You are on page 1of 23

CSL491: Cloud Compuung

Abdur 8ahman Adnan


lmran Possaln Shaon
Course Agenda
Cloud Compuung
Amazon Web Servlce (AWS)
LucalypLus (prlvaLe cloud soware)
Pands-on cloud experlence
Course Coal
undersLandlng cloud compuung
1heoreucal denluon of cloud compuung
Cloud compuung vs. vlrLuallzauon
Cloud compuung models
undersLandlng publlc cloud
Amazon Web Servlce
8ackspace
1radluonal l1 ueploymenL Model
!"#$ $#&!#"'" "#$(#$
"#$(#$ )$*(+"+*,#-
./+'
,#0*1/'# $#"*!$2#
2*,30!$/1*,
,#0*1/'# 4!"+,#"" 2/"# /))$*(#-?
56 /(/+7/47#8
,*
,*
9#"
9#"
Cloud ueploymenL Model

Self provlslonlng
AglllLy
!"#$ $#&!#"'" "#$(#$
"#$(#$ )$*(+"+*,#-
56 -#3,#" $#"*!$2#
)*7+2+#"
.+':+, )*7+2+#"8
9#"
,*
Cloud Compuung uenluon
nauonal lnsuLuLe of SLandards and 1echnology
(nlS1) denluon:

!" " " $ %&'() *&+ (,$-).,/ 0-.10.2&034 5&,6(,.(,24
&,7'(%$,' ,(28&+9 $55(33 2& $ 3:$+(' ;&&) &*
5&,</0+$-)( 5&%;0=,/ +(3&0+5(3 " " " 2:$2 5$, -(
+$;.')> ;+&6.3.&,(' $,' +()($3(' 8.2: %.,.%$)
%$,$/(%(,2 (?&+2 &+ 3(+6.5( ;+&6.'(+
.,2(+$5=&,"@
Cloud CharacLerlsucs
Accordlng Lo nlS1:
Cn-demand self-servlce
8road neLwork access
8esource poollng
8apld elasuclLy
Measured servlce
Cn-uemand Self-Servlce
!"#$ $#&!#"'" "#$(#$
"#$(#$ )$*(+"+*,#-
8road neLwork Access
Cloud compuung resources are avallable over
Lhe neLwork and accessed Lhrough sLandard
mechanlsms LhaL promoLe use by
heLerogeneous plauorms.
Cen Lhrough a browser uslng xML, SCA, P1ML,
eLc. over P11
8esource oollng
Compuung resources are pooled Lo serve
muluple consumers wlLh resources
dynamlcally asslgned and reasslgned
accordlng Lo consumer demand.

27*!-
8apld LlasuclLy
1he ablllLy Lo qulckly scale ouL and
qulckly scale ln.
Measured Servlce
Cloud resources can be monlLored,
conLrolled, and reporLed, provldlng
Lransparency for boLh Lhe provlder and
consumer of Lhe uullzed servlce.
Cloud Compuung Models
Servlce Models:
! Soware as a Servlce (SaaS)
! lauorm as a Servlce (aaS)
! lnfrasLrucLure as a Servlce (laaS)
ueploymenL Models:
! ubllc
! rlvaLe
! Pybrld
Soware as a Servlce
rovldes appllcauons Lo users
no user managemenL or conLrol of appllcauon
or lnfrasLrucLure
Slmple and qulck Lo lmplemenL
Lxamples: Salesforce.com, Coogle uocs
lauorm as a Servlce
rovldes a soware developmenL
envlronmenL Lo users
Soware developmenL Lools
necessary Cu, memory, sLorage, and neLwork
resources
Lxamples: Mlcroso Azure, Coogle App
Lnglne, Appscale
lnfrasLrucLure as a Servlce
rovldes raw compuLe resources Lo users
Cu cycles, memory space, sLorage space,
neLwork connecuvlLy
user manages and conLrols everyLhlng excepL
Lhe underlylng hardware and neLwork
resources
LxLreme exlblllLy, fasLer deploymenL
Lxamples: Amazon Web Servlces, Coogle
CompuLe Lnglne, LucalypLus and oLhers
ubllc Cloud
1he cloud lnfrasLrucLure
ls made avallable Lo
Lhe general publlc
buL ls owned by Lhe
Crganlzauon selllng
cloud servlces.

