FINAL DEGREE PROJECT
ECONOMIC
MODEL AND
ANALYSIS OF
THE CLOUD
BUSINESS MANAGEMENT DEGREE
UNIVERSITAT POLITÈCNICA DE VALÈNCIA
FACULTAD DE ADE
2015/2016
AUTHOR: PABLO PINÓS ORDIÑANA
ADVISOR: JOSEP DOMÈNECH DE SORIA
SPANISH TITLE:
ANÁLISIS Y MODELADO
ECONÓMICO DE LA NUBE
Economic model and analysis o he cloud
BBA Uni e si a Poli ècnica de València
1
INDEX
Figu es Index 3
Tables Index 4
Acknowledgemen s 5
1. In oduc ion 6
1.1 Abs ac 6
1.2 In oduc ion 7
1.3 Objec i es 9
2. Cloud Compu ing 10
2.1 De ini ion 10
2.2 Elemen s 11
2.3 Se ice Models 13
2.3.1 SaaS 13
2.3.2 PaaS 14
2.3.3 IaaS 14
2.3.4 XaaS 15
2.4 Deploymen models 16
2.4.1 P i a e Cloud 16
2.4.2 Communi y Cloud 17
2.4.3 Public Cloud 17
2.4.4 Hyb id Cloud 17
2.4.5 Cloud Fede a ion 18
3. Supply 19
3.1 Amazon Web Se ices (AWS) 21
3.1.1 His o y 21
3.1.2 P icing 22
3.2 Google Cloud Pla o m 25
3.2.1 His o y 25
3.2.2 P icing 25
3.2.3 Cloud Compu ing Cos s: Moo e’s Law 27
Economic model and analysis o he cloud
BBA Uni e si a Poli ècnica de València
2
3.3 Mic oso Azu e 29
3.3.1 His o y 29
3.3.2 P icing 29
3.4 Summa y 30
4. Demand 32
4.1 Global aspec s o Cloud Demand 32
4.2 Posi i e Fac o s 32
4.3 Nega i e Fac o s 33
4.4 Public s P i a e Cloud 35
4.5 Cloud si ua ion in Spain 37
4.6 Cloud Compu ing si ua ion in he wo ld 41
5. Scena ios 49
5.1 Scena io 1 50
5.2 Scena io 2 56
6. Conclusions 60
6.1 Conclusions 60
6.2 The u u e o Cloud Compu ing 63
A achmen s 67
Glossa y 69
Re e ences 71
Economic model and analysis o he cloud
BBA Uni e si a Poli ècnica de València
3
Figu es Index
Figu e 1 Cloud Compu ing Models 16
Figu e 2 AWS S3 s o age se ice p ice o I eland 23
Figu e 3 AWS EC2 compu ing se ice paymen op ions 24
Figu e 4 G aphic Cloud p ices s. Ha dwa e Cos e olu ion 28
Figu e 5 2015 Use s o Cloud Compu ing Wo ldwide 41
Figu e 6 2016 Use s o Cloud Compu ing Wo ldwide 41
Figu e 7 2015 Responden s by Region 42
Figu e 8 2016 Responden s by Region 43
Figu e 9 2015 Public Cloud Usage 43
Figu e 10 2016 Public Cloud Usage 44
Figu e 11 Public-P i a e Cloud Wo kload (En e p ise s SMB) 45
Figu e 12 Cloud Bene i s (2015 s. 2014) 45
Figu e 13 Cloud Bene i s (2016 s. 2015) 46
Figu e 14 Cloud Challenges (2015 s. 2014) 46
Figu e 15 Cloud Challenges (2016 s. 2015) 47
Figu e 16 G aphic (S o age Scena io) 52
Figu e 17 Azu e S a is ics, Oc obe 2014 60
Figu e 18 Cisco In e cloud 64
Figu e 19 Amazon Web Se ices (HDD SAN+NAS) 67
Figu e 20 Mic oso Azu e (HDD SAN+NAS) 68
Figu e 21 Google Cloud Pla o m (HDD SAN+NAS) 68
Economic model and analysis o he cloud
BBA Uni e si a Poli ècnica de València
4
Tables Index
Table 1 Amazon Volume discoun o Rese ed Ins ances 24
Table 2 GCP ins ances % o cha ge pe mon hly usage le el 26
Table 3 GCP P e-emp ible Ins ances p ices pe hou 27
Table 4 Th ee majo Cloud p o ide s’ se ices/p icing policies 30
Table 5 AWS, GCP and Azu e se ice compa ison 31
Table 6 Spanish Cloud ma ke si ua ion (INE da a) 38
Table 7 To al cos s o 36 mon hs o con inuous se ice 51
Table 8 Scena io 1, ini ial expendi u es compa ison cos 52
Table 9 Scena io 1, 3-yea compa ison able 53
Table 10 Scena io 1, ini ial in es men compa ison 53
Table 11 Scena io 1, 3 yea s cos wi h 80% o use 54
Table 12 Scena io 1, sa ings om 20% use educ ion 55
Table 13 Decision-making able 62

Economic model and analysis o he cloud
BBA Uni e si a Poli ècnica de València
5
Acknowledgemen s
To my pa en s and my amily, because hey always guided and
suppo ed me in my decision e en when hey disag eed.
To all he p o esso s I me du ing my uni e si y pe iod, and
especially o my di ec o P o . Josep Domènech, because hey
made me app ehend wi h hei bes in en ions and helped me
g ow as a s uden and pe son.
To Uni e si a Poli ècnica de València (UPV), Singapo e
Managemen Uni e si y (SMU) and o all hei pe sonnel, s uden s
and o he people I ound on my way h ough hem, o
con ibu ing o make mysel he bes way possible.
To all my dea iends, because hey a e pa o my li e and make
i be e .
Economic model and analysis o he cloud
BBA Uni e si a Poli ècnica de València
6
1. In oduc ion
1.1 Abs ac
This inal deg ee p ojec analyzes he Cloud Compu ing
si ua ion in he wo ld wi h a ocus on he Spanish case, and models
he Cloud Compu ing supply and demand. Di e en analysis
abou he Cloud Compu ing economic model ha e been ca ied
be o e bu ei he hey ocus on he gene al e ms o Cloud, o hey
ocus on speci ic se ice models. We ied o gi e a global
pe spec i e including all h ee main se ices.
In o de o do ha , i ocuses on he de ini ion o he Cloud
Compu ing and i s elemen s, and he ac o s ha may in luence
he demand and supply. Then, some heo ies a e conside ed, and
a scena io-case has been calcula ed o help unde s and how he
supply mee s he demand and i s applica ions.
A he inal sec ion, he conclusions abou he di e en Cloud
Compu ing elemen s, and ac o s ha a ec i s adop ion ha e
been summa ized and a o ecas abou u u e de elopmen s in he
indus y closes he pape .
Keywo ds: Cloud Compu ing, Ne wo k, SaaS, XaaS, Iaas,
PaaS, In o ma ion Sys ems, In o ma ion Technologies, economic
analysis
Palab as cla e: SaaS, XaaS, Iaas, PaaS, Red, Tecnología en la Nube, sis emas de
in o mación, análisis económico
Economic model and analysis o he cloud
BBA Uni e si a Poli ècnica de València
7
1.2 In oduc ion
Cloud Compu ing is a name gi en o he ne wo ked echnology
ha , sha ing esou ces in o de o imp o e economies o scale
wi h con e gen in as uc u es, educes cos and simpli ies and
op imizes managemen . Basically, i is a ansi ion om indi idual
ne wo ks, s o age acili ies, and compu ing acili ies, o he same
se ices accessible om anywhe e, wi h educed cos s and many
mo e oppo uni ies.
Cloud Compu ing a e he Se ices and Solu ions deli e ed and
consumed by he cus ome s o e he in e ne in eal ime. I is a
deli e y model o compu ing se ices (unde s anding as
“compu ing” e e y possible se ice ela ed wi h compu ing and
so wa e enginee ing, in o ma ion sys ems and echnologies, and
compu e science).
Using an e-mail se ice, s o ing iles in online sha ed olde s,
unning web-based applica ions o hos ing hem, e c. all a e Cloud
Compu ing solu ions o se ices ha we use in daily basis wi hou
ealizing i . Cloud Compu ing in ol es many concep s on i s
de ini ion, so some s a i ica ion is needed in o de o analyze he
main cha ac e is ics o each pa .
Du ing he de elopmen o his p ojec (June 2014- July 2016),
di e en announcemen s ha e been made by some en e p ises
ega ding u u e Cloud Compu ing p ojec s. New imp o emen s
o he ac ual models appea ed, and new se ices oo. Acco ding
o Fo bes, he e a e 18.239.258 jobs ela ed wi h Cloud
Compu ing wo ldwide; in he US, he e a e 3.9 million, and om
hose, 384.478 a e IT jobs ( he o he s a e ela ed wi h Cloud
Compu ing bu no wi h IT). Cloud Compu ing is a de eloping
indus y wi h a g owing ma ke and wi hou an ac ual limi on he
scope o g ow h.
As new se ices and applica ions a e launched, new
p o essionals a e needed. In 2012 he employmen o Cloud
Compu ing p o essionals in he US g ew an 80%. This indus y is
a end now, a good in es men oppo uni y and because o ha ,
some hing ha many people alk abou . Wi h his scena io is no
easy o ocus in o de o come up wi h a clea analysis o he
indus y wi h app op ia e de ails and ha is why his p ojec ’s
analysis ocuses on he Public Cloud.
Economic model and analysis o he cloud
BBA Uni e si a Poli ècnica de València
8
This Final Deg ee P ojec p e ends o analyze he Cloud
Compu ing ma ke ’s si ua ion and model i . The e a e di e en
pa adigms in he Cloud Compu ing: Public, P i a e o o he kind
o Cloud Compu ing models ha e di e en cha ac e is ics. Each
pa adigm has di e en cus ome s, ad an ages and disad an ages,
possibili ies, e c.
We ocused on he Public Cloud model, because i is he one
ha has mo e in o ma ion a ailable and allows compa ison, he e
is a public ma ke place om whe e da a is a ailable and he e a e
many compe i o s and cus ome s o ob ain da a om.
Fi s , a li e a u e e iew has been done o analyzing he
si ua ion o he Cloud Compu ing echnology and ma ke . A he
in o ma ion sys ems and compu a ion ield, new echnologies and
applica ions a e c ea ed e e y day, so i is impo an o cla i y wha
we a e ocusing on, in o de o ge a ixed image o a as changing
indus y.
