HySOR: A Simulation Model for the Sharing of Risk in a Service Level Agreement-Aware Hybrid Cloud

Sei e , Michael; Kuehnel, S ephan
A icle — Published Ve sion
HySOR: A Simula ion Model o he Sha ing o Risk in a
Se ice Le el Ag eemen -Awa e Hyb id Cloud
Business & In o ma ion Sys ems Enginee ing
P o ided in Coope a ion wi h:
Sp inge Na u e
Sugges ed Ci a ion: Sei e , Michael; Kuehnel, S ephan (2024) : HySOR: A Simula ion Model o he
Sha ing o Risk in a Se ice Le el Ag eemen -Awa e Hyb id Cloud, Business & In o ma ion Sys ems
Enginee ing, ISSN 1867-0202, Sp inge Fachmedien Wiesbaden GmbH, Wiesbaden, Vol. 67, Iss. 4,
pp. 495-510,
h ps://doi.o g/10.1007/s12599-024-00890-7
This Ve sion is a ailable a :
h ps://hdl.handle.ne /10419/330624
S anda d-Nu zungsbedingungen:
Die Dokumen e au EconS o dü en zu eigenen wissenscha lichen
Zwecken und zum P i a geb auch gespeiche und kopie we den.
Sie dü en die Dokumen e nich ü ö en liche ode komme zielle
Zwecke e iel äl igen, ö en lich auss ellen, ö en lich zugänglich
machen, e eiben ode ande wei ig nu zen.
So e n die Ve asse die Dokumen e un e Open-Con en -Lizenzen
(insbesonde e CC-Lizenzen) zu Ve ügung ges ell haben soll en,
gel en abweichend on diesen Nu zungsbedingungen die in de do
genann en Lizenz gewäh en Nu zungs ech e.
Te ms o use:
Documen s in EconS o may be sa ed and copied o you pe sonal
and schola ly pu poses.
You a e no o copy documen s o public o comme cial pu poses, o
exhibi he documen s publicly, o make hem publicly a ailable on he
in e ne , o o dis ibu e o o he wise use he documen s in public.
I he documen s ha e been made a ailable unde an Open Con en
Licence (especially C ea i e Commons Licences), you may exe cise
u he usage igh s as speci ied in he indica ed licence.
h ps://c ea i ecommons.o g/licenses/by/4.0/
RESEARCH PAPER
HySOR: A Simula ion Model o he Sha ing o Risk in a Se ice
Le el Ag eemen -Awa e Hyb id Cloud
Michael Sei e •S ephan Kuehnel
Recei ed: 3 July 2023 / Accep ed: 17 Ap il 2024 / Published online: 13 Decembe 2024
The Au ho (s) 2024
Abs ac Mo e and mo e o ganiza ions a e conside ing
public cloud se ices o hei business. Func ional
imp o emen s, inno a i e ea u es, and s a egic ac o s
a e u he d i ing demand. The adop ion o public cloud
se ices ypically in ol es he in eg a ion in o exis ing IT
a chi ec u es and an es ablished se ice s uc u e ha is
ideally aligned wi h he non- unc ional equi emen s o he
business p ocesses o be suppo ed. Hyb id cloud p o ide s
mus be able o accommoda e a a ie y o di e en public
cloud p o ide s while ensu ing con inui y o se ice o
app op ia e compensa ion p io o implemen a ion. Exis -
ing li e a u e ocuses on he calcula ion and simula ion o
se ice a ailabili y, bu less on se ice c edi o business
p ocess ou age cos s o se ice composi ions. In conse-
quence, his pape p esen s a calcula ion and simula ion
model o he concep o ‘‘sha ing o isk’’ in Se ice Le el
Ag eemen (SLA)-awa e hyb id clouds (HySOR), ocusing
on he isk-sensi i e simula ion o he inancial impac on
hyb id cloud p o ide s and cus ome s. The model was
implemen ed as an R-based applica ion and e alua ed wi h
12 leading expe s in he ield, yielding in e es ing impli-
ca ions o heo y and p ac ice.
