Microsoft Word - WebServerModeling-Project-2-2.docx

(1)

" # $ %" &

" ' (

**)*%**

2.3

$ '. / ) # $ " 1391

(2)

!

18 /

4 /

91

24 /

4 /

91

2.3

0$ 1 2 3 4 18 / 4 / 91 1.0 18 / 4 / 91 1.1 ! " # $ %& '( ) * & 19 / 4 / 91 1.2 $! ' ) * & $ , - ./ 19 / 4 / 91 1.3 0 12 345 $! ' 6. 7 - ./ 21 / 4 / 91 1.4

) * & - ./ 8 & 1 ! ! &

21 / 4 / 91 1.5 : ; <& , - ./ QBD MATLAB 21 / 4 / 91 1.6 ) ;> , - ./ Arena 22 / 4 / 91 1.7 '( 6 " ) * & , 22 / 4 / 91 1.8 6 " @ A& '( 22 / 4 / 91 1.9 $ . B ! &C $ & ! " # D E < F! G E > E $ H& 23 / 4 / 91 2.1 4I 4I J 6 K ) * & E < L H ! 2 345 23 / 4 / 91 2.2 M F! G N E /O - ./ 8 & 1PI! &

24 / 4 / 91 2.3 E /O

(3)

List of abbreviations

MMPP Markov Modulated Poisson Process

MAP Markovian Arrival Process

BMAP Batch Markovian Arrival Process

QBD Quasi Birth Death

ML Maximum Likelihood

PH Phase Distribution

Hadoop _{$ I}_-_{6> G ! 2 ) & $}_{# ! G} _6@ ₁ _{& Q (}_6P"_J ARENA _{M & H " C ( :! 5} _# _{) ;>} _{! '( Q}

SPNP ₆ _{" $ /;> - P<} _{) * &} _{! '( Q}

SHARPE _{... S5 NSI! & N} _{" $ /;> T - P<} _{) * &} _{! '( Q}

(4)

: 9% 87 7 5 6 ... ... ... ... ... 3 1 -$ # " ... ... ... ... ... ... 4 2 " 4 9 @ ' $ # " ... ... ... ... 6 2 -1 --U & 3 V& ! * 2009 ... ... ... ... ... 6 2 -2 --U & 3 V& ! * 2011 ... ... ... ... ... 7 3 -$ # " ... ... ... ... ... ... 9 4 -$ $ 9 C" ... ... ... ... ... .. 11 5 -' $ " !78 " / ) ' ... ... ... ... ... 12 6 -( ' " @ ' . $ E 5 ... ... ... ... 13 6 -1 -0,H& $ ( J 6 ... ... ... ... ... 13 6 -2 -* & ) C ( $ ! ... ... ... ... ... 13 6 -3 -* & MMPP ... ... ... ... ... ... 14 6 -4 -* & MAP ... ... ... ... ... ... 15 7 -G % ' $ " ... ... ... ... ... 17 7 -1 -: DK4& &) ! ... ... ... ... ... 17 7 -2 -* & ) MAP 6 K J 4I ... ... ... ... 18 7 -3 -* & ) MAP 6 K 4I ... ... ... ... 18 8 -/ ) ' I ' ... ... ... ... ... 20 8 -1 -W ! ; <& 1 ;> -X & ... ... ... ... ... 20 8 -2 --K ) E ;> -X & E % ) ;Y ! ' SMCSolver ... ... ... 21 8 -3 -W ! ; <& 1 MAP ... ... ... ... ... 22 8 -4 -;Y ! ' MAP QN & * ) P< -/;> $ S5 MAP ... ... 22 8 -5 -E % ) KPC-Toolbox ;V Z $ & ! " 8 ! . $ ... ... 25 8 -6 -;> ) * & ! Arena ... ... ... ... .... 27 8 -7 -;> ) * & ! /;> " ... ... ... ... 31 8 -7 -1 -;> ) * & ! ! MMPP /;> " ... ... ... 31 8 -7 -2 -;> ) * & ! ! MAP /;> " ... ... ... 34 9 -)K L ... ... ... ... ... ... 37 CM ... ... ... ... ... ... 38

(5)

9 $ 5 -/> 1 -.> 1 PI 1 ) ! T \ $ ] ... ... ... ... 4 -/> 2 -P& / 1 # $ ; <& 1 ... ... ... ... ... 9 -/> 3 -. 1 ) T 'I & E S ( ) & ) 'I & E .Y& 1 [ ( ) & ) 'I & E " ... 10 -/> 4 -.> PI 1 ) & ) 'I & E ... ... ... ... ... 10 -/> 5 -Q ' & 6 <5 L _ & ... ... ... ... ... ... 11 -/> 6 -` > ! aM _ & J > & 8 ( I' 1 > & 8 $ ) & ... ... ... ... 12 -/> 7 -;> ) /P.b [ ! ... ... ... ... ... ... 12 -/> 8 -Mc 1 Fractal ( J 6 ... ... ... ... ... . 13 -/> 9 -F ! $ * & ) ! d ;& 1 * & SI! & ( J $ [ ... ... ... 14 -/> 10 -E % ) * & MMPP * & ) e ! f & &) ... ... ... 14 -/> 11 -* & S5 MMPP/M/1 Yg 1 ... ... ... ... ... 15 -/> 12 -* & ) [ ! E % ) S5 MMPP/G/1/K ... ... ... ... 15 -/> 13 -* & MAP Yg 1 ... ... ... ... ... ... 16 -/> 14 -* & S5 MAP/M/1 Yg 1 ... ... ... ... ... 16 -/> 15 -&) ] <& h & 6 E) @ C ... ... ... ... 17 -/> 16 -&) ] <& " h & 6 E) @ C E 4Y &) <& ... ... ... 17 -/> 17 . ( J ! Randomization . $ ... ... ... ... 18 -/> 18 . ( J ! Randomization . $ ... ... ... ... 19 -/> 19 . ( J ! Randomization . $ ... ... ... ... 19 -/> 20 -@ A& 1 ) . $ SP ,& * & ! MAP ... ... ... ... 24 -/> 21 -@ A& 1 ) . $ SP ,& * & ! " ... ... ... ... 25 -/> 22 . ( J ! -& > E) &) 8 ! J & P . ... ... ... 26 -/> 23 -& ] D0 * & MAP E > ... ... ... ... ... 26 -/> 24 -& ] D1 * & MAP E > ... ... ... ... ... 26 -/> 25 --K & SP ,& " I & ! " h & ;V Z .& Q . $ ... ... ... 26 -/> 26 -.> PI 1 * & ! [ ... ... ... ... ... ... 27 -/> 27 -D : . * i & Create ... ... ... ... ... ... 27 -/> 28 -D : . * i & &) 6@ Y 8 G ! $ ... ... ... ... 28 -/> 29 -D : . * i & Process D # $ & ! " ! ... ... ... ... 28

(6)

