
et
z
ki
alo
n
l@
q
u
alit
y
-
w
iz
e.c
om
Prime
dir
ectiv
e o
f P
er
formance
T
esting
Pr
ov
e
appl
icat
io
n
is
st
ab
le
en
ough
t
o
go int
o
pr
oduction
(Ac
ce
pt
ance
)
Det
er
min
e
if th
e
new
v
ersion o
f t
he
so
f
twar
e
had
a
ne
gativ
e imp
act
on
r
es
po
ns
e
time
Det
er
min
e
at what
poi
nt
does
de
gr
adation
o
f
per
for
mance
occ
ur
(C
ap
ac
it
y Planning
)
Identi
f
y
appl
ic
atio
n
and infra
structur
e
bo
tt
lene
cks
Ev
aluat
e
pr
oduct
and/or
har
dwar
e t
o
det
er
mine
if
it
ca
n
handle
pr
oj
ect
ed
lo
ad
v
ol
umes
ht
t
p://
s
wqu
al
i
t
yas
s
ura
nc
e.co
m
Copy
ri
ght
©
Q
uali
ty
Wize
™
,
A
lon
L
in
et
z
ki
alo
n
l@
q
u
alit
y
-
w
iz
e.c
om
Ben
efits
o
f P
er
for
man
ce T
esting
Hi
gher
Pr
oduc
tion R
eadiness
Meeting
User R
equir
emen
ts
for doing bus
iness
Impr
oved efficiently
o
f
per
form
anc
e
R
educed c
ost
–
ma
int
enan
ce,
opera
tional
Meeti
ng
P
er
forma
nce S
tandar
ds/r
egul
at
or
y
needs
Contingent
Di
sas
t
er r
eco
ver
y
situations
Copy
ri
ght
©
Q
uali
ty
Wize
™
,
A
lon
L
in
et
z
ki
alo
n
l@
q
u
alit
y
-
w
iz
e.c
om
T
ype
s
o
f
per
formance
te
sting
Copy
ri
ght
©
Q
uali
ty
Wize
™
,
A
lon
L
in
et
z
ki
alo
n
l@
q
u
alit
y
-
w
iz
e.c
om
T
ypes o
f pe
r
formance
te
sts (
KPIs)
Lo
ad
t
estin
g
(c
onc
urrency)
-
ch
ec
ks
the
app
li
cation
'
s
abili
ty
t
o
per
fo
rm u
nde
r
anti
cipated
us
er load
s
.
The
ob
ject
iv
e is
t
o identif
y
per
forma
nce
bo
tt
leneck
s
befor
e
the
so
f
twar
e/
syst
em
appli
cation
goe
s l
iv
e.
S
tr
ess
t
estin
g
-
invo
lv
es
t
es
ting
an
app
li
cation
un
der
extr
eme
w
orklo
ads
t
o
see
how
it h
andles
high
traff
ic
or dat
a
pr
oces
sing
.
The
ob
ject
iv
e is
t
o identif
y
br
eaking
po
in
t
o
f
an
appli
cation
.
End
urance t
esting
-
is
done
t
o
mak
e
su
r
e
the
so
f
twar
e/sy
st
em
can
han
dle
the
exp
ec
t
ed
lo
ad
ov
er
a
long
pe
riod o
f ti
me
.
The
ob
ject
iv
e is
t
o e
v
al
ua
t
e
how
the
syst
em
work
s f
or
long
per
iod
o
f
time
, und
er
ex
pected
lo
ad
(r
egula
r
,
av
erage,
max,
abov
e
max).
Copy
ri
ght
©
Q
uali
ty
Wize
™
,
A
lon
L
in
et
z
ki
alo
n
l@
q
u
alit
y
-
w
iz
e.c
om
T
ype
s
o
f p
er
forman
ce
te
sts (
KPIs)
Spik
e
t
esting
-
t
est
s
the
so
f
twa
r
e
'
s
r
eaction
t
o
sudde
n
lar
ge
sp
ik
es
in
the l
o
ad
ge
nera
t
ed
by
use
r
s.
V
olume testing
-
Un
de
r
V
ol
ume
T
est
ing
lar
ge
no.
o
f
.