Keywo ds Risk simula ion Cloud adop ion Hyb id
cloud Se ice le el ag eemen
1 In oduc ion
Public cloud usage has g own s eadily in ecen yea s and
is o ecas o con inue o g ow e en u he , wi h So wa e
as a Se ice (SaaS) emaining he la ges segmen (Ga ne
Inc. 2024). As public cloud SaaS adop ions inc ease, hei
in eg a ion in o exis ing IT a chi ec u es owa ds hyb id
clouds mus be p ope ly managed (Sun e al. 2008).
Howe e , managing his in eg a ion poses a numbe o
challenges o o ganiza ions, e.g., changes o con ac ual
commi men s, di e en de ini ions and o mula ions o
Se ice Le el Ag eemen s (SLA), o he economic e alu-
a ion o SLA b eaches (Sei e e al. 2023), o name jus a
ew. The a ious SLAs o di e en cloud p o ide s o e
cus ome s a con ac ual basis o assessing he se ice
commi men (Aljoumah e al. 2015; Yuan e al. 2015). Fo
his pu pose, howe e , he con inui y- ele an componen s
mus i s be o malized (Sei e 2021) be o e being
agg ega ed ac oss he a ious e ical and ho izon al in e-
g a ion pa e ns (B ei e and Naik 2013) and he se e al
cloud a chi ec u e le els (Comuzzi e al. 2009) up o he
hyb id cloud composi ion (Theilmann e al. 2010).
An essen ial aspec o such hyb id cloud a chi ec u es is
he one- o-many ela ionship be ween he (one) p o ide o
public cloud SaaS and he (many) cus ome s, including he
p o ide o he hyb id cloud composi ion (Pan and Mi chell
2015; Sei e e al. 2023). The possibili y o oppo unis ic
beha io o se e al addi ional public cloud p o ide s (Pan
and Mi chell 2015) leads o inc eased isks in con ac ual
commi men and pa ne ship, which should be conside ed
in he inancial assessmen o he desi ed hyb id cloud
a chi ec u e (Sei e e al. 2023). Such an a chi ec u e o en
has o deal wi h di e en nego iabili y ypes o SLAs
(K i ikos e al. 2016) and a esul ing isk ega ding he
adequacy o he op-le el SLA o he se ice composi ion
Accep ed a e wo e isions by O
´sca Pas o .
M. Sei e (&)S. Kuehnel (&)
Chai o In o ma ion Sys ems, esp. Business In o ma ion
Managemen , Ma in Lu he Uni e si y Halle-Wi enbe g,
Uni e si ae s ing 3, 06108 Halle, Ge many
e-mail: [email p o ec ed]
S. Kuehnel
e-mail: [email p o ec ed]
123
Bus In Sys Eng 67(4):495–510 (2025)
h ps://doi.o g/10.1007/s12599-024-00890-7
(Comuzzi e al. 2013). This may u he lead o isks in he
execu ion o he business p ocesses being suppo ed.
Assessing he inancial isk o ailu e o ex e nal (cloud)
se ices is al eady e lec ed in a ious isk calcula ions and
simula ions, as can be seen, o example, in Jiang e al.
(2013), Yuan e al. (2015), and Wang and F anke (2020).
Howe e , exis ing models discussed in he li e a u e lack
he conside a ion o unce ain ies ha go beyond a ail-
abili y pa ame e s, as desc ibed as a ele an issue by
F anke e al. (2013) and Johnson e al. (2014). In addi ion,
he conside a ion o business-c i ical SLA pa ame e s in
ecen public cloud SLAs is conside ed a p omising
ex ension o isk calcula ion and simula ion app oaches
(Sei e 2021).