-/> 30 -E % ) sub-model * & ) [ ! ... ... ... ... 29 -/> 31 -U' : P2 1 sub-model [ ! ... ... ... ... ... 29 -/> 32 . F! 'M j & * k S5 6 2! $ SP ,& ! ! ... ... ... 30 -/> 33 . ;Y $! ' - 7 E % ! Arena * & ) > W $ SP ,& ... ... .. 30 -/> 34 -* & S5 M/M/1 E % ) /;> " ... ... ... ... ... 31 -/> 35 -* & S5 M/M/1/K E % ) /;> " ... ... ... ... .. 31 -/> 36 -& S5 MMPP/M/1/K E % ) " 6 ... ... ... ... 32 -/> 37 -& S5 MMPP/M/1/K E % ) " 6 ! * K ! Q ! '( SPNP ... ... . 32 -/> 38 -e $ !aM 1 $! aM SP ,& * A& MMPP ... ... ... ... 33 -/> 39 -& E % ) $! aM SP ,& * A& MMPP ... ... ... ... 33 -/> 40 -* . K 2 1 S5 * A& MMPP ... ... ... ... ... 33 -/> 41 -& G 8 Y $ & ! S5 * A& MMPP ... ... ... ... 34 -/> 42 -& S5 MAP/M/1/K E % ) " 6 ! * K ! Q ! '( SPNP ... ... ... 34 -/> 43 -e $ !aM 1 $! aM SP ,& * A& MAP ... ... ... ... 35 -/> 44 -& E % ) $! aM SP ,& * A& MAP ... ... ... ... 35 -/> 45 -* . K 2 1 S5 * A& MAP ... ... ... ... ... 35 -/> 46 -& G 8 Y $ & ! S5 * A& MAP ... ... ... ... 36 -/> 47 -<%5 D # : 0,H& $! I * ! E > ! ;> ) CloudReports ... ... ... 37 -/> 48 -_ j & LV E % ( 0 1 $! ! I ! CloudSim CloudReports ... ... 37

(7)

!N

! > V 1 ! .> #$ ( & 1 I C! I 1 E 'I & ! '( $ . : l l< I .6 E > Q -U & -K ! ) E % E < m 02 ! ! n C I o $ .$ 8 l< : l Y( 6 N6 $ 8 6 I ) _ & ) _ E > ! ) I ! ! LV 'I & E #$ ( E > 6 ! V AI K E % I . o $ I 6 ' O C N6 E > #$ ( ! I ! _ & ) M! / $ 8 E) K 8 : l l< .6 * & E < 1 ! 6 E > 1Y $ c" 8 ! ! E > ! ! I ! h & ) " /& N[ 6@ ! C 6 Yg 1 @ p 0, _ & ] 6 % I q%K . ! C #$ ( H d 6 @ ! 1& * & I $ MAP 1& * & 1 2 ! ! [ D ! I! $ ;> $! ' ! I * & ! & ) . M ! 7 E % ! 1& I 1%P ,& $! ' F! 'M 8 * & MAP . ! ) $ Q C E % ! & l& 6 E > L H 1 ! G ! 7 W E > $ . 8 I MAP * & " ! N6 E > Q H& 8 G & r ;V * & @ MAP 1& H ! [ J ( . $ ) * & E < 8 n.$ &) f & J ( ) E % ) ;> Arena E > E H " /;> ' .6

(8)

1 -" # $

P& / 1 & I 6 M : ; <& 1 0 b b 8%P ) M Nr N[C ) ]" ! C 1 # " s ( O ! . . 8 K N1 Y ! I 1 & 1 B I H $) / J ] I ! ! 7 ! < GO -& E 1 C N > 1 . . b & $ 1 ) # ; <& $ 1 E > U ! Y I 6 1 O ! 8 & 2 1 2 . $ U! ! I ! 1 ) 8 # ; <& $ 1 : ; <& :) ! ;b 4I : ; <& N ! ! ! . ) #$C ( ) _ & M I N62 & 1 > & J.I 8 -$ ) & b NE > /;> J F ! " - M! 7 E % ! & NE 'I & J ) 62 & : & 2 SP ,& T U ! o VY - 7 ; < 1 2u E 2 ! ) ! . > > * K O 8 > s ( 8 ) _ & 62 ( VY - 7 N1 ! / ! . ! / * / GO 8 _ & G ! / SP ,& $ ! I -U & -K ! # '( h; I C! I 1 ! . H -U & ! / Q 8 GO I 6 8 ) 62 $ ;@ Y ! # C 6 _ & ) E % ! ! E @ AI K . ! V .$ I ) -/> 1 6 j ; - 7 ' & v #$ 6. $ N 1 ! aM p,H& h; Pw & 3 k I O . > &) $ H ) Y 62 & 1 ) Y N > 62 @ &C #$C ( h; -U & & 1 . > O!/ 1 -" # $ ' 7 $ 8/ P$ E )$Q R

N_ & `! ! l D ] ; H" 1 & $ 6 Y( 6 .$ NE 'I & > X!'

(9)

1& ! $! E C ! I ! I! 1 " ) 6 % I 8 & _ & _7 & p 0, ! D & >

] P. ) . > 1& Q ! D & 8 .$ 1 @ ! I ! D ! & ] ! NE 'I & ! d ! $ * & ! D & 8 I 6 [ @ ! ) 1& ! 7 1 ! ! & F! 'M 8 ! ! [ ! I . M P. ) $ c" * & m 02 ! I I 1 $ & 6 E > Q ! [ ! I! ) * & ) ) E % * & MAP J ( * ! 2002 ) E % MMPP [ J ( ) * & * ! 2003 I E! > ] 4 16 17 [. E > ! 2 I 1 $! I ) 8 n.$ ) E % N6 MAP * & J ( ) ] ! 2 I 6 " 1 ) .6 ( M ! 7 1 ! ! & m 2 ! V F! 'M 8 ! I 6 E > U ! [ 1& E > Q $! I * & ! I! ) Bursty ] ! 18 * & 1> ! ! [ ) ! I! & $ O \ ] ! 19 . I E! > ' [ F ! T N C ! E > Q : l l< T ! 1 ! ) ]" & ! * & d ! $ ) F ! ! ) ]" # $ 2 ! 7 1 ! ! & ! 6 ! [ J ( ! [ J ( ) * & MMPP MAP 4& 1 ! N .62 " # $ 2 C ) -5 K d ) E " : DK

(10)

2 -4 -4;

9 '

@

# $

"