One o he poin s o his p ojec is o analyze he ma ke o
Public Cloud model. To do so, bo h, supply and demand, ha e
been conside ed, se ice p o ide s ha e been analyzed, and hei
p oduc s and p ices compa ed. Only h ee ha e been conside ed
because e en Cloud Compu ing is a as g owing ma ke , many
Public Cloud p o ide s a e s ill small, and ew big co po a ions
ha e la ge ope a ions. Wi h ha said, many IT- ela ed big
companies a e de eloping new p oduc s and se ices o en e his
ma ke in he ollowing yea s.
The ac o s ha in luence Cloud adop ion ha e been
cha ac e ized and de eloped in he demand analysis, Small and
Medium En e p ises (SMEs) do no ollow he same c i e ia han
la ge co po a ions, and new social media s a ups ha e no he
same lexibili y and needs han indus ial ac o ies.
And inally, wo scena ios de eloped o compa ing di e en
cloud adop ions and he p ices o he main cloud p o ide s
depending on he se ice model, ying o de e mine wha is bes
o a company ega ding i s needs and si ua ion.
The conclusion alks abou he si ua ion o he Cloud Ma ke ,
and also in oduces a o ecas looking in o he u u e o new
echnological de elopmen s ha will b ing new models and
sys ems in he indus y, eplacing and changing he ac ual ones.
Economic model and analysis o he cloud
BBA Uni e si a Poli ècnica de València
15
a any ime o he business, is he bes op ion when high
scalabili y is needed, he company does no ha e o wan o spend
money on ac i e asse s and he e a e empo al businesses o
p ojec s. F om small s a -ups o la ge co po a ions, hey all
bene i om he economies o scale o his se ice model.
An example could be he in as uc u e needed by a company
o hos i s websi e, i would need some i ual machines, CPUs,
s o age memo y and disk memo y (RAM) o hos he websi e and
un i s applica ions; he company could change one o some o he
in as uc u e cha ac e is ics i needed a any ime ins an ly
inc easing o educing he bill and he se ice ob ained. Some o
he main IaaS p o ide s a e Mic oso Azu e, Google Compu ing
Engine, Amazon Web Se ices, Joyen , e c.
2.3.4 XaaS
The e is a ou h se ice model ha NIST de ini ion does no
include ye bu ha has been in oduced ecen ly and become
popula a he li e a u e.
The E e y hing as a Se ice (XaaS), is he agg ega ion o all
se ice models, ha appea s as a ma ke need in o de o exp ess
he se ice model ha combines he h ee p e ious se ice
models. I is a se ice ha p o ides in as uc u e, so wa e and
pla o m a he same ime. The end consume only pays one bill
e en each sub-se ice may be p o ided by di e en companies
specialized in di e en se ices.
An example could be a Desk op as a Se ice (DaaS), a i ual
desk op unning on a i ual en i onmen wi h he OS,
applica ions, e c. ha he cus ome needs wi hou he need o
ha ing he physical ha dwa e in he same place (some ha dwa e
and so wa e would be needed in o de o in e ac wi h he i ual
desk op, bu hey could be only minimal). So he cus ome may
ha e a i ual compu e on he Cloud, accessible om anywhe e
om a minimal in e ace, bu accessing esou ces ha will no be
possible o ha e o he wise. Ano he one could be a Disas e
Reco e y as a Se ice (DRaaS), a se ice o eco e ing all
in o ma ion and in as uc u e hos ed in a cloud en i onmen
a e a na u al o human-made disas e happens (powe sho age,
sabo age, hacking…), in o de o ensu e business con inui y. The
se ice will p o ide he necessa y esou ces and backup o a
business o con inue wi h egula o educed se ice o i s
cus ome s, ha ing all i s da a secu ed and p o iding ins an ly he
esou ces ha we e los .

Economic model and analysis o he cloud
BBA Uni e si a Poli ècnica de València
16
Figu e 1 Cloud Compu ing Models
Sou ce: Sul an, 2010, A simple ep esen a ion o in o ma ion communica ion ia cloud compu ing.
2.4 Deploymen models
Acco ding o he NIST, he e a e ou di e en Deploymen
models ega ding he kind o p o ide o he cloud se ice, as in
he se ice models, we ha e in oduced one mo e by ou own.
A deploymen model de ines he way he se ices a e owned
and used. He e he e a e he main Cloud Compu ing deploymen
models:
2.4.1 P i a e Cloud
A cloud is said o be P i a e i one single o ganiza ion ( ha
inco po a es mul iple consume s) uses an in as uc u e supplied
o i s exclusi e use. I may be managed, owned and ope a ed by
a hi d pa y, he o ganiza ion i sel o a combina ion o bo h, and
may exis on o o he p emises.
Mos o he en e p ises use his model, because i is he i s
ha appea ed, i does no ha e dependence om o he s, and also
because o he secu i y and a ailabili y om he esou ces poin o
iew, i p o ides be e quali y i hey ha e he p ope esou ces.
Nowadays, e en i is no he bes op ion om he economic poin
Economic model and analysis o he cloud
BBA Uni e si a Poli ècnica de València
17
o iew, i con inues being chosen because he companies al eady
ha e hei sys ems buil a hei p emises and changing he model
would ca y some cos s and unce ain y.
P i a e Cloud is used widely bu i is no compa ible wi h all he
Cloud se ices and applica ions de eloped. The e a e companies
o e ing IaaS on he cus ome p emises, managed by he p o ide ,
o PaaS and SaaS he e a e many op ions bu he main p oblem
usually is in eg a ion be ween he p o ide and he consume
sys ems.
2.4.2 Communi y Cloud
I is a Communi y Cloud when i is c ea ed o he exclusi e use
o consume s om di e en o ganiza ions ha sha e conce ns
(policy, secu i y equi emen s, mission, e c.). I may be managed,
owned and ope a ed by a hi d pa y, one o mo e o he
o ganiza ions o he communi y o any combina ion o bo h, and
may exis on o o he communi y p emises.
This is no one o he mos common models, bu i has some
ad an ages as he p i a e cloud in e ms o secu i y, compliance,
e c. and a he same ime, sha ing he esou ces be ween di e en
o ganiza ions may sa e cos s and gi e mo e lexibili y and
capabili ies. I is a good op ion i di e en o ganiza ions ha e he
same goals and a e willing o sha e esou ces in o de o ob ain
highe bene i s.
2.4.3 Public Cloud
A Public Cloud is c ea ed o open use by he public. I may be
managed, owned and ope a ed by business, academic, o
go e nmen o ganiza ions, o any combina ion o hem, and exis s
on he p emises o he cloud p o ide .
This is he model used by mos cus ome s, i has some
p oblems wi h secu i y and law compliance e ms, bu i is he one
ha p o ides highe scalabili y, lexibili y and capabili ies. I allows
co po a ions wi h peak needs o sa e lo s o money on
in es men s and a he same ime, educes he ba ie s o use
Cloud Compu ing.
2.4.4 Hyb id Cloud
A cloud is said o be Hyb id when i is c ea ed by any
combina ion o wo o mo e o he h ee p e ious models (p i a e,
communi y and public), ha emain unique en i ies, bu a e
Economic model and analysis o he cloud
BBA Uni e si a Poli ècnica de València
18
bounded by p op ie a y o s anda dized echnology ha allows
applica ion and da a po abili y.
I is an op ion used by many co po a ions ha need o ha e
highe secu i y s anda ds o o comply wi h local laws and ha e
hei own p i a e cloud on hei p emises, bu a he same ime,
hey wan o bene i om he la ge scale o Public Cloud se ices.
Me ging bo h models is no easy bu some imes he se ices can
be di ided. I is he mos used o he deploymen models, ha is
because i combines he bene i s o he wo i s models and
p o ides he solu ions ha companies need igh now.
2.4.5 Cloud Fede a ion
Due o cons an changes and inno a ions in he Cloud ield,
he e is a i h model, a new p oposal o model combina ion, called
Cloud Fede a ion.
A Cloud Fede a ion is he unioniza ion o so wa e,
in as uc u e and pla o m se ices om dispa a e ne wo ks ha
can be accessed by a clien ia he in e ne . The ede a ion o cloud
esou ces is acili a ed h ough ne wo k ga eways ha connec
public o ex e nal clouds, p i a e o in e nal clouds (owned by a
single en i y) and/o communi y clouds (owned by se e al
coope a ing en i ies); c ea ing a hyb id cloud compu ing
en i onmen .
Economic model and analysis o he cloud
BBA Uni e si a Poli ècnica de València
19
3. Supply
The supply o Cloud Compu ing se ices and solu ions
wo ldwide is o med by housands o p o ide s o di e en sizes.
As he egula ions and accessibili y may a ec he decision o a
cus ome o con ac wi h one p o ide o ano he , and also o he
p o ide s o g ow in di e en loca ions o s ay whe e hey a e,
mos o he p o ide s a e loca ed in ce ain a eas, p o iding
se ices o limi ed loca ions.
The Cloud P o ide s could be sepa a ed in di e en g oups, o
public cloud, la ge in e ne companies, elecoms, and
go e nmen al o ganiza ions.
La ge in e ne companies ha e many la ge da a cen e s a ound
he wo ld, and can p o ide any amoun o esou ces o any
cus ome wi h p ices ha small companies canno compe e wi h.
Some o hem a e Amazon, Sales o ce, Google, Mic oso , e c.
They p o ide all kind o cloud ela ed se ices, mos ly he
con en ional IaaS, PaaS and SaaS, bu i is in hei pla o ms whe e
o he companies de elop new se ices and applica ions, and some
ge in eg a ed in he p o ide ’s po olio.
Small and Medium En e p ises ha e no size o p o ide IaaS o
PaaS se ice o mos o he cus ome s, hey may ope a e wi h
public cloud om la ge in e ne companies, o ha e hei own da a
cen e s (o a mix u e o bo h). They can o e hei se ices
wo ldwide bu hei p ices canno be as compe i i e as wi h la ge
co po a ions.
Public sec o agen s can be uni e si ies, go e nmen en i ies,
e c. They need o ha e hei own cloud o hei ope a ions and
use s, and some imes hey can o e also hei se ices o hi d
pa ies. Fo example, UPV has i s own da a cen e and has
de eloped i s own pla o ms o HR, accoun ancy, esea ch,
lib a y, e-lea ning, da a eco e y, mul imedia, e c. All i s use s ge
access o an email se ice and an online da a s o age, also hey can
eques access o a websi e c ea ion SaaS ha includes s o age, e c.
o a PaaS and IaaS, whe e hey can de elop hei own p ojec s.