Data
is populat
ed
in
dat
ab
ase
and
the
ov
er
all
so
f
twa
r
e
syst
em'
s
be
havior
is mon
it
or
ed
.
T
he
obje
ctiv
e
is t
o
check
so
f
tw
ar
e
appl
ic
ation
'
s
pe
r
formance
un
de
r v
ar
ying
da
tab
ase
v
olumes
.
Scala
bilit
y t
esting
-
The
objectiv
e
is t
o
de
t
er
mi
ne
the
so
f
twa
r
e app
li
catio
n
'
s
ef
fe
cti
v
eness
in
"
scali
ng
up" t
o
suppor
t
an
in
cr
ease
in
user lo
ad
.
It helps plan
cap
aci
ty
add
itio
n
t
o
y
our
so
f
twa
r
e syst
em.
h
t
t
p
:/
/
w
w
w
.
gu
ru
99
.c
om
/
p
er
f
or
man
c
e
-
t
es
t
in
g.h
t
ml
Copy
ri
ght
©
Q
uali
ty
Wize
™
,
A
lon
L
in
et
z
ki
alo
n
l@
q
u
alit
y
-
w
iz
e.c
om
Commo
n
per
for
mance problems
Copy
ri
ght
©
Q
uali
ty
Wize
™
,
A
lon
L
in
et
z
ki
alo
n
l@
q
u
alit
y
-
w
iz
e.c
om
Common
per
for
man
ce problems
Long
Lo
ad
time
-
Lo
ad ti
me i
s no
rma
ll
y the ini
tial
time it tak
es an
application
t
o star
t.
This sho
uld generally be k
ept t
o a
minimum.
Copy
ri
ght
©
Q
uali
ty
Wize
™
,
A
lon
L
in
et
z
ki
alo
n
l@
q
u
alit
y
-
w
iz
e.c
om
Commo
n
per
for
mance problems
P
oor
r
esponse
time
-
R
esp
onse
time is
the ti
me it
tak
es
fr
om when
a use
r i
nputs
dat
a int
o
the
ap
plic
ation
until
the
ap
plic
ation
outputs
a
r
esp
onse
t
o
that
input.
Gener
ally
this
shoul
d
be
v
er
y
quic
k
(dep
ending on
the
actual
tas
k done).
R
espo
nse
Time
= Pro
cessing
Time
+ W
aiti
ng T
ime
ht
t
p:
//lo
gi
cal
r
e
a
d.s
ol
ar
w
inds
.c
om
/r
e
s
pons
e
-
t
im
e
-
an
a
lys
is
/#
.Vjs
0
Db
crK
U
k
Copy
ri
ght
©
Q
uali
ty
Wize
™
,
A
lon
L
in
et
z
ki
alo
n
l@
q
u
alit
y
-
w
iz
e.c
om
Commo
n
per
for
mance problems
P
oor
scal
abi
lity
-
pr
oduc
t
canno
t
handle
the
ex
pe
ct
ed
num
be
r
o
f
user
s
or
when
it
does
not
acc
omm
odat
e
a
wide e
no
ugh
range
o
f
user
s
.
ht
t
p://
bl
o
g
.er
r
a
t
as
ec.c
o
m
/
2013
/
02
/s
ca
la
b
il
it
y
-
i
t
s
-
qu
est
io
n
-
t
ha
t
-
dri
ve
s
-
us
.h
t
m
l#.
Vjs
0
kr
crK
U
k
Copy
ri
ght
©
Q
uali
ty
Wize
™
,
A
lon
L
in
et
z
ki
alo
n
l@
q
u
alit
y
-
w
iz
e.c
om
Commo
n
per
for
mance problems
Bottlenecking
-
Bott
leneck
s a
r
e ob
str
uc
tion
s
in
syst
em
whic
h
de
gr
ade
ov
er
all
syst
em
pe
r
for
manc
e.
Bottlenecking
is wh
en ei
ther
coding
error
s
or
hardw
are
issues
cause
a decre
ase
o
f thro
ughput
under
cer
ta
in
lo
ad
s
.