1PU & ) . > >C ! E > ! ! I * A& E GH ) 1 $ $ * A& 8 .6 E > E ! C 3 V& & W I $ 1PU & j & $

NSF ( M ! 7 1 ; H" ! & E GH W .6 E > ! I $ C ! 3 V& $ 2 -1 -( S < OT % 2009 * ! 2009 & o k ) NSF K 5 ! ; b !\ P & 14 Ei " E GH ! @ C $ 2 ! Ei " E) K . > E m 0 - P< 1l ;V Y V& N 0 F) " j & * 8 ! $ E h 6 Y Y V& NX!' B l& $ 8 ) Q I $ . 1 Q i 1$ 14 6( ! E GH &) 8 ! ) 3 > E' 8 E I > 6 Y( $ ] 14 :[ -E GH Carnegie-Mellon N[ <& h & 6 K S 5 : C ( ! I 6@ [ -E GH Carnegie-Mellon : > 2 : ; <& ! .Y& Y O! {/ > & . ) E % ) 8 ! ) . J ) ) G ( -E GH Florida International : P< -0 1 $ <& 6x %K L $ I J.I o $ 1V -E GH $ Wisconsin-Madison N Massachusetts Yale : l& Y V& > ) E $ P< -E l& 4I B X!' -E GH Purdue : ) Y 1 6> G Q @%& & 6@ $ I 1 # $ l& B X!' b . Q Pb ! I 1 /;>) b . $ 1 N( # 5 $ NE $ & 1 /;>) ! I ) @H ( . > > ! I ... -E GH California-Irvine E 6 % I '( : -E GH California-San Diego / San Diego Supercomputer Center F) " 6 & : b . & -> 2 ! X!' Y O E $ ; <& $ 1 l& B X!' -E GH California-Santa Barbara o M F) " Ei " : 62 ) Y ! E GH 8 E ; $ -$ Q 1& I ! '( 6 2! ! I :! 5 b . & ! $ M E $ } " X!' ! o .6 'I . & $ -E GH Maryland-College Park ) 8 1P" E E b I 1. & 8 > & Ei " : _& ! I 1 $ G $ ( O 1 l # > E N . ! C -E GH Maryland-College Park # ! G Y : .Q i h Y - - P< ) & $ & I #$ .H 1 J $ # 6, )! 8 -& I Q i 1 J ! E -.Y ! @ O o x ) ! I P< . -8 E B $ ' @ C " E n 6 . J ) E J h

(11)

6 & -.Y ! 1& : b4k $

J ) hP~ I > 62 ) ! - 7 B l& '(

. ! - P< N) ! & & 1 ; <& -E GH Massachusetts-Amherst Ei W ! E! ! H 8 2 &C :

8 & ! L 5 :! 5 W ! 8 . $

; ! Ei " 8 NW ! 8 SHI .6 f & 4& I 6 E H p,H& I ! 1 ) d ! . I $ 2 J.I [ -E GH Virginia J / Y : X!' 1 $ E G " J ) I 1 0 .> $ 1 . E P & K * A& 8 ( 8 4 C 0 $ #HO \ 1 •( 6 % I J ! 1 ' O / ) 1 $ 1& 1& E % ! 2 . I -E GH Washington # ! G Y : .1& 0 - P< 1 N ! aM 6 @( $ -E GH Washington and University of Utah

Hl @ 62 ) J 62 : 1 ! $ .1 ; <& W : l l< 8 _ & IBM Google D ! & $ Ei " 6@ 1& $ b $ E GH :6 ( M ! 7 B ! ) :! 0 -E > _ ) - ( # € $ ) 8 & ! '( Q 1 ;& 4I HDFS 1 & * & ) I 6> G -. I 1& 1 ; H" $ I -W I ) 8 & ! '( Q IBM J.I & Y 6@ H I 1 $ 4I 1 $ ! '( Q 8 .6 E > 1K k I 1& ! € $ Eclipse 1& ! I . I -1 & ! '( Q W _ & " @ • ! & N6 & IBM Tivoli . > 1& Q 2 -2 -( S < OT % 2011 * ! 2011 1l( 8 ( / & 6 NSF ! ) E % m 02 ! ( / & 6 Ei " 6 NZ( 8 Z;k .6( M :! 5 $ Ei " W I 1 $ NSF ! 2011 8 8 ( M ! 7 6 .K ! & ] > 1& 3 > 15 :[ -E GH Cornell ! a" B l& . b : -& Craig Venter .Q Y ! $ 8 w " 8 -& Y Q' /& : U' @ 62 > : -1 E GH New York &) ;@ : ( ! ! `! E Azure /P.b ;@ & 8 G E $ -E GH California I E " E ; F) " M! / : -E GH Colorado Boulder ! ! ! $ ! I Y 1K k 1 O! @O Y : -E GH Michigan * & \ " : /;> ) E % ) $ SHI N[ B l& ! 1 ) $ : b4k ‚ ,

(12)

-E GH North Carolina at Charlotte $ I h 1 " :

DNA 1 I Q i !

-E GH South Carolina Research Fund # 6 & : Q G $ ! X!' B l& ! ! ' , C $ & ! $ D B K v f '( $ [C Y. 6 C _ & ! 1 -E GH Southern California 1K k : O! @O 1 & Y $ * & SP ,& $

(... a" B l& N1; I N1& .b N15 02) 62 ) -E GH Texas at Austin : 2u E . b -7 K ! E -E GH Washington : ) $ * & Y 62 $ 1 V ! E $! ) /& ! 7 6. ` > NE ! aM C ) E % E ! aM ' 2 /& I _; & O ) E h ... @ C -E GH Virginia Tech : ) : ; <& Q 1 G 8 -E GH Virginia Tech : O! @O U ! 1 ] " 1 1 _ & ) E T ) E % /& * & ) E % SP ,& " $ 1 '( " 67 1 " 6@ ! E % - 7 . 1 i _ & ) E M v & E % @ a" .$ o $ I n C 8 l< : l Y( 6 N6 $ 8 ) _ & ) I 6 _ E > ! ) ! I E > #$ ( E 'I & LV ! . I E % AI K ! V 6 8 K ! 1 I 6 ! & ) ! I $ 1 K . 1 W E > #$ ( ) J # M N I E % -& I ! V K 8 N > ( M D ! T g & 1M n " ) E % 8 E ; H & B K 1 > . 1_{Forecast as a Service}

(13)

3 -" # $

! @ 62 > ) ) N62 $ G 1 1P& / # ; <& $ 1 I ) & 1 ) # # E G ! ! C 8 G # ,H $ p $ # M . Mainframe $ * - Y ! # -; <-;&-; $ 1 G D ! # & N J # ! & X!' # k ) ! I I Z J & * K C 1 " & 1 ! & . I 8 # >) " > /O I $ 1 H ! :! 0 0,> $ ! 1 2 ! ]{ . ( M ! 7 ! I .$ 8 b . & * 0 I > #$ ( /& ) 8 # NJO I $ /;> >) " 1 H G, " . #$ ( $) >) " 1 H G 8 & . > $) >) " 1 -/> '( E '( * K ! ) # ; <& $ 1 7 M!' . > B K 8 Y ) ) 8 5 0 2 :! 0 $ /;> 1 ! 6 . ! C ! ! : ; <& /;> > -0 & #$ O!/ 2 -7 ! G % E 3 ! 8 8 & I > E $ H& P ! ! I $ 6 ! Q . ) : 7 AI ! I ! 2 E % 1. I # ; <& 1 G ! I 6( M -/> 1 . I -&) N > > $ ! / # 2 ! $! I ; <& 1 $ Y% . Q b . I 8 Y ! ) ; <& _; & 1 /;> ! JO I 6< b : ; <& " #$ ;Pk >) " 1 Db 1. _ & ) $ & . ! C ! ! ) G ! &) E 'I & $ 1 ( x Q . I 6> 6 . E H ( M! / -& I ! V @ C 8 . _ & 1 ! 5 0 2 :! 0 ! : ; <& /;> 1 ) N > ( M! / @ C x % G Y S b ! . E > 8 * K . ! $ ! I N @ C % P( O N6> #$ ;Pk /;> ! @ C ) E % /& 8 h ! / ! ( ) E % I 6( M -/> ) & $ ) 8 :! 0 ! _ & - 7 " o VY ( x Q . ) ! 7 E % ! & SP ,& $ ! I 6 C . I E % v & ! V @ ! ( 8 ! \ ! ) 62 2u : ; <& _ & ) E % /& I 6> Q E :! 0 ! ) ] J T h K ) & #$ ( 1 lK ! . ! C 6l J \ b ' 1 PI ! ( _ & 1/ ' J.I ) 2 ) & -& /-& ( ) 6 " ( _ & 1/ ' & -5 K 1 > .