Telecom companies o e also cloud se ices o hei clien s and
o o he en e p ises. Mos o hem o e some email accoun s and
s o age in he cloud o landline cus ome s, and cloud s o age o
mobile ones. As hey al eady ha e la ge da a cen e s in hei
Economic model and analysis o he cloud
BBA Uni e si a Poli ècnica de València
20
p emises, hey also o e cloud se ices o en e p ises o o he
cus ome s. Fo example, Tele ónica o e s SaaS and IaaS, as well
as DaaS and managemen o P i a e and Hyb id Cloud.
The p ices and kind o se ices a y a lo and a e no easy o
compa e and measu e. In his indus y, he quali y o he ha dwa e,
he echnicians o he echnical suppo can a ec he se ice
deli e y so, depending on he needs o he consume , some
p o ide s may no be sui able o hei needs.
So he compa able measu es ha one could conside in his
indus y a e he objec i ied quan i a i e ones, as ne wo k ans e ,
s o age, CPU, e c. cos s o IaaS. In he case o PaaS and SaaS, as
each se ice is unique in i s own way, he e is no possibili y o ha e
a gene al compa ison. In PaaS each p o ide suppo s di e en
p og amming languages, and in SaaS he e a e housands o
di e en applica ions ( he e may be a possibili y o compa e
simila applica ions as D opbox/iCloud/Box o
O ice365/Google D i e, bu ha is ou o he scope o his
p ojec ).
Compa ison be ween p o ide s in di e en loca ions and wi h
di e en sizes will no p oduce esul s ha could be o use o his
p ojec , a selec ion o he main global p o ide s wi h simila
se ices and loca ions has been done.
To illus a e he supply, only he h ee majo p o ide s ha e
been selec ed. They own da a cen e s a ound he globe wi h
capaci y o scale esou ces enough o any o i s clien s in a ce ain
momen o ime because o hei size, and hey ha e clea
in o ma ion abou he kind o se ice, p icing model, and suppo
se ice.
They p o ide he se ices h ough Se ice Le el Ag eemen s
wi h he consume s, and hose ollow di e en models ega ding
he kind o se ice and he p o ide ’s policy. The di e en models
a e:
-Pay pe use: I is he egula and mos common op ion. The
consume pays pe iodically he bill o he se ices ha has used
du ing he pe iod. The ac ions o he se ice can be he same
leng h o he pe iod ( o s o age capaci y, e c.) o may ha e sho e
e ms as minu es o hou s ( o compu ing, i ual memo y, and
so).
-Spo /P e-emp ible ins ances: This sys em allows cus ome s
wi h high needs in ce ain momen s o ime o wi h no need o

Economic model and analysis o he cloud
BBA Uni e si a Poli ècnica de València
21
ha ing a con inuous se ice, o “be ” o he p ice o compu ing
ins ances, and hey ge a cheape p ice han wi h pay pe use. This
sys em is based on he p emise ha he p o ide will always ha e
some ee esou ces, hen, ha amoun o esou ces ( ha a ies
on ime) a e “auc ioned” con inuously, and he cus ome s ha
ha e highe needs a ha ime bu wan o pay a cheape p ice, o
he ones ha do no need o ha e a con inued se ice and wan o
sa e on cos s, can ge cheape esou ces. The p ice changes as he
o e and demand changes, bu his sys em allows a mo e pe ec
ma ke .
-Rese ed: E en one o he main cha ac e is ics o Cloud
Compu ing is scalabili y and on-demand esou ces, some
cus ome s may ind hei needs cons an and may wan o ge long
e m SLA wi h he p o ide s. As ha ing pa o he demand
known in ad ance helps p o ide s o p edic and op imize hei
ope a ions, hey can o e discoun s o he cus ome s ha
commi o a minimum expendi u e o a pe iod o ime, so he
cus ome ensu es he esou ces and ge s a discoun , and he
p o ide ensu es a con inue income.
-F ee allowance: Ha ing some ee se ices, wi hou he ull
capabili ies and wi h a limi ed use, as in he eemium models,
allows companies o a ac new cus ome s, and allows possible
cus ome s o y di e en p o ide s and se ices be o e deciding.
3.1 Amazon Web Se ices (AWS)
3.1.1 His o y
Amazon Web Se ices was c ea ed in 2006, and i has been one
o he ea ly adop e s p o iding Cloud Se ice. Amazon s a ed in
1997 as a digi al books o e, and a e he in e ne c isis, ins ead o
going bank up , i con inued g owing. Nowadays, i is one o he
mos impo an e-comme ce companies in he wo ld and i has
de eloped i s ac i i y ange o new indus ies as small elec onic
goods (eBooks, TV’s, pen d i es, cables, ...) and i is one o he
majo Cloud Compu ing p o ide s.
In 2003, i s a ed o de elop i s Cloud se ice ha launched in
2006. I s a ed because a wo ke ealized ha hey we e expe s
on building P i a e Cloud, so hey could become a public cloud
Economic model and analysis o he cloud
BBA Uni e si a Poli ècnica de València
22
p o ide by o e ing hei expe ise o new companies ha wan ed
o en e he e-comme ce and cloud h ough public cloud.
Nowadays, aside om he ype o se ice, Amazon highligh s
because o hei global p esence and ope a ion capabili ies, ha ing
da a cen e s ha co e all geog aphical egions in he wo ld.
Amazon Web Se ices, conside ed he g ea es cloud p o ide
(www.s g esea ch.com) wi h mo e han 30% o he wo ldwide
ma ke sha e, p o ides IaaS, PaaS and SaaS se ices wi h a wide
and inc easing ange o applica ions and speci ic se ices.
3.1.2 P icing
I s p icing policy includes a 12-mon h ee ie se ice wi h
limi ed esou ces including mos o i s se ices. This allows
po en ial cus ome s o y i s capabili ies and unc ions be o e
mo ing hei business in o i , o o s a small be a e sions o
s a -up and new p ojec s. The limi a ions o each se ice a e
es ablished by Amazon, and can be opped up by he cus ome s i
hey need o inc ease he limi o any se ice.
Amazon o e s a pay-pe -use se ice in all hei p oduc s and
ese ed ins ances in mos o hem, and i has pe -hou billing in
i s compu ing se ices. The Spo Ins ances ma ke o compu ing
se ices is a i s deg ee p ice disc imina ion ma ke place whe e
Amazon Web Se ices o e s compu ing esou ces. The p ice is
de e mined by he consume s’ bids and he amoun o esou ces
Amazon Web Se ices can o e . The consume s se he amoun
o esou ces hey wan and he p ice hey a e willing o pay o
hem, and Amazon Web Se ices alloca e he spa e compu ing
esou ces o he highes bids. I a cus ome success ully bids bu
du ing his use o he esou ces o he cus ome se s a highe bid
and he e a e no mo e a ailable esou ces, he cus ome wi h he
lowe bid ge s his se ice in e up ed (bu does no ge cha ged by
he las pe iod lowe han one hou ). Tha p ice disc imina ion
allows he mo e p ice-elas ic cos ume s o ind a be e p ice o
he esou ces hey need.
Fo some se ices (compu e, s o age and da a ans e ), i has
second deg ee p ice disc imina ion, ha allows consume s wi h
big accoun s, o educe hei uni a y cos o big olumes,
eplica ing he economies o scale ha hey would ha e wi h a
p i a e cloud. A he same ime, his second deg ee p ice
Economic model and analysis o he cloud
BBA Uni e si a Poli ècnica de València
23
disc imina ion, allows Amazon o ob ain highe p o i s and
imp o e he ela ionship wi h i s cus ome s.
Figu e 2 AWS S3 s o age se ice p ice o I eland
Sou ce: aws.amazon.com
Fo compu ing esou ces besides he second deg ee p ice
disc imina ion o on-demand se ice, hey allow p e-pu chase o
compu ing capaci y o pe iods be ween 1 and 3 yea s. Fo he 1-
yea e m, hey allow h ee paymen op ions: no capi al up on ,
pa ial capi al up on and o al capi al up on , wi h di e en
ange o sa ings o each scheme.
As one can obse e in he ollowing able, p ice can a y up o
70% o e on-demand cos .
Economic model and analysis o he cloud
BBA Uni e si a Poli ècnica de València
24
Figu e 3 AWS EC2 compu ing se ice paymen op ions
Sou ce: aws.amazon.com
Also o Rese ed Ins ances, he e is a olume discoun policy:
Table 1 Amazon Volume discoun o Rese ed Ins ances
Sou ce: Sel -made om AWS
Rese ed Ins ances allow much g ea e discoun s han wi h he
o he p o ide s’ p ices. Bu he e is a ac ha should be
conside ed: ese ed ins ances ie he cus ome o ha minimum
expendi u e a he ini ial p ice du ing he pe iod con ac ed, wi h
he possible disad an ages ha will be explained la e .
Amazon Rese ed Ins ances Volume Discoun s
Expendi u e
<500K
<4M
<10M
Discoun
5%
10%
No disclosed
Economic model and analysis o he cloud
BBA Uni e si a Poli ècnica de València
31
Amazon Web Se ices p o ides a ee ie wi h limi a ions,
while Google Cloud Pla o m and Azu e p o ide a cash coupon
ha he cus ome can alloca e by his own c i e ia.
All h ee companies ha e mo e simila i ies han di e ences, bu
Google Cloud Pla o m is clea ly mo e di ec ed owa ds SMEs,
s a -up companies and cus ome s who seek simple decisions o
hei se ices, and Azu e and Amazon Web Se ices di ec hei
se ices mo e o big co po a ions o cus ome s wi h la ge
accoun s ha seek also a wide ange o ex a se ices in he same
con ac .
Table 5 AWS, GCP and Azu e se ice compa ison
P o ide
AWS
GCP
Azu e
Billing ype
Hou s
Minu es
Minu es
Cus ome
Ma ke
All, specialis s
in la ge
co po a ions
SMEs and
S a -up
La ge
co po a ions
P ice
educ ion
based on
Expendi u e
commi men
Inc emen al
use
Expendi u e
commi men
F eemium
Se ice?
Yes
Yes
Yes
Sou ce: Sel -made wi h Cloud p o ide s’ public in o ma ion

Economic model and analysis o he cloud
BBA Uni e si a Poli ècnica de València
32
4. Demand
4.1 Global aspec s o Cloud Demand
The e a e many di e en ac o s companies should conside
be o e o decide abou mo ing o a p i a e, public, o a hyb id
cloud. I is clea ha hose ac o s will a ec di e en ly SMEs and
la ge en e p ises (also hey will no a ec he same way companies
om di e en indus ies, bu his s udy will ocus on he company
size, as mos o he li e a u e pape s ollow ha di e en ia ion).