Bottle
-
necking
is g
ene
ra
lly
fix
ed
by
e
ither
f
ixing
poor
runni
ng
pr
oc
ess
es
or
add
ing addition
al
Har
dw
ar
e
Some
common
per
f
or
ma
nce
bot
tl
eneck
s
ar
e
CPU util
izat
ion
Memor
y
util
izat
ion
Ne
tw
ork
util
izat
ion
Operat
ing
S
yst
em
li
mit
ati
ons
Disk
usag
e
h
t
t
p
:/
/
w
w
w
.
gu
ru
99
.c
om
/
p
er
f
or
man
c
e
-
t
es
t
in
g.h
t
ml
Copy
ri
ght
©
Q
uali
ty
Wize
™
,
A
lon
L
in
et
z
ki
alo
n
l@
q
u
alit
y
-
w
iz
e.c
om
Measur
ing per
for
mance
(KPIs
)
Copy
ri
ght
©
Q
uali
ty
Wize
™
,
A
lon
L
in
et
z
ki
alo
n
l@
q
u
alit
y
-
w
iz
e.c
om
What
is a
K
ey P
er
formance
Indica
to
r?
A K
ey P
er
form
anc
e
Indicat
or (or
KPI)
is
usua
ll
y
defined
and us
ed t
o
ev
alua
te
the
system
’
s
ab
ility
t
o mee
t
a
cer
tain
attr
ib
ut
e
o
f
charact
er
istic
(the
“
How
”
)
.
The out
com
e
o
f
the meas
ur
ement is used t
o
trigg
er
any
tuning
and pr
oc
ess impr
ovements
.
Copy
ri
ght
©
Q
uali
ty
Wize
™
,
A
lon
L
in
et
z
ki
alo
n
l@
q
u
alit
y
-
w
iz
e.c
om
What s
hould
w
e measu
re?
If an
y o
f
the below r
eso
ur
ces
appea
r
s t
o be
going
thr
ough an ex
cessive
intak
e
–
this mig
ht
point o
ut o
n
bottlenec
ks
R
esource
Utilizatio
n
CPU
,
usef
ul when b
r
ok
en
out
by
p
r
oc
ess
(W
ait
time,
User
/S
yst
em
time,
idle
time)
Me
mor
y
, usef
ul when b
r
ok
en
out
by
p
r
oc
ess
(U
sag
e,
br
ok
en out
by
pr
oc
ess
Disk
activi
ty
(R
eq
uests
,
by
t
es
r
ead
/wr
it
e/se
c)
Ne
two
rk b
y
t
es
(By
t
es
r
ead
, Thr
oughput,
est
ab
li
shed
co
nnecti
ons)
Copy
ri
ght
©
Q
uali
ty
Wize
™
,
A
lon
L
in
et
z
ki
alo
n
l@
q
u
alit
y
-
w
iz
e.c
om
What s
hould
w
e measu
re?
P
er
formance
Co
unters
Ther
e ar
e
2
k
ey
co
mmo
n
co
unt
er
s
w
e c
an
measur
e
and
mo
ni
t
or
:
Time
T
o
Fir
st
Byte
(T
TFB)
Time
T
o
Last
Byte
(T
TLB
)
Copy
ri
ght
©
Q
uali
ty
Wize
™
,
A
lon
L
in
et
z
ki
alo
n
l@
q
u
alit
y
-
w
iz
e.c
om
What s
hould
w
e measu
re?
User Lo
ad can
be sta
tic v
alue
or ra
mp
-
up
and
down
during
per
form
ing th
e lo
ad t
est
Lo
ad
Co
unters
User
Lo
ad
-
The
num
be
r
o
f
co
nc
urr
ency
user
s
that
ar
e a
cc
ess
ing the s
yst
em
in
a g
iv
en t
im
e
R
eque
sts/sec
-
Displays
the
nu
mber
o
f
r
equest
s
be
ing i
ssued
ea
ch
se
co
nd
Err
or
s/sec
-
Displays
the n
umber
o
f
er
r
or
s th
at
oc
cu
rr
ed
during
the
lo
ad
t
est
run
each se
co
nd
(I.
E.