(14)

O!/ 3 -8 $ ' R U V: ( X " ' K ' $ U : 8 & ( " ' K ' $ U I $ ! ) ) N62 M # ) $ _ & r4k ) & E > 1 M b . & I ) ! I ! ) & $ G /

& -& > $ -0 & 1 . > 8 ! " ! V $ E > g b b J O _; & ; <& 1 O! {/ LV : l( B ] & U ! 1 . > 8 : I a& Q * k ! : l( ] M $ o 0& & ! 7 M I 1 M . ! Y 1 ! " - $ 'I & ) N ! I U ! ] > & ! : & 2 $ 8 " :! 5 E > /;> ) & $ N M I ) 8/.& U ! : & 2 M $ ! I ] & $ a" o VY ! ! I $ 6 @ H I U ! ! I ' ) $ 1 ) & ! I $ 1 ! ) E % . > > & & 1 I & # E ' > > # PI I ( _ & 1/ ' ! C J \ r ( -/>) > x 4 b ' :! 0 N ( 1 ) J LV \ & 6 & 1 > 8 h 6, _ & :! 0 N ) & -/H ! '( N] g l h K 6 @ H 2 ! ! & E ! 7 ! I 1 . > O!/ 4 -8/ 7 U ' ' K 8 ! 8 > W & #$ ( .> ! 1 I C! I 1 '( ! E 'I & $ E % . ) & N ) ) & 1 ! I N .I ! '( 6, H $ Q . 8 '& 6 Y ƒb _ $ ! I ] -$ 1 ! / :! Y J ! C $ ! ! 8 &) .6 E >

(15)

4 -C" 9$ $

O ) 1/ _ & 6 & ] T _ & Y O $ .6 E 'I & ! 1P5 $ H ! I $ 1& N > 1& H T g & 8 6 .$ Y( I ) > 6 ! ! . > Q J & :! 0 H $

_ & E ; ) # M I O .6 E I " ) ! 6 .$ T g & 8 : & 2 _ & " ! - . ! 1& ! $ ) '$ N# I E % L <5 ! V E > #$ ( O!/ 5 -C" Y 3 9$ $ 2 U: 6 Y( ) 1/ ' ] 6 % I 8 .• I 6 1 $ " ) C 6 & _ & o 0& 1 ! " 1 ! I! ] ! $ . & ! ! 8 > & E > ! ! I! # ! ! I ) & $ ! _ . 1/ ' ( ! 6 ) & $! ! C ) 1 . I -/> 6 ]6 E > E . 2 [. 6 & 8 :6 1 ! - 7 6@ ) ] 6 % I 8 .• 6@ _ & 1 -.# $ p 0, 1/ ' ( $ 8 > & ! ) & $ 8 > & GO

2 -8 > & ! 1/ ' ( -8 > & J _ & GO .# $ p 0, ) & $

' ! 1/ ' ( 8 > & ! 6 # $ 2 N# I 1 " ! ) & 8 > & $ E ! ! I ! * & # M .# I [ K @ C ! ) * & 8

" ! ) & 8 > & $ ) ! & _ & I J.I & & ) # I 1

C ) &) ! - . ! N ! ) _ & ) [ ! I 1 &) ! 6 8/.& * A& .# $ m 0 2 @

_ & .]/b > > ) .I &) ! Q I $ _ & p 0, E &C 6 o 0& * & B 8

a" Q ) ! & .

M * K T . & ) ! & _ & 1/ ' ( ! J ! l & ) & $ 8 > & W

: @& $ J / ) 1& N > E > > aM ` > _ & ) E )

! J ) & 8 > & J * l

I E % G 1/ ' ( .

(16)

O!/ 6 -C" ZL [ / )" \$ ] / )" U ] / ' K E

5 -' ) / " !78 " $-'

n, ! ! ) o 0& G ‚ , N! /P.b 1 )! E % - 7 E $ J / ) 1/ 6 ! /P.b I " 6 1Y7 6 K * Y& * & J

. & & F ! 8 ) E % 1 ! !

" I " 1& 1 C! I '( & I $ & !

. I J.I > /& 8 ) ;> ) E % 8 n.$ ! I SP ,& ! I $! ! ! # ! (! $ e! E C ! 6 8/.& I 1V > ! 7 1 ! ! & . . ;> ) 1/ ) E > E ! I ! E > Q $ Q W I # ! '( Arena ! 6 E > Q -/> 7 .6 E > E H O!/ 7 & !78 ' ) / F ! & N " $ /;> ) E % 4A& . ! C ! (! ) ;> # ) * & G $

1PI ! V . > * & o 5 ) &) I 6 8 E > ) ;> * & ) E % -/H&

(17)

;> * & E > ) I E % 1 ; <& $ * & ) 1g ! V ! J I 6> $ 2 /& 8 1 " ! m 2 6 Yg J ! # 8 ( M ! 7 * . K # $ Q ! ) ! & : ; <& v ) I ! .# I 1& & ! Z 7 1 ! 8.g # $ 2 * & N[ 5 02 6 J ( ! (! E > ! ! I ! ) .# I 1 " ! C /P.b 6 Yg # 6 @ ! # ) { [ ! J

6 -5 E $. '

@

" ' (

6 -1 -9 $ \ E )5* _ ) J E > Q $ 1 ! Z;k 6 J ( _ ) Heavy Tail -& > E n " ! (! ! I 6 $ G Long-range-dependent N Fractal N Self-Similar Bursty > 1& ! $ G 8 . I -/> 8 H 6 E > E . O!/ 8 -L^$ Fractal $ \ 9 6 -2 -' ( E $. # E > ! ! I! G W <& # -& Y I 6 C ( W 1( 0 $ C ( ! $ ! 2 1& * & . > $ _ ) N6 J ( m 2 _ ) 1( 0 1. b * & 1 ! ! C ! (! ) & I * & 1G I # ! 1 $ 8 1 &) $ ! . * & 8 ) . O . I * & ! $ $ ! C ( -& > E > 1( Y& : l l< ! I MMPP N MAP BMAP 1& > E ) * & 1 ;& I . $ SI! & 2_{Arrival Process}

(18)