4.2 Posi i e Fac o s
Some economic bene i s o cloud adop ion s a ed by Talukde
e al. (2010), a e:
-S a egic Flexibili y: SMEs need lexibili y o quickly ge ing
o ma ke and deploy hei p oduc s; la ge en e p ises can bene i
oo, i he e a e no so wa e in eg a ion and da a issues. I allows
any company o keep IT cos s as a iable, allowing escala ion
when needed and d as ically educes he capi al in es men .
-Cos educ ion: o SMEs, pay-as-you-go p icing can be
c i ical i en u e o ope a ing capi al is limi ed, and ha ing cu en
expenses ins ead o long e m dep ecia ion may p o ide ax
ad an ages. Fo la ge en e p ises, low se up cos s allow quick and
inexpensi e explo a ions; non-c i ical applica ions can ha e be e
economic pe o mance in educed cos en i onmen s wi hou
jeopa dizing he ope a ions. Fo bo h, being able o aise o d op
he IT expendi u e a any ime, educes cos s and op imizes he
expendi u e.
-So wa e a ailabili y: SaaS and PaaS p o ide so wa e and
in as uc u e a a low cos , SMEs bene i om d ama ic cos
sa ings. La ge en e p ises usually wi h legacy desk op licenses may
be e icen o mo e o SaaS e sions wi h less unc ionali ies
(despi e he cos di e ence). Th ough cloud solu ions any
so wa e can be a ailable anywhe e, as only an in e ace is needed,
om a able in one side o he wo ld, a supe compu e can be
managed emo ely. Any so wa e ega dless o i s needs, can be
un in a cloud en i onmen and accessed wi h a s anda d de ice.
-Scalabili y: i is one o he bes bene i s om cloud
compu ing; all companies can au o scale hei applica ions and
Economic model and analysis o he cloud
BBA Uni e si a Poli ècnica de València
33
ollow demand. La ge en e p ises may need o modi y hei
so wa e o adap i o he cloud, bu he scalabili y educes he
isk exposu e o any business and may inc ease he p o i abili y.
-Skills and s a ing: bo h la ge en e p ises and SMEs bene i
om educ ion o s a cos s on any ask o IT main enance and
suppo . La ge en e p ises will need o analyze hei needs and
adap hei s uc u e o new models, so hey may ake longe o
implemen changes. Cloud solu ions a e eplacing many epe i i e
asks ha a e done by wo ke s; also he ela ed cos s o owning IT
in as uc u e a e elimina ed and included on he cos o he
se ice.
-Ene gy e iciency: cloud compu ing gene a es cos sa ings
and en i onmen al bene i s o SMEs, and o la ge en e p ises,
bu in he case hey ha e p i a e cloud hey s ill ha e hei own
egula cos s. Cloud p o ide s loca e hei acili ies in cos -
e icien loca ions in e ms o ene gy, ans e and land p ice.
-Sys em Redundancy and da a backup: ha dwa e ailu e and
disas e eco e y a e expensi e o SMEs and la ge en e p ises;
bo h can bene i om cloud s o age solu ions ha p o ide sa e
sys ems a lowe cos s, allowing hem access o much mo e
edundan solu ions and indes uc ible backups.
4.3 Nega i e Fac o s
The main cos s o adop ing cloud a e:
-Da a secu i y: La ge en e p ises conside da a hei mos
impo an IT asse , and unce ain y abou he secu i y ha public
cloud p o ides is one o he main ba ie s o public cloud
adop ion. Fo SMEs i is easie o use hi d-pa y secu i y se ices,
bu e en cloud solu ions can p o ide be e secu i y s anda ds
han SMEs could a o d on hei p emises. The ex e naliza ion o
he secu i y o some da a is no accep able always. La ge
co po a ions ace he same p oblems and he e o e hey will no
adop public cloud solu ions o c i ical da a.
-Da a con iden iali y: SMEs and la ge en e p ises ace he
same issues wi h his poin , bu cloud compu ing allows
enc yp ion ha adds sa e y o his aspec . As wi h he secu i y, any
sys em ha is mo e accessible, can ha e mo e chances o being
ulne able.
Economic model and analysis o he cloud
BBA Uni e si a Poli ècnica de València
34
-Da a egula ions: Depending on he indus y and loca ion o
he company he e may be egula o y issues ha may p e en i o
mo e pa o i s applica ions o he cloud (e.g. EU has es ic ed
laws o pe sonal da a p o ec ion). Cloud p o ide s comply wi h
he legisla ion o he coun ies whe e hey ope a e, bu da a
p o ec ion laws may equi e o he owne s o he da a o hold hei
cus ome s’ da a on hei p emises.
-Da a in eg i y: Da a co up ion is likely o happen while
mo ing da a i p o ocols and sys ems a e no consolida ed. Cloud
p o ide s ha e edundan in as uc u e in o de o a oid e o s,
bu his could be expensi e o p i a e clouds, and i he p o ocols
ail o da a is missed in a ans e , i epa able e o s may occu .
-Da a ans e cos s: SMEs usually do no need o ans e
la ge amoun s o da a. La ge en e p ises, on he o he hand, could
need o mo e la ge amoun s o da a i mo ing o public cloud wi h
he consequen isk ha his can ca y. All h ee majo cloud
p o ide s cha ge only ou bound da a ans e s, and companies
ha e o conside also hei own ne wo k cos s. I he se ice is
ou side he company p emises, ans e cos s will become mo e
ele an and may be conside ed on he economic impac o
adop ing cloud.
-In eg a ion cos s: SMEs bene i om SaaS and PaaS se ices;
la ge en e p ises need o ind solu ions ha i hei s uc u e wi h
no inc emen al cos o e hei p e ious en e p ise ag eemen s.
Se ice p o ide s la gely ha e adop ed cloud solu ions and
p o ide SaaS al e na i es ha can be in eg a ed wi h p i a e and
public clouds. Sys em in eg a ion can be cos ly and di icul some
imes.
-Cloud a ailabili y: Slow pe o mance due o ne wo k o
se ice p oblems, and se ice una ailabili y a e se ious conce ns
o SMEs and e en mo e o la ge en e p ises. This is because only
big cloud p o ide s can a o d he s anda d demands o la ge
en e p ises ha equi e 24/7 se ice co e ing all he globe and
immedia e echnical se ice. A powe o ne wo k sho age in he
company o he cloud p emises loca ion can a ec he business,
and when mo e pa ies a e in ol ed, isk is inc eased i
edundancies a e no well managed.
Khajeh-Hosseini e Al. (2011) a hei Cloud adop ion oolki
pape , also men ion s akeholde impac as a ac o o conside
due o sociopoli ical bene i s and isks associa ed.
Economic model and analysis o he cloud
BBA Uni e si a Poli ècnica de València
35
4.4 Public s P i a e Cloud
Acco ding o Cloudonomics heo y, Weinman (2008); he e
a e h ee main di e ences be ween public and p i a e cloud
se ices, and hose di e ences a ec he bene i s hey gene a e
wi h hei adop ion.
The main di e ences a e ha public cloud p o ides ue on
demand se ices, ope a es in much la ge scales han he g ea e
p i a e en e p ises, and ha he i s s educe cos s ia dispe sion
and spli , and he seconds, ia concen a ion and consolida ion.
Weinman collec ed hese key compe ences ha di e ence
public and p i a e cloud and he named hem he 10 laws o
Cloudonomics:
1s : U ili y se ices cos less, e en hough hey cos mo e.
Using a public cloud company, a company pays a highe cos pe
uni o ime han i he esou ces a e inanced, leased o owned
by hem h ough a p i a e cloud. Bu a he same ime, hey cos
no hing when hey a e no needed. The highe he gap be ween
he peak and a e age need, he g ea e he sa ing wi h a public
cloud.
2nd: On-demand umps o ecas ing.
The abili y o eac ins an aneously o o ecas ing de ia ions,
ei he inc easing he p o isions o dec easing hem, sa es cos s
and c ea es e iciency.
3 d: The peak o he sum is ne e g ea e han he sum o he peaks.
A single company can ha e eally high peak needs in a ce ain
momen o ime, e en many companies can ha e hei peak
pe iods a he same ime, bu a public cloud p o ide will ne e
need o ha e enough capaci y o p o ide peak se ice o all i s
Economic model and analysis o he cloud
BBA Uni e si a Poli ècnica de València
36
cus ome s a he same ime, so i s in es men can be conside ably
lowe .
4 h: Agg ega e demand is smoo he han indi idual.
By agg ega ing he demand, he coe icien o a ia ion will be
lowe han he coe icien o a ia ion o a single cus ome
demand.
5 h: A e age uni cos s a e educed by dis ibu ing ixed cos s o e
mo e uni s o ou pu .
Economies o scale bene i he p i a e cloud, bu bene i much
mo e public cloud, as hei olume o pu chasing and ope a ions
is much highe .
6 h: Supe io i y in numbe s is he mos impo an ac o in he esul
o a comba (Clausewi z).
In case o being a acked by hacke s o expe iencing sys em
ailu es, an en e p ise solu ion will be bea en much as e han he
same solu ion in a public cloud ( he a ack o ailu e should ake
down a much g ea e sys em).
7 h: Space- ime is a con inuum (Eins ein/Minkowski).
Nowadays, decision-making depends on compu ing (big da a,
Business In elligence, isk analysis, e c.); ha ing la ge esou ces
a ailable allow business o espond o changing condi ions and
oppo uni ies as e han he compe i ion.
8 h: Dispe sion is he in e se squa e o la ency.
La ency is he delay be ween making a eques and ge ing a
esponse, i is essen ial o ha e educed la ency in o de o deli e
as enough se ices. To educe he la ency o hal i s alue, needs
ou imes mo e esou ces. Then, in o de o be e icien and sa e
esou ces, i is be e o hi e ha capaci y only when needed.

Economic model and analysis o he cloud
BBA Uni e si a Poli ècnica de València
37
9 h: Do no pu all you eggs in one baske .
The eliabili y o a sys em wi h n edundan componen s, each wi h eliabili y
, is 1-(1- )n. So i he eliabili y o a single da a cen e is 99%, wo da a
cen e s p o ide ou nines (99.99%) and h ee da a cen e s p o ide six nines
(99.9999%). While no ini e quan i y o da a cen e s will e e
p o ide 100% eliabili y, we can come e y close o an ex emely
high eliabili y a chi ec u e wi h only a ew da a cen e s. I a cloud
p o ide wan s o p o ide high a ailabili y se ices globally o
la ency-sensi i e applica ions, he e mus be a ew da a cen e s in
each egion.
10 h: An objec a es ends o s ay a es (New on).