The ocus o his pape is on he assessmen o hyb id
cloud se ice composi ions p io o implemen a ion, which
is mo i a ed by a unc ional o s a egic decision (Sei e
e al. 2023). The p o ide – whe he in e nal o ex e nal –
has o deal wi h he isk o newly added (‘‘incoming’’)
public cloud se ices. The decisions on p icing o penal ies
o ailing he op-le el SLA mus be made in a long- e m
pa ne ship be ween he p o ide and he cus ome (Ben-
lian e al. 2011; Goo e al. 2009), e en i his leads o
unaccep abili y o a ‘‘don’ do i ’’ decision. In his con ex ,
his pape p esen s a new isk calcula ion and simula ion
model and examines he ollowing esea ch ques ions
(RQ):
•RQ1: Wha a e he co e elemen s and business-c i ical
SLA pa ame e s d i ing isk in hyb id cloud
a chi ec u es?
•RQ2: How can he impac o simula ed IT ou ages o
pa icipa ing componen s be aken in o accoun in he
isk calcula ion o hyb id cloud a chi ec u es and how
does his a ec he expec ed cos s o p o ide s and
cus ome s?
•RQ3: How can he inancial impac o p o ide s and
cus ome s be calcula ed and ela ed o he a ailabili y
om he se ice composi ion’s op-le el SLA o
p o ide decision suppo o he sha ing o isk?
The u he s uc u e o his pape is based on he
Design Science Resea ch (DSR) p ocess o Pe e s e al.
(2007). Acco dingly, we i s p esen ela ed wo k ound
wi hin his mul i-s age esea ch p ojec , add essing he
‘‘Objec i es o a Solu ion’’ phase o DSR in Sec . 2. Nex ,
he calcula ion model o he sha ing o isk o SLA-awa e
hyb id clouds (HySOR) is p esen ed in Sec . 3, which
co esponds o he i s pa o he ‘‘Design & De elop-
men ’’ phase. In addi ion, he simula ion aspec s a e also
p esen ed he e. The second pa (wi h a s onge ocus on
de elopmen ) is ep esen ed by he implemen a ion o he
HySOR model in an R-based applica ion buil on he
‘‘Shiny’’ lib a y, along wi h he demons a ion using a case
s udy in Sec . 4. A su ey o 12 leading expe s in he ield
was conduc ed o ob ain a summa i e e alua ion, which is
documen ed in Sec . 5and ep esen s he ‘‘E alua ion’’
phase o DSR. A second pa o his phase can be ound in
he subsequen Sec . 6, which con ains implica ions o
p ac ice condensed om he ollow-up expe in e iews.
The limi a ions o his wo k, as well as oppo uni ies o
u he esea ch, a e p esen ed in Sec . 7. The pape closes
wi h a conclusion, answe s o he esea ch ques ions, and a
p esen a ion o he con ibu ions o heo y and p ac ice in
Sec . 8.
2 Theo e ical Backg ound and Rela ed Wo k
In his chap e , we d aw on ou own p e ious esea ch (see
Sei e e al. 2023), in which we conduc ed a comp ehen-
si e sys ema ic li e a u e e iew (SLR) o esea ch on
hyb id clouds esul ing om he adop ion o SaaS. The
s udy iden i ied six majo challenges in dealing wi h hyb id
clouds esul ing om public cloud adop ion, which we use
as a basis o his concep ualiza ion. I is assumed ha
u he esea ch is needed, especially in he a ea o mod-
eling business-c i ical Quali ies o Se ice (QoS), such as
se ice commi men o se ice c edi and hei agg ega ion
in hyb id cloud composi ions (Sei e e al. 2023). Fu -
he mo e, he conside a ion o business and echnological
unce ain ies as well as he changing commi men o he
in ol ed cloud p o ide s se es as a esea ch gap o he
ex ension o he exis ing knowledge base (Sei e e al.
2023).