O!/ 9 -' ( E 1 _$ V ( " & E \ " ! I ! V .$ -/> 9 * & 6 E > E H MMPP 1& " F ! J Y e f & ! $ F ! I * & ! &) MAP 1& I 6 C ) 1& .b 6 K E n " $! (! 1G 1 &) 1 7 & $ ' ! * & . I * & BMAP ) 1& .b 6 K J ' MAP 1& I 6 6 K ! $ * & ' ! * & L g & ! . I MMPP MAP .62 " # $ 2 6 -3 -( MMPP * & MMPP * & SI! & E ) B ! I " E F ! J

6 K ) f & e ! $ 6 1& p,H& ! e J E $ H 6 Yg $ C ! I 1 Yg * & * A& . > -/> 10 D ! ! ! e G %5 6 Yg I G ! e G J 6 Yg ! e N6 Yg f ]" . > . I $ 2 f # ! 6 Yg f 8 ) -5 K E > E $ H& $ . ) 1 A& -/> 10 .6 E > E H O!/ 10 -( ' @ MMPP X ' " ` a ' ( "

(19)

S5 * & M * K MMPP/M/1 ! E > E H 6 Yg 1 ;& -/> 10 H * & N# I # ! ! ! E > E -/> 11 $ 2 6 Q b I &C 6 Yg $ ! $ . Y G . > 1& # ! & O!/ 11 -V ( MMPP/M/1 4 ) E % E > Q $ ) * & . ) * A& J Y MMPP ] & 4 E! > [ J * & I I S5 ) E % ! <& MMPP/G/1/K ! 6 Yg 1 )! 1 ! ! & )6 E ! 7 -/> 12 .( O!/ 12 -V ' @ " & ' ( MMPP/G/1/K $ S5 -K E < ! & ! MMPP ;> * & 1 ;& I -! $ X & L g ) E " 6. 7 .6 E > E 6 -4 -( MAP * & MAP ) 1& .b -/> J MMPP N6 6 Yg I ! e f C ! I 6

. $ 1& e! $ 6 Yg 8 m 2 !aM J &) ! ! J ! C ) . -/> 13 6 Yg E > E H ! .6 * & 8 ) Q @%& $ E ) E ) ! .6 E > E % @ " SI! &

$!aM ) M @ " SI! & $

1. E $ H& # ! v N > ! ;b @ " I 1 &) >

(20)

O!/ 13 -( MAP 4 6 K ! MAP H 6 Yg $ 2 1 ' & E $ . E $ H& 8 $ . > 1& E $ H& - 7 !aM J ) M I 1 &) 6> # $ 2 G 2 ( 5 N# I ! ;b @ " !aM J ) M > $ 2 ! . J N# I ! ;b # E $ H& - 7 !aM J . > E $ H& . 8 " ! E > E $ H& $ . ) 1 A& -/> 13 E H .6 E > O!/ 14 -V ( MAP/M/1 4 S5 * & MAP/M/1 1 Yg ! -/> 14 .6 E > E H Q b 6 Yg $ ! Y E $ H .6 # ! & $ .

(21)

7 -" $'

G %

! [ X\ %$ J j & 1 )! ! & # Apache J / 1P" E GH Milano I 6 &) E) ! : b4k 3 ;& { 2006 6b 4:27 1 L;5 10 6b ;& { 5:53 -& > : b4k 8 . > 1& L;5

/& 8 E 4b . > 1& 6 h K 6 4I E > E M „1> E) v J 67 6 2! $ &) <& p ,H 3 " 4 ' . ! $ 6 2! I I s ( & } " 1P5 D( K ) . l & 1& E 6 2! / ! 5 ! > $ $! " back-end * ! E E G " 1 F) " 6@

! & . ! I 6. E > E M „1> E) 6 h & $ 6 2! ] &) . > 1& J<& B E > Q $ $ c" Z;k I 6( M D ! ' ! <& &) " $ 6 2! m 02 TPC-W N _ ) ) <& &) Lognormal 8 G & 3.275 ] ! v 1P & 11.56 . I 1& " O!/ 15 -P$ X ' 3 . ' " 9 ; %$ O!/ 16 -P$ X ' 3 . ' " 9 ; $ I 3 : X ' " 7 -1 X ' bc N > E! > I ! V .$ $ X\ 67 ; <& ! %5 ) Y & 8 . > 1& 6;v v J

J ! 6 2! 5 1 K $ E N\ J ( ! 6 8/.& I 6 1 K ! 8 . > 1& $ . ! 8 &) :! 0 ! G & ! 7 v J ! I 1 $ 6 2! PI & NT g & 8 ) -5 K ' $ I .# > & v ] 8 62 / 1( 0 0N1 # I 1& _ ) [ ) -/> 17 ( I > 1& Q 1 . K _ ) & ! I 8 . $ & ! " C h & E 1g ! $ * & ) E % . > 1& ; <& 3_Static 4_Dynamic

(22)

O!/ 17 \ ) 8 Randomization ) 8 E 7 -2 -" ' ( MAP ) \$ 9: c " 6 2! .$ N 4I J 6 K ! 8 $ &) ) * & ! > 1& ( M D ! T J ) $ ! ' ;Y ) E % I ( " $ 6 2! ) -l &) KPC-Toolbox ! MATLAB Y , ! I N6 E > E L g _ ) MAP . M 1& ; <& @ C h &

7 -3 -" ' ( MAP ) 9: c "

E > ( M D ! ' & " $ 6 2! N 4I 6 K ! 8 $ &) ) * & ! . _ ) B4I $ ]{ PH F ! ) E % C Maximum Likelihood M 1& ; <& ' ! I 8 I ! ' ;Y W KPC-Toolbox ! MATLAB _ ) _. I ! w 8 8 ( M D ! ]{ .6 Q - 7 * Y& ) ( MAP 6 K N > 1& * & > 1& _. #$ ) :! 5 MAP . C 6 • • ,

W E &C 6 * & l& MAP F ! E % ) ;> ) ! 8 &) G $ ! d I 6 E > -/> 18 .6 E > E H I 6 Iu Q)\ K . > 1& B4I Y E $ H :6 E > E % ) V ! ) I 6 E > ( M D ! S5 * k N V2 ; <&

(23)

O!/ 18 \ ) 8 Randomization E ) 8 1& H 1 .g ! k d 1& .I V2 E \ ! F ! 67 N] B4I '( I $ . > ) ]"

! I ( M ! 7 1 ! ! & ] &) I ! 1 @ S5 * & N $ ! 8 &) 67 l& -/>

19

! SP ,& B4I %$ ) ! N1Y7 $ . E &C 6 * & Z ;V ' & .6 E > E H <& ) E " m 02 ! H : U' .6 E > E H …O 6. $! ' ! I SP ,& $ 6. 7 ; .6 E > ! Y , ! E % ! & O!/ 19 \ ) 8 Randomization E ) 8

(24)

8 -'

I

' ) /

E " , 8 ! ;> ) , ) j & $! ' 1 ! ! & # SP ,& $

C - P< .6 E > L H $ * & ; <& E < , 8 ! 542 ! V QBD ; <& E < N MAP I " E < MAP . B h & ) E % E > E $ H& $ MATLAB * & E < 8 n.$ .6 E > E H ) E % f & ! ! [ ) ;> I ! '( Q Arena /;> E % ) * & E < ' $! '( Q ) E % " SPNP SHARPE .6 E > E H 8 -1 -: ) / 3 d" -e -/> 11 E ) ) 1 K $ -X & ;> C I 6 X & 5 QBD SI! & $ E ) . > 1& %M