Un
autho
riz
ed
-
401
,
Ac
cess Deni
ed
-
403
,
File
or
Dir
ect
or
y
not
found
-
404
et
c
’
)
30
Copy
ri
ght
©
Q
uali
ty
Wize
™
,
A
lon
L
in
et
z
ki
alo
n
l@
q
u
alit
y
-
w
iz
e.c
om
Ex
er
cise
High P
er
for
mance D
isc
ussio
n
…
20
min
70
Copy
ri
ght
©
Q
uali
ty
Wize
™
,
A
lon
L
in
et
z
ki
alo
n
l@
q
u
alit
y
-
w
iz
e.c
om
High P
er
for
mance D
isc
ussio
n
…
W
e wil
l
discu
ss
a f
ew
pe
r
for
manc
e t
opic
s
:
1.
Bus
ine
ss
driv
en
per
formance
t
es
ting
-
de
scrib
e
th
e
busin
ess go
als
and
busin
ess benefi
ts
t
hat
un
de
rl
ine
doi
ng
pe
r
forma
nce t
esting
2.
P
er
formance
Planning
aspects
–
wh
at
aspec
ts
sh
ou
ld
we
th
ink abou
t
for pl
ann
ing a
pe
r
formance te
st
?
3.
W
or
kl
o
ad
mod
el
s
-
de
scri
be
r
eq
uir
eme
nt
s
for a
pe
r
forma
nce
model
.
if
y
ou
kno
w
model
s
li
k
e
th
at,
de
scrib
e
with
an e
xam
pl
e
4.
As
pects
o
f r
unni
ng p
er
formance
t
es
ting
ear
ly i
n the
li
fecy
cl
e
-
de
scrib
e
in w
hich
ar
eas
de
v
,
t
est
and
ot
her
t
ec
hn
ic
al
expe
r
ts c
an
t
est
pe
r
forma
nce e
arl
y
, and
which kind
o
f pe
r
formance te
st
ing
is th
at.
20
min
Copy
ri
ght
©
Q
uali
ty
Wize
™
,
A
lon
L
in
et
z
ki
alo
n
l@
q
u
alit
y
-
w
iz
e.c
om
Baseline
& Ben
chmarking
Copy
ri
ght
©
Q
uali
ty
Wize
™
,
A
lon
L
in
et
z
ki
alo
n
l@
q
u
alit
y
-
w
iz
e.c
om
Baseline and
Ben
chmarkin
g
A
t the v
er
y beg
inning
o
f
a per
form
anc
e
t
est,
w
e
need t
o establish
a
st
ab
le
b
aseline
A
per
formance
b
aseline
is per
form
anc
e
informa
tion g
ather
ed
at the beginning
o
f
the
ini
tiativ
e fr
om w
hich v
ariations
ar
e meas
ur
ed
thr
ougho
ut
the per
form
anc
e
li
fecycle
Aft
er e
v
ery benchmark
(ne
w
t
es
t run)
w
e w
il
l w
an
t
t
o c
ompar
e
it t
o the
baseli
ne
…
Copy
ri
ght
©
Q
uali
ty
Wize
™
,
A
lon
L
in
et
z
ki
alo
n
l@
q
u
alit
y
-
w
iz
e.c
om
Baseline and
Ben
chmarkin
g
The
b
aseline
is used fo
r
comp
ar
ison
betw
een
test r
uns
,
mo
nit
or
ing
the
pr
oject
pe
r
fo
r
ma
nce
over
the
period o
f
time
that has
p
ass
ed in pr
oject ex
ecutio
n
Ben
chm
arking
is the pr
oces
s
o
f
com
p
ar
ing
the
system
per
formance
against
the
b
aseline
Baselin
e
is r
e
flecti
ng the perf
ormance
r
equi
r
emen
ts
Copy
ri
ght
©
Q
uali
ty
Wize
™
,
A
lon
L
in
et
z
ki
alo
n
l@
q
u
alit
y
-
w
iz
e.c
om
Why Baseline
per
for
man
ce?