Fo consis en applica ion o he exis ing li e a u e o he
iden i ied esea ch gap, we conduc ed a second sho SLR
in p epa a ion o his design cycle wi h a special ocus on
isk calcula ion and simula ion o cloud a chi ec u es and
se ice composi ions. We included esea ch wi h a ocus
on he ollowing opics (inclusion c i e ia): (I) ‘‘se ice
down ime’’, ‘‘se ice ou age’’, and ‘‘se ice ailu e’’ o
add ess a ailabili y simula ion and agg ega ion. (II) ‘‘Se -
ice c edi ’’ and ‘‘se ice penal y’’ o e lec business
p ocess ou ages and o calcula ing he inancial impac on
he cus ome side as well. We delibe a ely do no conside
he con ac ual dis inc ion be ween penal y and c edi , as
we see bo h as a paymen om he p o ide o he (sub-
sc ip ion) cus ome . And (III), ‘‘cloud’’, o ensu e ha
bo h he cha ac e is ics o he di e en dep hs o nego ia-
bili y ( isking he business con inui y), and mul iple cloud
componen p o ide s in ol ed in he
composi ion a e add essed.
We excluded esea ch wi h a ocus on he ollowing:
(I) Op imiza ion in e ms o pe o mance and cos
(e.g., Gue
´ ou e al. 2014; Ma eo-Fo nes e al. 2019), (II)
dynamics and SLA (e.g., Al-Ghuwai i e al. 2016; Faniyi
123
496 M. Sei e , S. Kuehnel: HySOR: A Simula ion Model o he Sha ing o Risk..., Bus In Sys Eng 67(4):495–510 (2025)
e al. 2012), and (III) isk assessmen based on ansac ion
his o y (e.g., Hussain e al. 2010).
Based on he esul s o he SLRs, he ollowing sub-
sec ions de ine and de i e bo h e minology and equi e-
men s ha a e used o concep ualize he HySOR isk
calcula ion and simula ion model. The equi emen s a e
highligh ed in i alics a he end o each sub-chap e .
2.1 Cloud Se ice Componen s, Cloud Se ice
Composi ions, and he Hyb id Cloud A chi ec u e
When desc ibing hyb id cloud a chi ec u es, a dis inc ion
mus be made be ween indi idual componen s and di e -
en composi ions o componen s in ol ed in he p o ision
o a cloud se ice. To his end, we build on he wo ks o
Comuzzi e al. (2009), Labidi e al. (2016), and Sei e and
Kuehnel (2021), ha use a chi ec u e modeling o design
se ice composi ions in o de o o mally map hyb id
clouds and cap u e SLA-c i ical isk aspec s (e.g., a ail-
abili y). The ela ed business and echnology pe spec i es
a e also p omising o unde s anding he impac on eco-
nomic isks (Comuzzi e al. 2013; Sei e and Kuehnel
2021).
A se ice componen is a single IT se ice ha ul ills a
speci ic unc ion, is equi ed o he execu ion o a business
p ocess, and has an SLA. Se ice componen s can occu in
any cloud se ice o cloud deploymen model. In con as ,
a se ice composi ion can be desc ibed as a combina ion o
se e al se ice componen s, whe eby hese a e di ided in o
ho izon al and e ical componen s depending on he ype
o in eg a ion (Sei e e al. 2023). Ho izon al in eg a ion is
he in e ac ion o IT se ices o he same ype om di -
e en sou ces (B ei e and Naik 2013), such as wo SaaS
componen s, whe e one is ob ained om an exis ing p i a e
cloud and one om a public cloud. In con as , e ical
in eg a ion means ha one se ice depends on ano he o
consumes i (B ei e and Naik 2013), such as when SaaS is
buil on in as uc u e as a se ice (IaaS).
To pu i simply, se ice composi ions e e o he a -
ious possible combina ions o di e en clouds and can
desc ibe, o example, mul i-clouds o he hyb id clouds
ele an o his wo k. In addi ion, he simple use o public
cloud se ices alongside he exis ing IT landscape usually
al eady leads o hyb id clouds, whe eby he indi idual
componen s di e in e ms o hei nego iabili y (e.g.,
a ailabili y). Since ou isk model is applicable o bo h
ypes o cloud deploymen s, we use he e m cloud se ice
composi ion as a gene ic e m.