S5 E > E . $ MMPP MAP . E ) ) 1 $ QBD :! 0 C !aM :\ . K ] & . $ ] > 1& ) 3 :[ ! " " " " " # I : > 1& . - 7 ) :! 0 C ! & $ ` P G $ % & & ' ( & ' & ₍ ' ' ' ' ' ' )

E ) 1PI 6 K H& ; l C 1 ; <& 1PI W ! 6@ I $ 2 ; <& - 7 b X & $ : < g H 1& ] 3 Y & [ . . • * & • * * * *( + • *_, *_, *_, *_{, (} + • *_- * . -• (/ . / .( & • * 0 / . 1 & • * 02 .13 _{' 4} _&

1& $ E ) 8 j & : ; <& Q D & $! ' ) MAMSolver SMCSolver E % ) $ I I C Matlab ) E % E > Q : ; <& . . $ & SMCSolver ! Matlab

(25)

- ( .6 E &C & ! * A& J ! ! ' ;Y 8 ) E % ) ! & $ Matlab 1& ! ) ] 5 [ ZPY & E GH {& I Q Pb 1g ! E /H ! : l l< E M J €! J cP I 6( ! ! W; & :\ l& I ! C ! ' ) . 8 -2 -: ) / K ' Oc -e U" ) M ' @ " SMCSolver ] ) * A& 8 b l& 3 $ ] & . > l& - 7 C d 6<5 6 E > [ , [ B0 A0 A1 A2 D ! ) :! 0 ! : G B0: A0: A1: A2:

f & ! I ; <& ! ! " 6 Yg & ) : ! W ]{ pi : ! 7 >> [G,R]=QBD CR(A0,A1,A2); >> pi=QBD pi(B0,A0,R); * * * ( * 5 * 6 * 7 * 8 * 9 * * * ( … 0.0381 0.0675 0.1350 0.4500 0.0282 0.0412 0.0414 0.0900 0.0158 0.0211 0.0112 … ] ! E > E H d I 3 <%5 [ 140 .6 Z;V & 8 ) E % m 02 ! H : U' E $ H& 6@ ] & ! ' ;Y 12 . I Y & ' [

(26)

8 -3 -3 d&#-34; MAP F ! m 02 ! MAP :! 0 / QBD N @ " $!aM & - > -K ] & :! 0 !aM :\ . K :;< / C ! I > D0 @ " $!aM j & D1 $!aM j & & ) :! 0 W ! ]{ .6 ! />C . > ; • *_{:;< :;<} & • *_:;<' 4 4 • = *:;<' ' 4 • * 4 >? @ 8 -4 -U" ) M MAP QN " V E )! / O 73 ' ( MAP ! ' ;Y 8 W ) I ] 6 [ N6 6( ! - 7 1& ! 1%P ,& : ; <& S5 $ MAP ) ! & P. ) : Q • SI! & $ E ) -K MAP • _ ) # D MAP h & • 1G ;.$ h g j & $ ; <& MAP • !aM ] & MAP •

! 8 $ &) ] ! 8 G & ; <&

• SP ,& $ * A& ! $ . ) MAP • - ; MMPP MAP • ... ( f 1 : 6 K ! 8 ! I . ) ! & N l& 1Y7 1 Q b H 1g (

! ] &) W & .# > > 1 . ! &) #$ 6H" S5 I s ( .6 E > E S5 $

J ' 6 Y. -I . > v 10

.# I ; <& ! * & 1$ !aM 8 G & # $ , & >

>> P=[0,1;1,0]; >> mapqn_ezsolve({map_exponential(1);map_exponential(1)},10,P) 1$ !aM ' & 8 I 0.9091 -.b ) :! 0 & E % LV S5 * k ; <& . $ 2 : I >> [XN,QN,UN]= mapqn_ezsolve({map_exponential(1);map_exponential(1)},10); C ! I XN 1$ !aM N* S5 1 2 ! QN(i,h) S5 * k i ) ( ! F) " 1 7 Q h N > UN(i,h) S5 E % LV i ) ( ! F) " 1 7 h N > V ! ) o 0& ' & sum(UN(2,:)) I C 1& 6 0.9091 V ! ) S5 * k 6 sum(QN(2,:)) I C 1& 6 5 .6

(27)

( f 2 : ] &) G S5 J * A& 8 ! hyper-exponential # I 1& ( g 8 G & 7 ] ! 25 J ) ( ! ! •K * . K 0.99 . M 1& ! 7 Q S5 ) & I ) :! 0 ' ! :\ . K ] & :# I 1& S Y >> P=[0.1,0.7,0.2;1,0,0;1,0,0]; ! V # $ , & * K :6> # $ 2 8 .# I [ K C 1$ !aM H& >> Q1=map_exponential(1); >> Q2=map_exponential(1); >> Q3=map_hyperexp(7,25,0.99); >> [XN,QN,UN]=mapqn_ezsolve({Q1;Q2;Q3},10,P); ) 1$ !aM e XN :! 0 ( 0.6465 . > $ 2 ( f 3 : C ( J Q S5 M -;7 * A& ! MAP N) ( skewness=10 RMAP(1)=0.35 D ! :6> # $ 2 N# G >> P=[0.1,0.7,0.2;1,0,0;1,0,0]; >> Q1=map_exponential(1); >> Q2=map_exponential(1); >> Q3=map_mmpp2(7,25,10,0.35); >> [XN,QN,UN]=mapqn_ezsolve({Q1;Q2;Q3},10,P); 1$ !aM e 0.5122 . > $ 2

: Q ! ) ! & & 6 E % - 7 I G #@& _ P. ) 1G ;.$ _ ; <& LAGS D ! & map_acf(MAP,LAGS) 1Y. _ ) _ ; <& map_cdf(MAP,POINTS) • \! C ( J - ; MAP map_erlang(MEAN,k) " C ( J - ; MAP map_exponential(MEAN) & ] & MAP map_infgen(MAP) ! e 8 G & ; <& map_lambda(MAP)

! 8 &) 8 G & ; <&

map_mean(MAP) - ; MMPP(2) MAP map_mmpp2(MEAN,SCV,SKEW,ACF1) ! 8 $ .& ; <& map_moment(MAP,ORDERS) B 1( 0 $ . MAP map_sample(MAP,NUM,S0) ] ! ; <& MAP map_var(MAP)

(28)

-;7 * A& N1( 0 $ . _ ) E % # $ , M $ . ) 1%P ,& ! Q S5 N# I ! SP ,& $ 6 K d I # I E % ) ! ) # & -/> 20 .6 E > E H >> Plot(map_sample(Q3,100)) ([ 1000 . . 100(S ( 100000 . . 10000(‚ O!/ 20 -( f V7 * E ) 8 : ' $ E ( " MAP l& W E > $ . MAP _ ) B ! . Y 8 .$ N " ! 6 K " & ! " 8 G & # I 1& H& ! d I -/> 21 ! e H & E $ H& I .6 E > E H E > * & MAP . ! [ J ( H 1/ ' >> Plot(poissrnd(10,[100,1])

(29)

(S 100 . ([ 1000 . (‚ 10000 . ( 100000 . O!/ 21 -( f X I ( " V7 * E ) 8 : ' $ E 8 -5 -' @ KPC-Toolbox E I h < " ] " E ) 8 ! ' ;Y KPC-Toolbox ! Matlab ! &C $ & ! " I " /& 6 E > !