T
o
establish
go
als and
st
andar
ds
T
o
set SL
A
metrics and per
for
mance
th
r
esho
lds
T
o
mak
e
impor
tant
decisions
Copy
ri
ght
©
Q
uali
ty
Wize
™
,
A
lon
L
in
et
z
ki
alo
n
l@
q
u
alit
y
-
w
iz
e.c
om
P
er
for
man
ce requir
emen
ts
Copy
ri
ght
©
Q
uali
ty
Wize
™
,
A
lon
L
in
et
z
ki
alo
n
l@
q
u
alit
y
-
w
iz
e.c
om
Defin
e per
formance
r
equir
emen
ts
A specific
business
pr
oc
ess or
a mix o
f
business
pr
oc
esses
A specific
r
espo
nse time
thr
esho
ld
A specific
lo
ad
A specific
set o
f
application
cha
ract
eristics
Compu
ting
pow
er needed
and a
spec
ifi
c
co
nstrains if any
Copy
ri
ght
©
Q
uali
ty
Wize
™
,
A
lon
L
in
et
z
ki
alo
n
l@
q
u
alit
y
-
w
iz
e.c
om
Defin
e per
forman
ce
r
equi
r
emen
ts
T
ypical
conv
er
sat
ion
o
f
a
per
for
mance
tester
wit
h
the
pr
oject
manager
-
Copy
ri
ght
©
Q
uali
ty
Wize
™
,
A
lon
L
in
et
z
ki
alo
n
l@
q
u
alit
y
-
w
iz
e.c
om
Defin
e per
forman
ce
r
equi
r
emen
ts
T
es
t
er
:
What is
the
load you
wan
t
t
o
re
plic
at
e
on
the
t
es
t
ed
s
ys
t
em?
Manag
er
:
well,
I
gue
ss
2000
c
on
curr
en
t
use
rs
!
Copy
ri
ght
©
Q
uali
ty
Wize
™
,
A
lon
L
in
et
z
ki
alo
n
l@
q
u
alit
y
-
w
iz
e.c
om
Defin
e per
forman
ce
r
equi
r
emen
ts
T
es
t
er
:
What ar
e
the
mos
t freque
n
t
busin
es
s pr
oc
es
se
s
you
wan
t
t
o
t
es
t?
Mana
g
er: I
belie
v
e th
e
mos
t
c
ommon
ones a
r
e
-
(
1
)
Login,
add
it
em,
chec
k
out
.
(
2
)
Login,
search
,logo
ut.
(
3
)
Login,
fil
l
a
f
orm,
send
a
f
orm ,
logout.
Copy
ri
ght
©
Q
uali
ty
Wize
™
,
A
lon
L
in
et
z
ki
alo
n
l@
q
u
alit
y
-
w
iz
e.c
om
Defin
e per
forman
ce
r
equi
r
emen
ts
T
es
t
er
:
and wha
t
sho
uld
be the
tr
ansac
tion
r
a
t
e?
Manag
er
:
All
a
t
the same
time
…
.?#
$@?
Copy
ri
ght
©
Q
uali
ty
Wize
™
,
A
lon
L
in
et
z
ki
alo
n
l@
q
u
alit
y
-
w
iz
e.c
om
Defin
e per
forman
ce
r
equi
r
emen
ts
T
es
t
er
:
W
ell
-
c
an you
t
ell
me
the
rat
e
of
all
the busine
ss
pr
oc
es
se
s you
wan
t
t
o
t
es
t,
o
ver
an ho
ur
, at pea
k
and
no
rmal times?
Ma
nager
:
I don
’
t kno
w the
tra
nsactio
n ra
t
e,
I
’
ll
have
t
o
t
alk
with our
C
T
O
…
.
Copy
ri
ght
©
Q
uali
ty
Wize
™
,
A
lon
L
in
et
z
ki
alo
n
l@
q
u
alit
y
-
w
iz
e.c
om
Defin
e per
forman
ce
r
equi
r
emen
ts
T
es
t
er
:
??
##
@@!!??