HySOR has o conside an a chi ec u al model con-
sis ing o (i) se ice componen s ha may ep esen exis -
ing IT componen s used by an o ganiza ion and (ii)
incoming public cloud se ices ha esul in se ice com-
posi ions ha suppo he business p ocess.
2.2 Cus ome and P o ide in Di e en Roles
The e a e also o ganiza ional aspec s o hyb id cloud
a chi ec u es ha need o be aken in o accoun and can
become a challenge, e.g., wi h ega d o he di e en oles.
‘‘These challenges become e en mo e complex when cloud
p o ide s ake on he dual ole o p o ide and cus ome ,
o example, when hei own cloud se ice o e ings build
on he se ices o ex e nal cloud p o ide s’’ (Sei e and
Kuehnel 2021). Rega dless o whe he i is an in e nal o
ex e nal cloud se ice p o ide , someone needs o be
esponsible o he op-le el SLA and discuss he isk o
business p ocess ou ages wi h hei cus ome s (Sei e and
Kuehnel 2021).
Subsc ip ion is a e y common p icing model o public
cloud se ices (Maz ekaj e al. 2016). ‘‘Wi h subsc ip ion
p icing, use s pay on a ecu ing basis o access so wa e as
an online se ice o o bene i om a se ice’’ (Maz ekaj
e al. 2016). The ole dis inc ion in his model is c ucial o
de e mining who holds he subsc ip ion o he cloud
(Zhang and Zhou 2009). Subsc ip ion in he open a chi-
ec u e o cloud compu ing consis s, among o he hings, o
a p ocess and he oles (Zhang and Zhou 2009). Whoe e
subsc ibes o he cloud se ice is a con ac ual pa ne ,
which is also a decisi e isk elemen o he calcula ion o
penal ies.
HySOR has o conside (i) he hyb id cloud cus ome
and (ii) he hyb id cloud p o ide as he op-le el con-
ac ing pa ies, wi h (iii) one o hem subsc ibing o each
o he cloud se ices in ol ed in he se ice composi ion.
2.3 Se ice Le el Ag eemen and Ope a ional Le el
Ag eemen Fo maliza ion
A co e ace o hyb id cloud a chi ec u e in ol es SLA and
ope a ional le el ag eemen (OLA) aspec s (Sei e e al.
2023). To quan i y a ailabili y as a isk me ic, we need o
o malize SLA/OLA pa ame e s, including hei measu -
able and calculable pa ame e s called QoS (Suakan o e al.
2012).
A ailabili y is one o he mos impo an a ibu es o
cloud se ice quali y, and mos popula public cloud se -
ices claim hei a ailabili y p omise (Base 2012; Gulia
and Sood 2013; Sei e 2021). A ailabili y is o en
desc ibed no only by a numbe bu also by di e en
pa ame e s.
The ca ego ies o Yuan e al. (2015) desc ibe his
app op ia ely o ou con ex . The i s pa ame e is he
measu emen pe iod, usually de ined as one mon h. The
se ice g anula i y de ines which se ice scope is mean ,
and he ime g anula i y is usually speci ied in he o m o
1, 5, o 10 min. The co e age desc ibes wha mus be
unning co ec ly and which se ices mus be included. In
123
M. Sei e , S. Kuehnel: HySOR: A Simula ion Model o he Sha ing o Risk..., Bus In Sys Eng 67(4):495–510 (2025) 497
addi ion, exclusions o una ailabili y a e commonly
de ined.
Yuan e al. (2015) p o ide a use ul o maliza ion o he
penal y unc ion: a) ‘‘ o al cha ge a io’’, b) ‘‘ ixed alue a
di e en iola ion le els’’, o c) ‘‘down ime a io’’. We can
also see he public cloud penal y in he ‘‘se ice c edi ’’
ca ego y o Sei e (2021). Mo eo e , we can ind he
ypical se ice penal y calcula ion o public cloud SaaS in
Sei e (2021) as a combina ion o he me hods om Yuan
e al. (2015). An example is he so wa e company SAP
wi h he a io o o al cha ge and he a io o down ime:
‘‘pe 1% below a ailabili y (99.5) you ge a c edi o 2% o
you mon hly ee’’ (Sei e 2021). In addi ion, he maxi-
mum c edi olume is gi en, which is also c ucial o isk
assessmen .