8 &) j & ! ! 1& #$ ( ! E b . & J I ] 7 [. ] & C 6( ! 8 . I Y & [ . J 6{ / $ & ! " # D ! ' 8 /P.b E < E $ H& demo_run.m 8 O I > 1& #$ ( # D -K & NE . . $ 1& . ! $ & ! " I I s ( * A& S

) > ! 8 $ &) -& > ! E b . & K ] & -/> 26 .( ]{ NE * A& J Y N ) : ! & * & MAP ! h & I " Q .& . >> trace = kpcfit_init(S); >> MAP = kpcfit_auto(trace,'OnlyAC',1);

(30)

O!/ 22 -) 8 X 7 \$ " ] " X ' ' " O / \ ) 8 O!/ 23 -P$ D0 ( MAP / : O!/ 24 -P$ D1 ( MAP / : $ ] & D0 D1 j & MAP E > 16 .6 E > E H ! 6 Yg ) E % :! 5 !

! I ; N I E $ H& 8H . :! 0 ! -K & & ) ! ! ! I Q 1 & -K & . > $ 2 I -/> 25 .6 E > E H >> MAP = kpcfit_auto(trace,'OnlyAC',1,'AnimateAC',1); O!/ 25 -Oc V7 * 2 X 8 " h < " ; I X I ) 8 E

(31)

8 -6 -' ) / ( " Arena E * & E < ! [ ! ) f & ! ) E % &) ! '( Q Arena E H ! * & 8 1PI 3 k .6 E > -/> 26 ) & 1P5 <%5 ! .6 E > E H * i & Create 1& ! 7 F) " ! & ! [ W I I E % [ " $ 6 2! . M O!/ 26 -8/ & ( 7 * i & ) Q I $ : . D ! m 2 G J B 6 v N1( 0 :! 0 ! e 8 Y /& $ ! G I p,H& /& ! [ X\ ! ) / - ! . ! 'M 8 8 N ! $ 0,H& I # E I [ , ! Schedule ! -/> 27 .6 E > E H Q b - I 6 Iu Q)\ l& $ & ! " $ 1 .6 E > E % 1g ( l& ) O!/ 27 -( i 8 b Create * i & ! & ! ! $ G Schedule ! I I 8 Y -/> 28 - ! .6 E > E H * i & J T $ N ! # ! T / Create $ ! G I # E I . I S Y & ' & ! V ! T $ )

(32)

O!/ 28 -E ) %" : ] 9 M " X ' ( i 8 b * i & J N! [ F) " ) * & Process Y & C : . D ! I # I 1& E ) " ] e $ _ & ! ! I I 8 Y -/> 29 .6 E > E H O!/ 29 -( i 8 b Process " E I G b

(33)

) # & N# I * & E n " ! ! # $ , I 1 K ! sub-model

# I E % I h 8

. > : % & ] T N6 2! T $ 4A& I > E > - /H 1P2 * i & O ) & ! 2 ! T g & 8 Q E < -/> 30 -/> 31 .6 E > E H O!/ 30 -' @ sub-model & ' ( " O!/ 31 -76 TUM sub-model &

(34)

! -/> 31 > " / h K ! 6 2! & I 6 E > E % # .0 * i & J )

> $ 2 #$ ( ' ! * I /& h 8 . I * ! k & , ! 1 % & ] Q & 4A& .

.6( M D ! 6 2! T $ , ) E % ) ;> ) ]" & 6 @ !

Report @>! 'M N

) ! C d ) ;> 8 K ! I ! C 6 ... _ & o 0& NS5 * k D 1 $ & ! " m 02 ! D ! & . > 1& O!/ 32 -1 UL ) 8 V7 * E 9 6 V ( )" j "

$! ' ;Y ) E % ' E n " ! & ) * & /& I I E! > 6 @ ! Arena ! @ C ) 12 I ! -/> 33 .6 E > E H O!/ 33 -@ O" k E U" ) M ) 8 Arena V7 * E d$ / ' ( "

(35)

8 -7 -I )! / " ( ' ) / ) E % W ! :! 0 K #$ ) ;> :! 0 #$ # - P< 1> ! " $ /;> ! ' ! .6 b SHARPE SPNP * & /& H& :! 0 $ I 6( M ! 7 1 ! ! &

! I : % 8 ! C 1& #$ ( ! " SPNP ! 0 ! E $ H& /& ! 1 6> ' 8H . : E > E Y ` E GH {& I r E /H ! $ / - -K E < ] 9 [ N ) * & C ( D ) . $ / : ; <& - P< ! '( Q ! ; SHARPE $ F ! W ) * & /& " /;> E 4b & D G 1 ! ... ) S5 $ /;> ' SI! SPNP . I 1& 1 ; H" " /;> ) Wl( O!/ 34 -V ( M/M/1 I )! / ' @ " O!/ 35 -V ( M/M/1/K I )! / ' @ " E S5 J 1 ! ) E " ! M/M/1 $ S5 ) * & F ! ]{ # 2 " E n " MMPP MAP S5 . .# E Q ! M/M/1 ! -/> 34 E H ]6 E > 3 E " Q G $ .[ 6 K •( -I " $ /;> ! " 6 K -K - P< / - Nr ( 1 ; <& $! ' ! * & 8 ) •( -K - 7 . ! ! C -K :! 7 N6 <& * & 8 6 K •( 62! I 1& # ! ! C

. > <& 6 K ! I ! V .$ 8 -/> 35 J '( 6 E > E H * Inhibitor S5 J * & <& M/M/1/K $ 2 - P< - 7 ! ' W I > 1& - ; I K * E > p,H& h g Inhibitor > 1& ] 10 $ * & 1 ! .[ ] > 1& @ H " 1P& Y :! 0 H 11 . > Y & [ 8 -7 -1 " ( ' ) / MMPP I )! / " S5 E > * & . MMPP/M/1/K 6 Yg ! -/> 36 6 E > E H * & 8 I

! E > ! SI! & E ) x & -/> 10 . > 1& C - ; > %M I ! V .$ ; MMPP/M/1 @ & Inhibtor -K - 7 ! ' W <& 6 K •( > - :! 5 8 ! 1 N6> ! . $ , A A k

(36)