Da a cen e s consume la ge amoun s o powe and need p ope
cooling sys ems, co e ne wo k connec ion and cheap land. Public
cloud p o ide s loca e hei da a cen e s in places ha ul ill hese
needs. En e p ises, mo eo e , loca e hei s a hei headqua e s
o egion o ices, whe e he e iciency condi ions a e no gene ally
me . Public cloud, hen, would ha e lowe pe o mance cos s and
will be able o p o ide cheape p ices.
F om hose 10 “laws”, we can conclude ha he e a e signi ican
di e ences be ween Public and P i a e Cloud; Public Cloud
p o ides mo e oppo uni ies han P i a e. I has a highe deg ee
o lexibili y ha diminishes he cos s and inc eases he e iciency
and secu i y.
4.5 Cloud si ua ion in Spain
The Spanish Ins i u o Nacional de Es adís ica (INE), s a ed o
include Cloud Compu ing in i s su eys h ee yea s ago. Table 9
collec s he mos ele an da a.
Economic model and analysis o he cloud
BBA Uni e si a Poli ècnica de València
38
Table 6 Spanish Cloud ma ke si ua ion (INE da a)
En e p ises
Households
2013*
2014
2015
2013
2014
2015
In e ne use
98,00%
98,30%
98,40%
69,80%
74,40%
78,70%
Use o cloud
solu ions
19,80%
15,00%
15,40%
N/A
N/A
N/A
S o age
86,90%
69,00%
63,60%
N/A
32,40%
N/A
Da abases
N/A
54,70%
56,50%
N/A
N/A
N/A
eMail
N/A
61,40%
70,60%
N/A
N/A
N/A
Paid use
N/A
N/A*
N/A*
N/A
6,40%
N/A
*Da a om 2014 and 2015 om he en e p ises, e e s o
companies wi h 10 o mo e employees ha pu chased any kind o
solu ion du ing he las h ee mon hs. Da a om 2013 e e s o
companies ha used any ee o paid solu ion.
Sou ce: Sel -made om INE p ess eleases
The da a p esen ed om he Ins i u o Nacional de Es adís ica
(INE) om he Households, shows ha in e ne use has g own
om 70% o 79% om 2013 o 2015. Rega ding he use o
S o age Cloud Compu ing se ices, we only ha e da a om 2014,
when ha module was added o he su ey, bu i was no
con inued in 2015. In 2014 he da a collec ed also included he use
o S o age Cloud Se ices and he eason o use/no o use hem.
Du ing 2014, 32.4% o he in e ne use s used any cloud s o age
se ice du ing he las h ee mon hs, and a 6.4% paid o hem.
Anyway, a 60.5% o he in e ne use s ha did no use any cloud
s o age se ice decla ed no o know i s exis ence, bu a he same
ime some decla ed o use he email as s o age se ice.
The da a p esen ed om he En e p ises side shows ha he
in e ne use has g own om 98% o 98.4% om 2013 o 2015. In
2013, 20% o he en e p ises wi h 10 o mo e employees used
some kind o cloud solu ion, and 87% o hose used s o age
solu ions. In 2014 15% o he companies wi h 10 o mo e
employees used some kind o paid cloud solu ion, and in 2015
hey we e a 15.4%. Ou o hose, 69% used S o age se ices in
2014 bu only 64% did i in 2015. In 2014 also 61% o he
en e p ises ha paid o Cloud se ices did i o email se ices,
and 56% o Da abase se ices; in 2015 hose numbe s g ew o
71% and 57% espec i ely.
Economic model and analysis o he cloud
BBA Uni e si a Poli ècnica de València
39
The use o in e ne is no ully ex ended ye among he
Households bu i is g owing i mly. Among he en e p ises wi h
10 o mo e employees i is almos a 100%, showing ha nowadays
is a mus ha e o any business.
F om he da a, i is clea ha he in o ma ion eleased by he
INE is no s anda dized ye . Rega ding he use o cloud by
en e p ises, da a om 2013 says ha 19.8% o he companies wi h
mo e han 10 employees used Cloud solu ions, and 2014 and 2015
da a e e s o he pe cen age o companies ha pu chased any
cloud solu ion (15%). So on he en e p ises side a e 2013 we a e
ge ing da a on paid cloud se ices only, as we ha e seen ha some
p oduc s a e p esen ed wi h a eemium model, and ha some
ins i u ions p o ide ee cloud se ices, … So he da a om 2013
canno be compa ed wi h he 2014 and 2015 ones.
The use o cloud solu ions is a 32% a he Households, being
he use o s o age a 31% and he only one he e is da a om. We
also ha e ha 6% o he use s paid o he se ice. The igu es a e
lowe han a he en e p ise le el, and ha is unde s andable o
he di e en needs o an indi idual and an en e p ise.
We ha e ha 87% o he en e p ises in 2013 used cloud s o age
se ices; in 2014 and 2015 a 69% and 64% o he use s ha paid
o cloud se ices used s o age se ices, ha is consis en wi h he
models ha Google, Mic oso o D opbox ha e, gi ing a ee
s o age o some GB and hen cha ging o ex a ea u es and mo e
capaci y.
All his da a makes sense conside ing a s udy by he Na ional
Obse a o y o IT and IS (ONTSI) ha ecognized in 2012 ha a
15% o he companies used some kind o Cloud se ice. Also, a
o ecas om IDC se he pe cen age o Spanish companies ha
would use SaaS in 2012 a 18%.
This di e s wi h he INE da a, conside ing also ha he lack o
speci ic knowledge abou Cloud Compu ing may lead su eys o
miss da a om ee SaaS (as Gmail, Ho mail, Kindle, e c.). In
2013, a 42.99% o he Spanish companies wi h mo e han 10
employees mani es ed no o ha e a high knowledge abou Cloud
echnology. In 2014, he pe cen age was 34.87%. Tha numbe could
a ec he esul s o he su ey because o he companies ha do
no ha e a high knowledge abou Cloud echnology, bu ha use
cloud based o di ec ly Cloud echnology wi hou no icing i .
Economic model and analysis o he cloud
BBA Uni e si a Poli ècnica de València
40
Conside ing he use o in e ne and Cloud solu ions by he
households, he numbe s do no ma ch. I 8.5 million people use
some kind o Cloud Compu ing se ice in Spain (32% in 2014),
bu on ha yea 66.3% o he popula ion used mail se ices o
send a ached iles, and in he p e ious yea 79.5% o he
in e iewed did he same, some hing is w ong. Fi s , conside ing
he echnology gap, he su ey may no be accu a e o de e mine
he numbe o eal use s, and second, ha ing hose pe cen ages on
he use o email and on sending iles a ached, p obably hey could
use some Cloud Compu ing based o Cloud Compu ing solu ion
wi hou knowing i .
The changes in he su ey om he INE do no help o
de e mine he eal numbe o Cloud solu ions use s in 2014 and
2015. Conside ing ha he p e ious yea almos a 20% o he
Spanish companies ecognized o use Cloud Compu ing, and
conside ing ha only a 6.3% o he use s o s o age sys ems paid
o he se ices, i is no easy o hink ha a 15% o he Spanish
companies used paid Cloud Compu ing solu ions in 2014 ( ha
would make 2/3 o he Cloud Compu ing solu ions en e p ise
use s in he p e ious yea ).
Ha ing a 15% o Spanish companies paying o he se ice may
be possible, bu he numbe o companies using some kind o
Cloud se ice would be much highe .
The p oblem esides in he unde s anding o Cloud se ice by
he use s, nowadays mos o he Apps ha he use s ha e a hei
phones a e connec ed o he cloud, Mic oso launched i s O ice
365 package based in he cloud ha includes s o age, da a backup,
oIP, e c. All public email use s a e using cloud based se ices,
ha usually include some s o age se ice as well and some
messaging/con e ence se ices as well, bu as hey a e using he
se ice o many yea s on daily basis and o ee, hey do no
ecognize i as a Cloud se ice.
So Spain is a coun y wi h high use o in e ne , and a g owing
numbe o paid cloud se ices, he use o ee se ices is also
ex ended bu i could no be de e mined om he da a collec ed.
Anyway, conside ing social media, email, s o age, pho o sha ing,
e c se ices he numbe o Cloud se ices in Spain is also high and
g owing.
Economic model and analysis o he cloud
BBA Uni e si a Poli ècnica de València
47
Figu e 15 Cloud Challenges (2016 s. 2015)
The challenges o he Cloud use s a e he ma e o discussion
o di e en o ums, and he p o ide s a e ying o ind sui able
solu ions o b eak he ba ie o en y o many new possible
cus ome s.
Secu i y was he main issue ha conce ns he cus ome s in
2015, inc easing om 2014 o 2015; bu in 2016 he lack o
esou ces and expe ise o e ook secu i y as main issue. The e
we e no main issues ha ha e dec eased he conce n o he
cus ome s in 2015, and in 2016 only pe o mance and complexi y
did i . The ones ha inc eased mos we e he lack o
esou ces/expe ise, and managing mul iple cloud se ices. Those
a e ela ed wi h he lack o enough ained p o essionals, due o
he as e olu ion and changes o his echnology. The con inuous
changes and he lack o s anda diza ion make i di icul o he
companies o bene i comple ely om he new op imiza ions and
changes, so o he ollowing yea s, he demand o Could
Compu ing p o essionals will con inue g owing, and he p oblems
de i ed o managing mul iple se ices o di e en ones, will be
sol ed by Cloud Compu ing solu ions ha will make i easie o
he IT echnicians.

Economic model and analysis o he cloud
BBA Uni e si a Poli ècnica de València
48
Economic model and analysis o he cloud
BBA Uni e si a Poli ècnica de València
49
5. Scena ios
Wi h he ollowing scena ios, he au ho p e ends o simula e
di e en clien demand si ua ions in o de o e i y which cloud
p o ide is mo e sui able o each kind o consume and also, i
Public Cloud adop ion is mo e bene icial o he consume .
The me hodology used o he compa ison is cos di e ence,
because he alue gene a ion o he in es men should be he
same in all cases. Each p o ide de ends ha hey p o ide ex a
quali y se ice, and he e a e di e en addi ional se ices included
in he p ice.
The e a e many quali a i e ac o s ha may di e ence one
p o ide om o he , bu in o de o simpli y he model, we ied
o educe hose ac o s by selec ing simila se ices.
Fo he demand pa , he e a e consume s ha no only ake in
accoun he cos ac o when deciding on mo ing hei esou ces
o cloud en i onmen s, o o choose one o o he p o ide . Due
o he complexi y ha in oducing all hose quan i a i e a iables
would a ec he model, he au ho has decided no o ake hem
in he economic decision, bu o conside hem as quali a i e
aspec s o ake in accoun by he consume .