Ano he impo an aspec o he SLA is i s nego iabili y
(Comuzzi e al. 2013; K i ikos e al. 2016), which only has
a seconda y e ec on he o maliza ion, i.e., as a o mu-
la ion o ixed pa ame e s in public cloud SLAs (Sei e
2021). The nego iabili y o he SLA plays a decisi e ole in
he desc ip ion o he ‘‘sha ing o isk’’ in Sec . 2.6.
HySOR has o (i) simula e se ice con inui y based on
a ailabili y commi men pa ame e s and (ii) enable in e-
g a ed calcula ion o ypical penal y unc ions om cloud
SLAs so ha hese (iii) can be agg ega ed in he se ice
composi ion.
2.4 Unce ain y as a Risk D i e
Unce ain y has o be aken in o accoun in hyb id cloud
a chi ec u es (Johnson e al. 2014). In he ca ego y
‘‘unce ain y in echnology and business’’, Sei e e al.
(2023) dis inguish h ee dimensions o isk o unce ain y
in his con ex . Fi s , angible isks, such as a ailabili y
(Paque e e al. 2010), can be ep esen ed as a p obabili y
dis ibu ion, e.g., om he QoS his o y (Johnson e al.
2014). Second, he e a e in angible isks when he business
p ocess depends on an ex e nal cloud elemen (Paque e
e al. 2010). This may lead o oppo unis ic beha io by he
public cloud p o ide due o a one- o-many ela ionship in
he hyb id cloud con ex (Pan and Mi chell 2015). Thi d,
unce ain y ega ding he knowledge o he a chi ec u e
suppo ing he cloud composi ions and business p ocesses
(F anke e al. 2013; Johnson e al. 2014; Rockmann e al.
2014) is an in angible isk d i e and mus also be
conside ed.
HySOR has o conside (i) angible isks, (ii) in angible
isks a ising om possible oppo unis ic beha io o he
cloud componen p o ide s in ol ed, and (iii) in angible
isks a ising om he business echnology a chi ec u e.
2.5 Rela ed Wo k on Risk Calcula ion and Simula ion
Wi hin he SLRs, we ound h ee ela ed pape s on isk
calcula ion o simula ion models based on a chi ec u es
wi h s ong ele ance o ou con ex .
Yuan e al. (2015) mo i a e hei app oach wi h he lack
o cla i y in a ailabili y commi men and penal y o cloud
consume s, and he business model o cloud p o ide s o
ind he op imal penal y le el. The ipa i ion o possible
penal y unc ions, i.e., (I) a io o o al cha ge, (II) ixed
alue, and (III) a io o down ime, seems p omising o he
de elopmen o isk calcula ion and simula ion models.
The inancial impac o cus ome cos as p o ide p ice
oge he wi h he impac o down ime appea s easonable.
The impo an inding ha he p o ide will educe he
penal y in o de o compensa e o highe a ailabili y
equi emen s o claims leads us o he assump ion made
la e in his pape ha be e isk sha ing (e.g., mo e
equally sha ed isks) is a gap in esea ch o da e. Due o he
lack o conside a ion o he composi ion (and he compo-
si ion p o ide ), i.e., he combina ion o se e al compo-
nen s and hei agg ega ion, his isk calcula ion is no
adequa e o ou con ex wi h he abo e-men ioned
concep s.