O!/ 36 -V MMPP/M/1/K 9 I ' @ " $ 6 Yg P0 P1 $!aM E .$ T2 T3 j & J $ @ C x & $ e ) 1/ I 6 ! . $!aM T1 T0 . $ 1& H ! 6 Yg 8 8 1 ' P2 1& * & ! ! S5 ' I T4 ] e C 1& 8 Y ! . > 1& # ) ‚ 2 h; I I O!/ 37 -V MMPP/M/1/K U 2 M ( c 9 I ' @ " SPNP ! 1P& Y :! 0 C ( N8H . ! ) E % * & :! 5 ! -/> 37 6 E > E H ! ! 7 S5 ! N-/> 8 ! .6 ! I ! 2 ! $!aM I * Y( /& I e 1PY( T3 1& . > 8 /P.b j & $! . $ & ! " S5 ) ! -/> 38 1 -/> 41 H .6 E > E T0=5, T1=1, T2=10, T3=5, T4=8, K=100

(37)

O!/ 38 -( f " V7 * E ZL E ZL a MMPP O!/ 39 -( f " V7 * E ZL ' @ X MMPP O!/ 40 -( f " V X " : 6 ( 8 c MMPP

(38)

O!/ 41 -V M E ) %" ] ( f " MMPP 8 -7 -2 " ( ' ) / MAP I )! / " S5 " /;> * & MAP/M/1/K ! 6 Yg -/> 42 .6 E > E H * K ! x & * & 8

! E > ! SI! & * & -/> 13 . > 1& 4;7 I ! V .$ N > E! > ' " /;> * & ' ! 8 > @ Inhibitor S5 MAP/M/1 $!aM . > 1& - ; T2 T3 6 K ! ! e P0 $!aM T4 T5 6 K ! ! e P1 . $ P2 ! S5 T6 1& C ] e . > 1& E $ H& N ) D< 8 ! I > . ! ! 7 S5 ! 6H$ O!/ 42 -V MAP/M/1/K U 2 M ( c 9 I ' @ " SPNP ! ) $ & ! " S5 8 /P.b j & $! . -/> 43 1 -/> 46 .6 E > E H T0,T3,T5=5, T1,T4=1, T2=10, T6=8, K=100

(39)

O!/ 43 -( f " V7 * E ZL E ZL a MAP O!/ 44 -( f " V7 * E ZL ' @ X MAP O!/ 45 -( f " V X " : 6 ( 8 c MAP

(40)

O!/ 46 -( f " V M E ) %" ] MAP $ ! $ Y 8 G & & 8 n.$ : > 1& ) 3 H I I [ K ! " 6 K ! ! 6 g

********* Outputs asked for the model: MAP2 ************** Steady-state average number of tokens in P0

etok(MAP2, P0): 1.66666669e-001

Steady-state average number of tokens in P1 etok(MAP2, P1): 8.33333331e-001

Steady-state average number of tokens in P2 etok(MAP2, P2): 9.37500004e-001

(41)

9 -L )K

E > ! ! I! L <5 1 " ) * & 6 % I 8 & _ & _7 & p 0, ! D & $!

" X!' @H O ) E) & 1 @ ! I ! D ! & ] 3 > : $ H& : l l< Z;k .6 E 'I & ! 1& NF! 'M 8 ! E > E * & F ! I I M ) MAP 1& J ( h & -/> $

1& 6 E > U ! I 1 $! ' ) E % I * & ! [ J ( D &) f & ) ! & C ( -I Q 1 K ! ! * & ) E % . . I ! 1( 0 $ O!/ 47 -E 5* G b )3@ ' ) / " )" / ( CloudReports O!/ 48 E " 5 @ Y< : )" j " C" CloudSim CloudReports ;> $! ' O M W <& ) D $ CloudSim CloudReports ) -/> 47 ) E % /& ( 1& #$ ( ! ! I! ) * & 1( 0 C ( * & 8 ! E > E % 1( 0 C ( 1 I 1( 0 _ ) $ )6 62 / -/> 48 ] h & I ( * & ) E % 8 .6 E > ! $ D 1 $ MAP & ;> W <& 8 ! E > Q $ ) d Q E % ! & ] h & H 67 ! $ 1Y7 " 6 . I

(42)

CM

[1] Pacheco-Sanchez, et al; "Markovian Workload Characterization for QoS Prediction in the Cloud," Cloud Computing (CLOUD), 2011 IEEE International Conference on , vol., no., pp.147-154, 4-9 July 2011

[2] The Art of Capacity Planning, J.Allspaw, O’Reilry, 156 pages, Sep 2008

[3] Queueing Networks and Markov Chains: Modeling and Performance Evaluation with Computer Science Applications, G.Bolch, S.Greiner, H. de Meer, K.S.Trivedi, WILEY, 896 pages, May 2006

[4] M. Andersson, et al, “Performance modeling of an apache web server with bursty arrival traffic.” In Proc. of ICOMP, 2003

[5] Performance Analysis of Telecommunication Systems, University of Antwerp, Department of Mathematics and Computer Science, PATS Research Group, http://win.ua.ac.be/~vanhoudt

[6] MAP Queueing Networks, http://www.cs.wm.edu/MAPQN/index.html, Accessed at 10 July 2012

[7] G. Casale, et al, "KPC-Toolbox: Simple Yet Effective Trace Fitting Using Markovian Arrival Processes", QEST’08

[8] KPC-Toolbox, http://www.cs.wm.edu/MAPQN/kpctoolbox.html, Accessed at 12 July 2012 [9] Kishor Trivedi home page, http://people.ee.duke.edu/~kst, Accessed at 12 July 2012

[10] M/M/m/b Queue, http://people.ee.duke.edu/~chirel/IRISA/SPNP/ex3.html, Accessed at 12 July 2012 [11] Reachability in Petri Nets with Inhibitor arcs, Klaus Reinhardt, University of Tübingen http://www-ti.informatik.uni-tuebingen.de/~reinhard/petr.html, Accessed at 12 July 2012

[12] D. Bini, et al, “The MATLAB toolbox SMCSolver for matrix-analytic methods,” University of Antwerp. Belgium, Slides: http://win.ua.ac.be/~vanhoudt/tools/SMCSolver_slides.pdf

[13] Will E.Leland, et al, “On the Self-Similar Nature of Ethernet Traffic,” IEEE/ACM Transactions on Networking, Volume 2 Issue 1, Feb. 1994

[14] National Science Foundation Awards Millions to Fourteen Universities for Cloud Computing Research, April 23, 2009 http://www.nsf.gov/news/news_summ.jsp?cntn_id=114686, Accessed 13 July 2012

[15] The Sky Is No Limit: 13 Research Teams Compute in the Clouds, April 20, 2011

http://www.nsf.gov/news/news_summ.jsp?cntn_id=119248, Accessed 13 July 2012

[16] A. Riska, et al. “Matrixanalytic analysis of a MAP/PH/1 queue fitted to web server data.” In MAM4, pp. 335–356, 2002.

[17] S. L. Scott, P. Smyth, “The Markov Modulated Poisson Process and Markov Poisson Cascade with Applications to Web Traffic Modelling”,Bayesian Statistics, Oxford University Press, 2003.

[18] N.Mi, et al. A”SIdE: Using Autocorrelation-Based Size Estimation for Scheduling Bursty Workloads,” IEEE Trans. on Network and Service Management, full paper, to appear in 2012.

[19] G.Casale, et al, “BURN: Enabling Workload Burstiness in Customized Service Benchmarks,” IEEE Trans. on Software Engineering, 2012.