The p ices o he public cloud se ices ha e been ob ained
om he cloud p o ide s si es’, he p ices o he P i a e Cloud
ha e been ob ained using Amazon Web Se ices To al Cos o
Owne ship calcula o , and he calcula o assump ions and
me hodology ha e been e iewed by he au ho in o de o e i y
ha he se ice p o ided was simila .
Two di e en scena ios ha e been simula ed in o de o show
he di e ences be ween di e en kinds o consume s and
se ices.
Economic model and analysis o he cloud
BBA Uni e si a Poli ècnica de València
50
5.1 Scena io 1
Public Sec o Agen , In as uc u e as a Se ice
As a Public Sec o Agen , he au ho has chosen a public
uni e si y, he Uni e si a Poli ècnica de València in his case.
Uni e si a Poli ècnica de València (UPV) is a uni e si y
loca ed in he ci y o Valencia, Spain. I is a Top 10 uni e si y in
he coun y, and he i s echnological uni e si y, wi h mo e han
39000 use s be ween pe sonnel and s uden s, and an annual
budge o mo e han 350M eu os.
Many o i s IT se ices a e al al eady in a P i a e G id Campus
ha p o ides compu ing, email, websi e and s o age acili ies o
all i s s uden s and pe sonnel.
As an In as uc u e as a Se ice, UPV could pa ne wi h any
cloud p o ide and ou sou ce mos o i s g id esou ces mo ing
hem o cloud solu ions.
Conside ing he In as uc u e as a Se ice, a mix u e o
se e s and s o age should be conside ed. As he di e en
se e s’ se ices ha he main cloud p o ide s o e ha e
quali a i e di e ences ha a e di icul o compa e, a s o age cos
compa ison has been done.
The uni e si y owns SAN and NAS Fuji su da a cen e s o
p o iding se ices as online ha d d i e s o age, he online ile
exchange o email.
In o de o compa e he cos o a p i a e owned in as uc u e
and a public cloud one, using he Amazon Web Se ices
calcula o , only s o age cos s we e accoun ed.
The assump ions a e he ollowing:
Fo he p i a e da a cen e , he ollowing ha dwa e needed o
se ing up he da a cen e , has been accoun ed ( om HDD o
cables, wi h a 50% discoun om he egula p ice, as indus y
policy).
Economic model and analysis o he cloud
BBA Uni e si a Poli ècnica de València
51
Powe , Cooling, and da a cen e space cos s ha e also been
accoun ed, he p ice o he kWh used is he cu en comme cial
p ice a US (10c$/kWh). The Spanish indus y, one wi h he
ac ual cu ency exchange a e, would be almos he same
(0.099*1.13=0.11$).
The p i a e cloud s o age needed in o de o p o ide he same
se ice ha Cloud P o ide s p o ide, edundancy is needed, so
a ailable capaci y is g ea e .
Fo example, wi h 100TB o physical memo y, a e he 7% o
memo y ha he Ope a ing Sys em measu es di e en han he
disk manu ac u e , plus he 50% loss ha comes wi h he RAID
10 dis ibu ion, equals 46.5TB.
RAID dis ibu ion duplica es he da a in di e en d i es in
o de o ob ain da a edundancy o pe o mance imp o emen ,
p o ec ing he da a om disk ailu e and un eco e able e o s.
Fo he simula ion he UPV ASIC has been con ac ed
ega ding hei sys ems and echnology, and public p ices ha e
been ob ained om Amazon, Google and Mic oso (see
a achmen s).
Table 7 To al cos s o 36 mon hs o con inuous se ice
3 yea cos
Resou ces
P i a e sys em
AWS
Mic oso Azu e
GCP
10TB SAN +
10TB NAS
100.640,00 €
19.851,00 €
18.046,36 €
13.534,77 €
100TB SAN +
100TB NAS
543.640,00 €
217.367,04 €
197.606,40 €
148.204,80 €
1000B SAN +
1000TB NAS
4.629.800,00 €
1.941.771,65 €
1.765.246,95 €
1.323.935,22 €
F ee esou ces
0
30GB
200$
300$
Sou ce: Au ho 's calcula ions based on da a p o ided by cloud p o ide s.
Fo he P i a e Sys em and Amazon Web Se ices esul s, he
AWS To al Cos o Owne ship Calcula o has been used
(aws cocalcula o .com). I conside s, Se e , S o age, Ne wo k
and IT-labo cos s, he calcula o has been e iewed and ce i ied
by F os & Sulli an.
Economic model and analysis o he cloud
BBA Uni e si a Poli ècnica de València
52
Fo Mic oso Azu e and Google Cloud Pla o m, hei own
se ice calcula o s ha e been used. As he se ice equi ed is
quan i a i e and he quali y di e ence be ween di e en p o ide s
is no being analyzed, i is assumed o be he same le el o all.
Sou ce: Au ho 's calcula ions based on da a p o ided by cloud p o ide s.
Table 8 Scena io 1, ini ial expendi u es compa ison cos
To al cos
Mon hly cos
Resou ces
P i a e sys em
AWS
Mic oso Azu e
GCP
10TB SAN +
10TB NAS
100.640,00 €
551,42 €
501,29 €
375,97 €
100TB SAN +
100TB NAS
543.640,00 €
6.037,97 €
5.489,07 €
4.116,80 €
1000B SAN +
1000TB NAS
4.629.800,00 €
53.938,10 €
49.034,64 €
36.775,98 €
Sou ce: Au ho 's calcula ions based on da a p o ided by cloud p o ide s.
A p i a e sys em will equi e o acqui e all he esou ces needed
since he beginning, so i he asse s a e dep ecia ed 1/3 yea ly, i
will equi e he same capi al o one mon h, one yea o h ee.
In he o he hand, a cloud solu ion will cos acco ding o he
SLA, i will be paid on mon hly basis and will ha e he possibili y
o upg ade o downg ade he esou ces ega ding he needs o he
business.
0 1.000.0002.000.0003.000.0004.000.0005.000.000
10TB SAN+10TB NAS
100TB SAN+100TB NAS
1000B SAN+1000TB NAS
S o age scena io
GCP
Mic oso
Azu e
Figu e 16 G aphic (S o age Scena io)
0 1M 2M 3M 4M 5M

Economic model and analysis o he cloud
BBA Uni e si a Poli ècnica de València
53
Table 9 Scena io 1, 3-yea compa ison able
3-yea cos sa ings o public cloud/p i a e cloud
Resou ces
P i a e sys em
AWS
Mic oso Azu e
GCP
10TB SAN +
10TB NAS
100.640,00 €
80,28%
82,07%
86,55%
100TB SAN +
100TB NAS
543.640,00 €
60,02%
63,65%
72,74%
1000B SAN +
1000TB NAS
4.629.800,00 €
58,06%
61,87%
71,40%
Sou ce: Au ho 's calcula ions based on da a p o ided by cloud p o ide s.
Acco ding o he di e en p ices ha he cloud p o ide s o e ,
and he cos o ha ing a p i a e cloud, he bes economic op ion
o he consume should be con ac ing he se ice wi h a cloud
p o ide . In his case, Google Cloud Pla o m is leading he p ices
o he ma ke .
Fo his model, we had di e en assump ions: he i s one,
opposed o he 10 laws o Cloudonomics, is ha s o age needs a e
no being cons an .
Fo example, a UPV, use s s a wi h 1GB o da a s o age in a
i ual disk. I hey each he limi o i eques ed, hey can inc ease
he amoun o space un il 5GB. Also he cou se pla o m,
Poli o maT, has s o age peaks a he end o each academic yea ,
and his is p edic able.
Table 10 Scena io 1, ini ial in es men compa ison
To al cos
Mon hly cos
Resou ces
P i a e sys em
AWS
Mic oso Azu e
GCP
10TB SAN +
10TB NAS
100.640,00 €
0 €
0 €
0 €
100TB SAN +
100TB NAS
543.640,00 €
0 €
0 €
0 €
1000B SAN +
1000TB NAS
4.629.800,00 €
0 €
0 €
0 €
Sou ce: Au ho 's calcula ions based on da a p o ided by cloud p o ide s.
The p i a e sys em will need he capi al ou lay a he beginning
o he h ee-yea pe iod, unless ag eemen s wi h he ha dwa e
Economic model and analysis o he cloud
BBA Uni e si a Poli ècnica de València
54
endo s a e done (Dell p o ides a 3 yea s inancing scheme wi h
mon hly paymen s), bu his will inc ease he cos wi h he
paymen ’s in e es s.
The public cloud will allow a a iable capaci y ha can be
changed e e y mon h, in o de o sa is y he needs o he cus ome
and would no equi e a capi al in es men on he ha dwa e.
Conside ing ha , UPV would need o ha e enough s o age
capaci y o p o ide use s who need an inc ease (unp edic able
s o age inc eases), and ha he e a e pe iods o he yea wi h
highe needs han o he s, using a public cloud solu ion will allow
hem o sa e cos s when less esou ces a e used.
I i was clea ha he e a e cons an esou ce needs, ha ing a
public cloud se ice will sa e cos s, adding ha he e would be
peak pe iods (so he e would be unde -used capaci y du ing he
non-peak pe iods), and ha a la ge amoun o se ice is needed
o unp edic able inc eases, a public cloud will be much mo e
e icien and cos e ec i e.
Table 11 Scena io 1, 3 yea s cos wi h 80% o use
To al cos
3 yea s cos conside ing 80% o use
Resou ces
P i a e sys em
AWS
Mic oso Azu e
GCP
10TB SAN
+ 10TB
NAS
100.640 €
15.880,80 €
14.437,09 €
10.827,8 €
100TB SAN
+ 100TB
NAS
543.640 €
173.893,63 €
158.085,12 €
118.563,8 €
1000B SAN
+ 1000TB
NAS
4.629.800 €
1.553.417,3 €
1.412.197,56 €
1.059.148,2 €
Sou ce: Au ho 's calcula ions based on da a p o ided by cloud p o ide s.
As men ioned be o e, Public Cloud allows business o a o d
peak needs wi hou ha ing unused esou ces du ing non-peak
pe iods, i UPV uses only an 80% o i s o al capaci y because i s
peak pe iods and con ingency ma gins o possible s o age
inc ease eques s om he use s, he cos o public cloud will d op,
bu he p i a e one will emain cons an .