)!47+2 27*!-
!"#$"
2*;)/,9
rlvaLe Cloud
1he lnfrasLrucLure ls
operaLed solely for
a slngle organlzauon.
Crganlzauon has
compleLe conLrol over
Lhe securlLy of daLa
ln sLorage and ln LranslL.


)$+(/'# 27*!-
!"#$"
2*;)/,9
Pybrld Cloud
1wo or more clouds bound LogeLher by Lechnology LhaL enables daLa and
appllcauon porLablllLy beLween Lhem
uullzes Lhe of cosL beneLs of a publlc cloud
When needed, proLecL condenual daLa ln prlvaLe cloud
Lnhances buslness conunulLy/dlsasLer recovery opuons
)!47+2 27*!- )$+(/'# 27*!-
:94$+- 27*!-
!"#$"
vlrLual Machlne
lnLeracLs wlLh a soware layer
called Lhe hypervlsor
AbsLracLs Lhe underlylng
physlcal resources
lnLeracLs wlLh a congurauon
le Lo presenL vlrLual hardware
Soware lnsLalls on Lhe vlrLual
hard dlsk
Cperaung sysLem
Appllcauons
rovlde servlces [usL llke
physlcal hosLs
<=>
):9"+2/7
;/2:+,#
(+$'!/7
;/2:+,#"
vlrLuallzauon 8eneLs
uramauc cosL reducuons ln:
CaplLal expenses (hosL hardware, swlLches, cabllng, eLc.)
Cperaung expenses (power, admlnlsLrauon cosLs, eLc.)
8euer per-hosL resource uullzauon (consolldauon)
Appllcauons remaln lsolaLed
lncreased exlblllLy ln Lhe daLacenLer
8euer workload moblllLy
Llfecycle managemenL noL ued Lo hardware refresh cycles
Lnhanced buslness conunulLy/dlsasLer recovery
capablllues
vlrLuallzauon and Cloud Compuung
Cloud compuung can leverage vlrLuallzauon,
buL vlrLuallzauon ls noL cloud compuung.
vlrLuallzauon provldes Lhe ablllLy Lo run
muluple, lsolaLed vlrLual machlnes on a slngle
physlcal hosL.
Cloud compuung provldes each auLhenucaLed
user wlLh self-servlce blllable access, CoS, and
prlvaLe secure lnLer-vM neLworks.
vlrLuallzauon vs. Cloud Compuung
6$/-+1*,/7 ?+$'!/7+@/1*, <7*!- <*;)!1,0
?A B*2/1*, on a speclc hosL on any hosL ln an
avallablllLy zone
?A C'*$/0# always perslsLenL ephemeral or perslsLenL
D#)7*9+,0 ?A E#"*!$2#" cusLomlzable across a
broad range of slzes
sLandard slzes (small,
medlum, large, eLc.)
<:/,0+,0 F))7+2/1*,
E#"*!$2#"
reslze exlsung vM launch new lnsLance
6+;# '* =$*(+"+*, mlnuLes Lo hours+ mlnuLes
?A G/+7!$# E#2*(#$9
A#':*-
auempL Lo recover falled
vM
dlscard and sLarL new
lnsLance
Summary
llve essenual characLerlsucs of cloud compuung
Cn-demand self-servlce
8road neLwork access
8esource poollng
8apld elasuclLy
Measured servlce
Cloud compuung models
Servlce models: SaaS, aaS, and laaS
ueploymenL models: publlc, prlvaLe, and hybrld
Cloud compuung vs. vlrLuallzauon

You might also like