Jiang e al. (2013) mo i a e a QoS-based isk app oach
combined wi h business-o ien ed a ge moni o ing. The
i e-s age p ocedu e o inding and pa ame e izing a
sui able ailu e p obabili y dis ibu ion con i ms he
necessi y o modeling based on dis ibu ions. In pa icula ,
pa ame e iza ion is an in e es ing aspec o isk sensi i i y
o assumed ailu e p obabili ies in o de o es decision
suppo in a ian s. The lack o ocus on he p o ide ’s
e enue and loss o e enue (e.g., by ocusing only on
educing usage) is an incen i e o de eloping new
me hods.
Wang and F anke (2020) p esen a model o analyzing
IT se ice ou ages o indi idual o ganiza ions and supply
chains. An impo an pa o hei model is he cos o
business p ocess down ime (i.e., economic impac on he
cus ome side) in e ms o h ee unc ion ypes: cons an ,
linea , and quad a ic. Ano he use ul aspec is he e-
quency o b eakdowns/down ime, ep esen ed by a Poisson
a i al model – as ound in he li e a u e. On his basis, he
down ime du a ion is modeled wi h a logno mal dis ibu-
ion, as is o en used when modeling down imes. The lack
o conside a ion o unce ain ies a ising om business and
echnology (e.g., suppo o business p ocesses by IT, see
Sec . 2.4) is a poin ha equi es adjus men . Fu he mo e,
he conside a ion o oday’s public cloud penal y unc ions
o compensa ion on he cus ome side and cos s on he
p o ide side can play an impo an ole.
HySOR needs o (i) conside di e en business p ocess
down ime cos unc ions and demons a e he inancial
123
498 M. Sei e , S. Kuehnel: HySOR: A Simula ion Model o he Sha ing o Risk..., Bus In Sys Eng 67(4):495–510 (2025)

implica ions o (ii) he hyb id cloud composi ion cus ome
and (iii) he p o ide .
2.6 The Concep s ‘‘Sha ing o Risk’’ and ‘‘Zone
o Possible Ag eemen s’’
Finally, ou app oach also di e s om ela ed wo k as we
do no aim a an economic op imiza ion model o cloud
p o ide s. Fo example, Wang and F anke (2020) s a e ha
hei ‘‘in ui ion behind he model is ha capi al K can buy
be e ha dwa e, he eby educing he equency o down-
ime’’. HySOR, howe e , does no p ima ily ocus on
in es men oppo uni ies o educe down ime, as we
assume a ypical ‘‘ ake i o lea e i ’’ scena io o public
cloud adop ion, as desc ibed by Comuzzi e al. (2013).
Combined wi h he assump ion o a desi ed pa ne ship
be ween he hyb id cloud cus ome and p o ide , his leads
o he need o app op ia ely judge he hyb id cloud a chi-
ec u e. The ollowing model he e o e ocuses on he
newly p oposed concep ‘‘sha ing o isk’’. This concep
opens up a nego ia ion co ido ha allows hyb id cloud
cus ome s and p o ide s o accoun o angible and
in angible decision ac o s by a ying isk pa ame e s. The
sha ing o isk e lec s he ange o isky inancial conse-
quences o he p o ide and he cus ome o a cloud se -
ice composi ion. To his end, we ely on one o he mos
well-known desc ip i e nego ia ion concep s – Rai a
(1982)’s Zone o Possible Ag eemen s (ZOPA) (see
Fig. 1). Wi h his concep , Rai a (1982) ep esen s a wo-
pe son dis ibu i e ba gaining p oblem ha is bounded by
he pa ies’ ese a ion p ices ( espec i ely hei bes cases/
al e na i es) (Ahle and S a
¨ e 2016).
I he buye ’s ese a ion p ice is g ea e han ha o he
selle , he e is a ZOPA con aining a possible ag eemen
alue (Ahle and S a
¨ e 2016; Rai a 1982). I he ese -
a ion p ices o bo h pa ies a e he same, he e is one
possible poin o ag eemen , whe eby in his case he gains
o bo h pa ies a e ze o. I he selle ’s ese a ion p ice is
highe han he buye ’s, he e is no ZOPA (Ahle and
S a
¨ e 2016).