Manag
er
:
The
C
T
O s
aid w
e
sho
uld
simula
t
e
BP
(B
usiness
Pr
ocess)
1
with
the r
a
t
e
of
X, BP
2
with
the
r
a
t
e
of
Y
…
Aft
er
a
f
e
w
da
y
s
…
Copy
ri
ght
©
Q
uali
ty
Wize
™
,
A
lon
L
in
et
z
ki
alo
n
l@
q
u
alit
y
-
w
iz
e.c
om
Defin
e per
forman
ce
r
equi
r
emen
ts
No
w
w
e
can s
t
art
…
Copy
ri
ght
©
Q
uali
ty
Wize™,
A
lon
L
in
et
z
ki
alo
n
l@
q
u
alit
y
-
w
iz
e.c
om
Scenario
s and
actions
Sc
ena
rio
A
ction
In
pu
t
Da
ta
Ou
tpu
t
Da
ta
Bro
w
sin
g
Prod
uct
Ca
talo
g
by
an
ex
istin
g
use
r
L
o
g
in
•
Un
iq
u
e
u
se
rn
a
me
•
Pa
ssw
ord
of t
he
use
rna
me
Brow
se
•
Ca
talo
g
T
ree
•
Use
r
T
y
pe
•
Prod
uct
de
scrip
tion
•
T
itle
•
Ca
teg
o
ry
Bro
w
sin
g
Prod
uct
Ca
talo
g
by
a
ne
w
use
r
L
o
g
in
•
Un
iq
u
e
u
se
rn
a
me
•
Pa
ssw
ord
of t
he
use
rna
me
Brow
se
•
Ca
talo
g
T
ree
•
Use
r
T
y
pe
•
Prod
uct
de
scrip
tion
•
T
itle
•
Ca
teg
ory
Copy
ri
ght
©
Q
uali
ty
Wize™,
A
lon
L
in
et
z
ki
alo
n
l@
q
u
alit
y
-
w
iz
e.c
om
Think
-
time
Sc
ena
rio
A
ction
In
pu
t
Da
ta
Ou
tpu
t
Da
ta
Th
in
k
T
ime
Brow
sin
g
Pro
d
u
ct
Ca
talo
g
by
an
ex
istin
g
use
r
Lo
gi
n
•
Un
iq
ue
use
rna
me
•
Pa
ssw
o
rd
o
f
the
use
rna
me
5
-
8
se
co
n
d
s
Brow
se
•
Ca
talo
g
T
ree
•
Use
r
T
y
pe
•
Prod
uct
de
scrip
tion
•
T
itle
•
Ca
teg
ory
4
-
30
seco
nd
s
Brow
sin
g
Pro
d
u
ct
Ca
talo
g
by
a
ne
w
use
r
Lo
gi
n
•
Un
iq
ue
use
rna
me
•
Pa
ssw
o
rd
o
f
the
use
rna
me
5
-
15
se
co
n
d
s
Brow
se
•
Ca
talo
g
T
ree
•
Use
r
T
y
pe
•
Prod
uct
de
scrip
tion
•
T
itle
•
Ca
teg
o
ry
10
-
60
seco
nd
s
Copy
ri
ght
©
Q
uali
ty
Wize™,
A
lon
L
in
et
z
ki
alo
n
l@
q
u
alit
y
-
w
iz
e.c
om
Usage P
atterns
U
ser
Scenario
Perc
en
t
Loa
d
D
ist
ribut
ion
Brow
sin
g
pr
oduc
t
ca
talog
40
%
Cr
eati
ng
a
user
acc
ount
5
%
S
earc
hing
for
a
produc
t
30
%
Login
to
applic
ation
15
%
Or
der
Placemen
t
10
%
T
otal
100
%
Copy
ri
ght
©
Q
uali
ty
Wize™,
A
lon
L
in
et
z
ki
alo
n
l@
q
u
alit
y
-
w
iz
e.c
om
Usage P
atterns
Us
er
Sc
ena
rio
Perce
n
t
L
o
ad
Dist
rib
u
ti
o
n
Brow
sin
g
pro
du
ct
cata
lo
g
40
Cre
atin
g
a
use
r
acco
un
t
05
Se
arch
in
g
for
a
pro
du
ct
30
Lo
gi
n
to
ap
pl
ica
tion
15
O
rde
r
Pla
ceme
nt
10
T
otal
100
40%
5
%
30%
15%
10%
P
e
r
c
e
n
t
L
o
a
d
D
i
s
t
r
i
b
u
t
i
o
n
B
r
ow
si
n
g
pr
oduc
t
cat
al
og
Cr
eat
i
ng
a use
r
acc
ount
Se
a
r
chin
g f
or
a
produ
ct
Log
i
n t
o app
l
i
c
a
t
i
on
Or
der
Pl
ac
em
e
nt
Copy
ri
ght
©
Q
uali
ty
Wize™,
A
lon
L
in
et
z
ki
alo
n
l@
q
u
alit
y
-
w
iz
e.c
om
Distribution
o
f User
s
logging
in
…
What
pr
o
file
are
t
hose
user
s
usi
ng?