Economic model and analysis o he cloud
BBA Uni e si a Poli ècnica de València
55
Table 12 Scena io 1, sa ings om 20% use educ ion
Sa ings om 20% use educ ion
Resou ces
P i a e sys em
AWS
Mic oso Azu e
GCP
10TB SAN +
10TB NAS
0,00 €
3.970,20 €
3.609,27 €
2.706,95 €
100TB SAN +
100TB NAS
0,00 €
43.473,41 €
39.521,28 €
29.640,96 €
1000B SAN +
1000TB NAS
0,00 €
388.354,33 €
353.049,39 €
264.787,04 €
Sou ce: Au ho 's calcula ions based on da a p o ided by cloud p o ide s.
The sa ings ob ained om using a public cloud and escala ing
he esou ces i needed, conside ing 3 yea s o use wi h a 20% o
educ ion om he i s o ecas , will p oduce a 20% o cos
sa ings.
S o age cos s ha e no discoun s o con inued use, p e-
paymen , e c. bu all h ee majo p o ide s ha e en e p ise
discoun s ha may a ec he p ice. As hose discoun s a e no
speci ied, we canno conclude i his will change signi ican ly he
p ice di e ence be ween p o ide s, so Google Cloud Pla o m
would be he cheapes op ion.
Using Public Cloud se ices is cheape han building P i a e
Cloud ones, bu his eason is no enough o make UPV mo e i s
s o age and compu e acili ies o a public cloud se ice p o ided
in I eland o o he coun y by a majo cloud p o ide .
A he uni e si y, hey alue o he bene i s om ha ing hei
own g id and compu ing acili y a hei p emises (e en i i is
economically ine icien ). Secu i y is he main conce n, hey
alleged ha e ms and condi ions a e no clea and ha non-
disclosu e ag eemen s a e no clea ei he because, e en he majo
cloud p o ide s ha e da a cen e s in EU coun ies, hei
headqua e s may be in o he non-EU coun ies whe e o he laws
apply.
Rega ding o his, all h ee majo Public Cloud p o ide s ha e
hei headqua e s in US, and acco ding o US law, he US
go e nmen can equi e any in o ma ion o hem, and hey should
comply.
Economic model and analysis o he cloud
BBA Uni e si a Poli ècnica de València
56
5.2 Scena io 2
In oduc ion o SaaS o educe human capi al expendi u e
Replacemen o human capi al by So wa e as a Se ice is
happening oday, and i manages epe i i e, low quali ica ion and
huge ime consuming ac i i ies wi h Cloud se ice solu ions ha
eplace wo ke s. Cloud compu ing is changing he way o wo king
o many co po a ions, om he in oicing sys em, pay oll,
accoun ancy, ma ke analysis, incidence managemen , CRM o a
call cen e ; all can be managed wi h a SaaS and less headcoun i
he business has enough scale.
Some cha ac e is ics and needs o eplace people by cloud
based sys ems a e:
In o de o be easible o be eplaced by a cloud solu ion (by a
So wa e as a Se ice), he ask has o bene i om one o mo e
o he main cha ac e is ics o Cloud Compu ing; b oad access, on-
demand, esou ce pooling, apid elas ici y and measu ed se ice.
Fo example, Mic oso la ely is p omo ing i s new Mic oso
O ice 365 p oduc o e i s adi ional Mic oso O ice
(Home/P o essional/…) e sions. Wi h Mic oso 365, he use
can access his documen s and a cloud e sion o he p og am om
any de ice and loca ion. Also i includes email and
ideocon e encing se ices and 1TB o cloud s o age, inc easing
he ange o cloud se ices ha come wi h he license.
O he majo Cloud p o ide , Sales o ce, is a company s a ed
as a SaaS se ices company ha now also o e s PaaS. I s main
solu ion, Sales o ce.com, is a CRM (Cus ome Rela ionship
Manage ), a ool ha allows co po a ions o manage housands o
cus ome s wi h hei con ac s, sales, e c om a sys em ha
o ganizes and i ualizes mos o he in e ac ions. So his ool,
used by housands o business wo ldwide, allows o ha e
housands o con ac s om housands o p o ide s in a da abase
in eg a ed wi h he sales da a om he clien s and hei espec i e
con ac s, educing he employee in e ac ion o minimal
supe ision and in e p e a ion o he analy ics p o ided by he
sys em.
A company implemen ing Sales o ce would equi e less
adminis a i e wo kload, allowing he employees o ocus on mo e
alue added asks. A he same ime, i will equi e mo e IT s a
and ained employees on handling he so wa e and analyze da a
and make app op ia e business decisions.
Economic model and analysis o he cloud
BBA Uni e si a Poli ècnica de València
63
6.2 The u u e o Cloud Compu ing
The e a e di e en heo ies abou he u u e o Cloud
Compu ing, ollowing he de elopmen o he indus y, and
compa ing i o o he echnologies’ pa h, he au ho hinks ha
u u e o cloud is ede a ed.
Cloud ede a ion goes one s ep u he , sugges ing he
uni ica ion o public, p i a e, hyb id and communi y clouds in a
cloud o clouds, whe e esou ces can be connec ed and sha ed
(wi h he p ope cha ge).
In he IEEE 3 d In e na ional Con e ence on Cloud Compu ing
(2010), some p o esso s om Uni e si a Poli ècnica de Ca alunya
p esen ed a pape abou Cloud Fede a ion (Goi i e al. 2010)
saying: “Cloud ede a ion has been p oposed as a new pa adigm
ha allows p o ide s o a oid he limi a ion o owning only a
es ic ed amoun o esou ces, which o ces hem o ejec new
cus ome s when hey ha e no enough local esou ces o ul il
hei cus ome s’ equi emen s. Fede a ion allows a p o ide o
dynamically ou sou ce esou ces o o he p o ide s in esponse o
demand a ia ions.”
Cloud Fede a ion concep was i s used in 2007 by Ke in Kelly
as a cloud o clouds, called In e cloud, as one machine comp ised
o all se e s and a endan cloudbooks on he plane , and i s
cha ac e is ics a e highe eliabili y, back up, a ailable 24/7/365
om any e minal in he wo ld, possibili y o hold in ini e apps
and s o age, and seamless sha ing and collabo a ion possibili ies.
Un il la e 2014 he e ha e no been eal ede a ion solu ions, bu
la ely, he e ha e been ew new solu ions ying o en e he cloud
ma ke .
The main issues o cloud ede a ion acco ding o Al-Tehmani
(2012) a e he in e ope abili y, secu i y, us , consume s,
moni o ing, legal issues, and quali y o se ice.
T us , secu i y, legal, moni o ing, and quali y o se ice issues
a e also issues o egula cloud adop ion. The in e ope abili y is
unique o cloud ede a ion, and i has o do wi h he p oblem o
connec di e en da a cen e s and make hem wo k as one
oge he .

Economic model and analysis o he cloud
BBA Uni e si a Poli ècnica de València
64
Each public cloud p o ide has his own p o ocols, and each
p i a e cloud is buil in o de o p o ec i sel om ou side a acks.
Cloud Fede a ion has o o e come all hose p oblems and be able
o connec any kind o public, p i a e, hyb id, and communi y
cloud, and make hem wo k as one.
The ollowing Cisco diag am shows how hei p oduc will
wo k, enabling use s o ha e access o di e en public and p i a e
cloud solu ions a he same ime.
Figu e 18 Cisco In e cloud
Sou ce: Cisco.com
In o de o achie e ha , hei sys em should be able o access
all hose di e en pla o ms and manage hei combined esou ces
as a unique cloud. Cisco has no launched ye i s Cloud Fede a ion
solu ion, bu hey a e wo king owa ds i and made i public.
Mic oso and HP also said ha hey in en o build an
“in e connec ed se o in e ope able clouds”, and OnApp, a cloud
managemen so wa e de elope , al eady launched ha allows i s
cus ome s o adop a ede a i e en i onmen in hei public o
p i a e clouds.
Being in e ope abili y he majo ba ie o a eal in e cloud, he
“in e connec ed se o in e ope able clouds” solu ion is a s ep
owa ds cloud ede a ion ha allows di e en clouds o adop
s anda ds and be connec ed.
Cloud Fede a ion claims o become a cloud o clouds, as
in e ne did a e he c ea ion o he Wo ld Wide Web and he
Economic model and analysis o he cloud
BBA Uni e si a Poli ècnica de València
65
TCP/IP p o ocols. Fi s , he uni e si ies, a my and big
co po a ions, c ea ed hei own in e nal ne wo ks, called in ane s.
A e ha , some uni e si ies and esea ch acili ies connec ed
hei egional academic ne wo ks and c ea ed an “in e connec ed
se o in e ope able ne wo ks”, and hen, s anda ds we e c ea ed
in o de o adop a global ne wo k o ne wo ks, known nowadays
as in e ne .
Cloud Fede a ion is expec ed o ollow he same pa h, bu o
s anda ds o be es ablished, all cloud p o ide s and owne s should
ag ee wi h ha , bu a he momen he e is no solu ion accep ed
globally.
Bene i s om Cloud Fede a ion:
Cloud Fede a ion allows a mo e e icien u iliza ion o he
esou ces. Theo e ically, as in any ma ke wi h pe ec
compe i ion, wel a e is a i s maximum poin because he e a e
many buye s and selle s, and none o hem has enough powe o
a ec p ice.
P i a e Cloud owne s, would bene i om ede a ion by en ing
hei unde used esou ces o hi d pa ies, he owne would ha e
he same ixed cos s when buying he in as uc u e bu , i he cos
o main aining he esou ces ope a i e is lowe han he bene i
om en ing hem, i s e enue will imp o e.
On he o he side, a p o ide who has no enough local
esou ces o a end his wo kload, would be able o ou sou ce his
esou ce needs o o he p o ide s, bu his expec ed e enue om
he ou sou ced esou ces should be highe han he ou sou cing
cos .
Cloud Fede a ion way o p o iding esou ces is likely he
elec ici y se ice ma ke , whe e he e a e majo se ice p o ide s
wi h s anda d p ices and le el o se ice, and he consume s can
eely choose he p o ide hey p e e . Also, i allows p i a e
p oduce s o he esou ce o sell hei esou ces o he ne wo k (in
he elec ici y ma ke could be sola panel owne s when hey sell
hei no -used elec ici y o he ne wo k, as in he ede a ed cloud),
Economic model and analysis o he cloud
BBA Uni e si a Poli ècnica de València
66
he p i a e cloud owne sells i s unused compu ing o s o age
capaci y o hi d pa ies du ing a pe iod o ime.