Empl
oy
e
e
s
–
lots
o
f
da
ta
e
ntr
y
(i
ns
e
r
ts
),
lots
o
f
que
ri
e
s,
lots o
f
upd
at
e
s
Manag
e
r
s
–
lots o
f
que
ri
e
s,
r
e
por
ts
,
small
am
ount
o
f
upd
at
e
s,
so
me
de
le
t
e
s
Copy
ri
ght
©
Q
uali
ty
Wize™,
A
lon
L
in
et
z
ki
alo
n
l@
q
u
alit
y
-
w
iz
e.c
om
Common traps
in per
formance
te
sting
Copy
ri
ght
©
Q
uali
ty
Wize™,
A
lon
L
in
et
z
ki
alo
n
l@
q
u
alit
y
-
w
iz
e.c
om
Don
’
t
fall
int
o
commo
n
tr
aps
…
Pr
ematur
e
Optimization
Don
’
t tr
y
t
o
optimize
th
e appl
ic
ation
befor
e
y
ou
hav
e
“
t
est
ed
th
e
wat
er
”
and
ran a
fe
w per
forma
nce
cy
cl
es
Rando
m
Optimization
In or
de
r
th
e
mak
e any
optimizat
ion
y
ou
hav
e t
o
chan
ge
only
one
set
o
f a
v
ari
abl
e
(i
solat
e)
, r
et
est
th
e
sys
t
em
and
th
en
comp
ar
e
th
e r
esult t
o
th
e b
asel
ine
Insufficient
R
equir
ements
Y
ou
h
av
e
t
o
mak
e
su
r
e
th
at
all
y
ou
r
pe
r
formance
r
eq
uir
ements
captu
r
e
all
th
e
need
ed
v
ari
abl
es,
in
case
th
ey
ar
e not
-
it wil
l
be
har
d
t
o
pr
oduce
v
al
id
concl
usions
Same
a
s
la
st
v
e
r
sion
…
Copy
ri
ght
©
Q
uali
ty
Wize™,
A
lon
L
in
et
z
ki
alo
n
l@
q
u
alit
y
-
w
iz
e.c
om
P
er
for
man
ce testi
ng modeling
Copy
ri
ght
©
Q
uali
ty
Wize™,
A
lon
L
in
et
z
ki
alo
n
l@
q
u
alit
y
-
w
iz
e.c
om
Lo
ad &
P
er
for
mance W
or
klo
ad
Modeling
Modeling
enables
us t
o
simulat
e
lo
ad
Lo
ad si
mulation
ena
bles us t
o
extr
ap
ola
te
and
ev
alua
te
pr
oduction
lo
ad
in
simila
r
or
expecte
d
conditio
ns
W
orklo
ad model is
built c
ollecting
the fo
ll
owing
data
it
ems
/ pr
ep
ara
tions
-
…
Copy
ri
ght
©
Q
uali
ty
Wize™,
A
lon
L
in
et
z
ki
alo
n
l@
q
u
alit
y
-
w
iz
e.c
om
Lo
ad
&
P
er
forma
nce
W
orklo
ad
Modeling
KP
Is
r
equ
ir
ed
Data
b
ase
& n
etw
ork traffi
c
v
olume
s
–
wi
th
exp
ec
t
ed
gr
owt
h
Applicati
on
modularity
and
sc
al
abil
ity
P
arall
elis
m
cap
abili
ties
(fo
r
eign
pop
ulatio
ns
)
Us
er
pr
o
fi
les (Us
er
Co
mm
un
ity
Mo
dels
)
–
iden
tif
ication
and
cr
eation
P
r
o
fi
le
sc
enario
r
epr
ese
ntatio
n (per
cen
tage)
T
es
t En
vir
onm
en
t
& c
onfigur
atio
n
–
p
r
odu
c
tion
lik
e,
ab
le
t
o
be
extr
apolat
ed
W
orklo
ad
(pr
oduc
tio
n) p
att
er
ns
identificatio
n
T
est en
vir
onmen
t
mec
hanism
s
(cache
manageme
nt,
netw
ork
bo
x
es
bu
ff
er
s m
anage
me
nt,
et
c
’
)
And
mor
e
…
Copy
ri
ght
©
Q
uali
ty
Wize™,
A
lon
L
in
et
z
ki
alo
n
l@
q
u
alit
y
-
w
iz
e.c
om
How early
should we
te
st
per
forma
nce?
Copy
ri
ght
©
Q
uali
ty
Wize™,
A
lon
L
in
et
z
ki
alo
n
l@
q
u
alit
y
-
w
iz
e.c
om
How ear
ly
should w
e
te
st
per
formance?
Com
mo
n
belief :
“…
pe
r
for
mance
t
es
ting ca
n
’
t
ef
fe
cti
v
ely
star
t
until
the
ap
plic
ation
is
stab
le
…”
An
ex
pe
rienced
pe
r
for
manc
e t
est
er
(and for
some
tas
ks
a de
v
eloper)
can
do a lo
t
o
f
thi
ngs:
NW
and/or
W
eb
ser
v
er
th
r
ou
ghput
li
mits
Indi
vid
ual
ser
v
er
r
esour
ce ut
il
ization
un
de
r v
ari
ou
s lo
ads
Sear
ch spe
ed
s,
quer
y
optimizat
ion,
table
/r
ow l
ocking,
and
spe
ed
v
ersu
s
cap
aci
ty
mea
sur
eme
nt
s
for
DBs
Spee
d cost and
r
esour
ce
cost
o
f security
measu
r
es
Measur
e and
ev
aluat
e
S
wap
rat
e,
I/O rat
e
for ce
r
tain
pr
ocesses
and
AP
Is
An
d
mor
e
…
Copy
ri
ght
©
Q
uali
ty
Wize™,
A
lon
L
in
et
z
ki
alo
n
l@
q
u
alit
y
-
w
iz
e.c
om
Which
r
esou
r
ces/
skills
ar
e
inv
olv
ed
in per
formance
te
sting?
Copy
ri
ght
©
Q
uali
ty
Wize™,
A
lon
L
in
et
z
ki
alo
n
l@
q
u
alit
y
-
w
iz
e.c
om
R
esou
r
ces
inv
olv
ed in the
per
formance
T
esting
…
T
ypes o
f
exper
ts tha
t might
be nee
ded:
Net
w
ork
, D
Bs ,
OSs, T
echn
ol
ogi
es
exp
er
ts
(W
eb,
cl
oud,
mobil
e,
big
-
dat
a),
Har
dw
ar
e, Siz
ing,
App
li
catio
n
…
Schedu
ling
them in or
der t
o g
et their input
is
cr
itical
–
their
cost
is
v
er
y
high
t
o the
pr
oject;
they a
r
e
not
needed
100
%
o
f
their
time
Copy
ri
ght
©
Q
uali
ty
Wize™,
A
lon
L
in
et
z
ki
alo
n
l@
q
u
alit
y
-
w
iz
e.c
om
Summa
r
y
Intr
oduc
tion
Benefits
Ri
sks
and c
om
mo
n
mistak
es
PT R
equir
emen
ts
PT
Measur
ements (KP
Is)
How early
?
R
eso
ur
ces
& scheduling
Copy
ri
ght
©
Q
uali
ty
Wize™,
A
lon
L
in
et
z
ki
alo
n
l@
q
u
alit
y
-
w
iz
e.c
om
The need
for Sp
eed
…
P
er
for
mance
&
Lo
ad
T
esting W
orkshop
Alon
Lin
e
tzki
alo
nl@
qual
ity
-
wiz
e.c
om
THANK Y
OU!
The Need for Speed - Performance Testing Challenges