Economic model and analysis o he cloud
BBA Uni e si a Poli ècnica de València
67
A achmen s
P ice samples o Amazon Web Se ices, Google Cloud Pla o m
and Mic oso Azu e simula ions o Scena io 1:
Figu e 19 Amazon Web Se ices (HDD SAN+NAS)
Sou ce: aws.amazon.com
Economic model and analysis o he cloud
BBA Uni e si a Poli ècnica de València
68
Figu e 20 Mic oso Azu e (HDD SAN+NAS)
Sou ce: azu e.mic oso .com
Figu e 21 Google Cloud Pla o m (HDD SAN+NAS)
Sou ce: cloud.google.com

Economic model and analysis o he cloud
BBA Uni e si a Poli ècnica de València
69
Glossa y
API: The Applica ion P og amming In e ace is a se o ou ines, p o ocols
and ools o building so wa e applica ions. I de ines unc ionali ies ha a e
independen o i s implemen a ions, allowing de ini ions and implemen a ions
o a y wi hou comp omising he in e ace. I helps he p og amme s educe
he ime hey need o de ine and build hei p og ams.
CPU: The Cen al P ocessing Uni , is he ha dwa e componen o a compu e
whe e mos o he calcula ions ake place. I is he b ain o he compu e , he
mos impo an elemen o a compu ing sys em, because de e mines i s
compu ing capabili ies.
CRM: Cus ome Rela ionship Managemen , a so wa e ha helps companies
o manage hei ela ionship wi h cus ome s and o manage he da a om he
sales, ma ke , e c allowing he business o au oma e analysis o business
decisions.
Ha dwa e: I is de ined as he physical componen s ha o m a compu e ,
ha de ine he capaci y, speed and capabili ies when combined wi h he
so wa e.
HDD: The Ha d Disk D i e is he mechanism ha con ols he posi ioning,
eading and w i ing o he ha d disk, which u nishes he la ges amoun o
da a s o age o a compu e .
Legacy: An old me hod, echnology, compu e sys em, o
applica ion p og am, "o , ela ing o, o being a p e ious o ou da ed compu e
sys em." O en a pejo a i e e m, e e encing a sys em as "legacy" o en implies
ha he sys em is ou o da e o in need o eplacemen .
NAS: A ne wo k-a ached s o age (NAS) de ice is a se e ha is dedica ed
o no hing mo e han ile sha ing. NAS does no p o ide any o he ac i i ies
ha a se e in a se e -cen ic sys em ypically p o ides, such as email,
au hen ica ion o ile managemen .
NAS allows mo e ha d disk s o age space o be added o a ne wo k
ha al eady u ilizes se e s wi hou shu ing hem down o
main enance and upg ades. Wi h a NAS de ice, s o age is no an
in eg al pa o he se e . Ins ead, in his s o age-cen ic design,
he se e s ill handles all o he p ocessing o da a, bu a NAS
de ice deli e s he da a o he use .
Economic model and analysis o he cloud
BBA Uni e si a Poli ècnica de València
70
A NAS de ice does no need o be loca ed wi hin he se e , bu
can exis anywhe e in a Local A ea Ne wo k, and can be made up
o mul iple ne wo ked NAS de ices.
OS: The Ope a ing Sys em, is he mos impo an p og am o a compu e , as
i is he one ha manages all he o he p og ams and applica ions, alloca ing
he a ailable esou ces and aking ca e o he secu i y o he compu e .
RAID: The Redundan A ay o Independen Disks p o ides a way o
s o ing in di e en places he same da a, so he eading and w i ing speed o
he ope a ions can be ca ied on as e , and aul ole ance and edundancy a e
inc eased.
SAN: A s o age a ea ne wo k (SAN) is a dedica ed ne wo k ha p o ides
access o consolida ed, block le el da a s o age. SANs a e p ima ily used o
enhance s o age de ices, such as disk a ays, ape lib a ies, and op ical
jukeboxes, accessible o se e s so ha he de ices appea like locally a ached
de ices o he ope a ing sys em. A SAN ypically has i s own ne wo k o
s o age de ices ha a e gene ally no accessible h ough he local a ea ne wo k
(LAN) by o he de ices.
So wa e: I is he a iable pa o a compu e , o med by ope a ing sys ems,
p og ams and applica ions ha un o e he ha dwa e.
KB, MB, GB, TB: All o hem a e uni s o measu emen o in o ma ion
(ne wo k o s o age capaci y o a sys em), mul iplie s o he uni by e.
Economic model and analysis o he cloud
BBA Uni e si a Poli ècnica de València
71
Re e ences
Aguila , Luis Joyanes. 2013. COMPUTACIÓN EN LA NUBE: No as pa a una
es a egia española en cloud compu ing. Re is a Del Ins i u o Español De Es udios
Es a égicos 1 (1).
ALACOT TORRES, MARCO. 2011. Implan ación De Una Pla a o ma De Cloud Compu ing.
Al o d, Ted, and Gwen Mo on. 2009. The economics o cloud compu ing. Booz Allen
Hamil on.
Al-Tehmazi, Dana. 2013. In e -cloud us model secu i y: Issues and challenges. Pape
p esen ed a P oceedings o he In e na ional Con e ence on Secu i y and
Managemen (SAM), .
Al mann, Jö. 667. A ma ke place and i s ma ke mechanism o ading commodi ized
compu ing esou ces. Annals o Telecommunica ions - Annales Des Télécommunica ions 65
(11): 653-67.
G id economics and business models ecu so elec ónico-en línea] : 6 h in e na ional wo kshop,
GECON 2009, del , he ne he lands, augus 24, 2009. p oceedings, eds. Rajkuma edi o
Buyya, Ome F. edi o Rana and Sp inge Link (Se icio en línea). Be lin,
Heidelbe g: Be lin, Heidelbe g : Sp inge Be lin Heidelbe g, 2009.
A am, Ma icela-Geo giana. 2014. Ad an ages and challenges o adop ing cloud
compu ing om an en e p ise pe spec i e. P ocedia Technology 12 : 529-34.
Boja, Că ălin, Paul Poca ilu, and C is ian Toma. 2013. The economics o cloud
compu ing on educa ional se ices. P ocedia-Social and Beha io al Sciences 93 : 1050-4.
Boja, Că. 1054. The economics o cloud compu ing on educa ional se ices. P ocedia -
Social and Beha io al Sciences 93 : 1050-4.
B umec, Sla en, and Ne en V ček. 2013. Cos e ec i eness o comme cial compu ing
clouds. In o ma ion Sys ems 38 (4): 495-508.
Chang, Vic o , Robe John Wal e s, and Ga y Wills. 2013. The de elopmen ha leads
o he cloud compu ing business amewo k. In e na ional Jou nal o In o ma ion
Managemen 33 (3): 524-38.
Chen, Yao, and Radu Sion. 2011. To cloud o no o cloud?: Musings on cos s and iabili y, eds.
Je ey Chase S., Am Abbadi ElACM.
De Al onso, Ca los, Miguel Caballe , Fe nando Al a uiz, and Ge mán Mol ó. 2013. An
economic and ene gy-awa e analysis o he iabili y o ou sou cing clus e
compu ing o a cloud. Fu u e Gene a ion Compu e Sys ems 29 (3): 704-12.
de Los Ángeles López, Ma ía, and Es e Albanese. 2014. Ges ión de iesgos pa a la
adopción de la compu ación en nube en en idades inancie as de la epública
a gen ina. Con adu ía y Adminis ación 59 (3) (, Vol.59(3), pp.61-88): 61-88.
Economic model and analysis o he cloud
BBA Uni e si a Poli ècnica de València
72
R. Peña-O iz, J. Domènech, J.A. Gil, and A. Pon . 2014. An app oach o economic
e alua ion o cloud-based applica ions.
Du kee, Da e. 2010. Why cloud compu ing will ne e be ee. (p ac ice)(essay).
Communica ions o he ACM 53 (5) (, Vol.53(5), p.62(8)): 62.
FONS GÓMEZ, FERNANDO JOSÉ. 2014. Cloud Compu ing: Ca ac e ización De Los
Impac os Posi i os Ob enidos Po La U ilización Del Modelo Cloud Compu ing Po Las
Pymes, Basado En La Tipología De Modelos De Negocio De Es e Tipo De Emp esas.
Fo es , Will, and Cha lie Ba hold. 2009. Clea ing he ai on cloud compu ing. Discussion
Documen om McKinsey and Company.
Ga cía Ga cía, And és. 2011. Cloud Compu ing PaaS Pla o m Cloud ComPaaS.
Ga cia Rod iguez, Pablo Jesus, Modes o Ramon Delgado T oya, and Giuseppe Blacio
Abad. 2014. Cen os de p ocesamien os de da os en la nube.
Gawai, NS, BV Chik e, VR Pandi , SM Ingale, and MS Bu le. 2014. Applica ion o
cloud compu ing as hpc pla o m o embedded sys ems.
Géczy, Pe e , No iaki Izumi, and Kôi i Hasida. 2012. Cloudsou cing: Managing cloud
adop ion. Global Jou nal o Business Resea ch 6 (2): 57-70.
Goi i, Inigo, Jo di Gui a , and Jo di To es. 2010. Cha ac e izing cloud ede a ion o
enhancing p o ide s' p o i . Pape p esen ed a Cloud Compu ing (CLOUD), 2010
IEEE 3 d In e na ional Con e ence on, .
Gup a, P ashan , A. See ha aman, and John Rudolph Raj. 2013. The usage and adop ion
o cloud compu ing by small and medium businesses. In e na ional Jou nal o
In o ma ion Managemen 33 (5): 861-74.
Hou, A us CY, Ching-Chin Che n, Houn-Gee Chen, and Yu-Chen Chen. 2011.
‘Mig a ing o a new i ual wo ld’: Explo ing MMORPG swi ching h ough human
mig a ion heo y. Compu e s in Human Beha io 27 (5): 1892-903.
Hsu, Jack Shih-Chieh. 2014. Unde s anding he ole o sa is ac ion in he o ma ion o
pe cei ed swi ching alue. Decision Suppo Sys ems 59 : 152-62.
Khajeh‐Hosseini, Ali, Da id G eenwood, James W. Smi h, and Ian Somme ille. 2012.
The Cloud adop ion oolki : Suppo ing cloud adop ion decisions in he en e p ise.
So wa e: P ac ice and Expe ience 42 (4): 447-65.
Kshe i, Ni . 2013. P i acy and secu i y issues in cloud compu ing: The ole o
ins i u ions and ins i u ional e olu ion. Telecommunica ions Policy 37 (4): 372-86.
Ku ze, Tobias, Ma kus Klems, Da id Be mbach, Alexande Lenk, S e an Tai, and
Ma cel Kunze. 2011. Cloud ede a ion. Pape p esen ed a P oceedings o he 2nd
In e na ional Con e ence on Cloud Compu ing, GRIDs, and Vi ualiza ion
(CLOUD COMPUTING 2011), .