Applied o he con ex o his s udy, we imagine a si -
ua ion in which a cus ome and a p o ide o a hyb id cloud
se ice composi ion en e in o nego ia ions abou isky
inancial consequences (ins ead o p ices). The si ua ion is
mo e complex han in Rai a (1982)’s o iginal model o
p ice nego ia ion, as he isky inancial consequences o
bo h pa ies a e in luenced by a ious aspec s.
Two main aspec s lead o inc eased isk and equi e
sha ing as pa o a long- e m pa ne ship be ween hyb id
cloud p o ide and cus ome . Fi s , he non-nego iabili y o
he incoming public cloud componen s in ol ed in he
composi ion inc eases he isk o an ou age. Second, he
unde lying penal y unc ions o hese componen s may be
insu icien o compensa e o he inancial consequences in
he wo s case. Fo example, passing he subsc ip ion o a
cloud componen on o he p o ide could educe he
p o ide ’s isk, while inc easing he cus ome ’s ( inancial)
isk o p ocess ou ages. Howe e , his could be aken in o
accoun when nego ia ing he se ice c edi o he p icing
o he se ice composi ion be ween he hyb id cloud
cus ome and he p o ide .
HySOR mus conside isk-sensi i e simula ion pa am-
e e s o p edic , e alua e, and compa e a ian s o he
desi ed hyb id cloud composi ion a chi ec u e.
3 The Risk Calcula ion and Simula ion Model HySOR
In he ollowing, HySOR and i s di e en elemen s a e
desc ibed conce ning he calcula ion and he simula ion
model, which ensu es he anspa ency and comp ehensi-
bili y o ou DSR p ojec . Mo eo e , a de ailed desc ip ion
o he a i ac allows he ields o u u e esea ch p esen ed
in Sec . 6and 7 o be add essed in a a ge ed manne . Fo
example, o he scien is s can exchange speci ic simula ion
componen s o in eg a e addi ional calcula ion modules.
Fig. 1 Rai a’s desc ip i e nego ia ion concep (illus a ion adap ed om Ahle and S a
¨ e (2016)
123
M. Sei e , S. Kuehnel: HySOR: A Simula ion Model o he Sha ing o Risk..., Bus In Sys Eng 67(4):495–510 (2025) 499
3.1 Co e Elemen s and Cha ac e is ics
The co e elemen s o HySOR a e a se ice composi ion s
consis ing o se ice componen s i ha a e in ol ed and
unc ionally necessa y o suppo he business p ocess and
i s execu ion. The model e e s o mmon hs o a con ac
du a ion M. The QoS o he pa icipa ing se ice compo-
nen s is hen simula ed o all hou s (n) o his con ac
du a ion and hen agg ega ed o he QoS o he se ice
composi ion.
Co e elemen s:
I: nonemp y ini e se o se ice componen s
i2I: a se ice componen om he se o I
sI: a se ice composi ion ep esen ing a subse o I
(i.e., consis ing o i= 1,2,3,...,q se ice componen s)
m2 1,2,3;...;Mg: mon h (m) o a con ac du a ion
(M) o he se ice composi ion (s)
n2 1;2;3;...;720 Mg: hou (n) o a con ac du a-
ion (M) o he se ice composi ion (s)
The basic cha ac e is ics o he se ice componen s
include he subsc ip ion owne ship, he mon hly cos o he
se ice, he a ailabili y pe mon h gua an eed in he SLA,
and he penal y amoun o missing he SLA (see he lis ed
se ice componen cha ac e is ics below). The angible
isks a e modeled using he isk pa ame e s lambda o he
Poisson a i al o he down ime and he expec ed alue and
a iance o he logno mal dis ibu ion o he down ime
du a ion. The speci ic cha ac e is ics o hese pa ame e s
can ei he be de i ed h ough benchma king and om
his o ical da a (e.g., QoS his o y, as desc ibed in Sec . 2.4)
o mus be es ima ed by expe s. As also desc ibed in
Sec . 2.4, insu icien knowledge o he business echnol-
ogy a chi ec u e and i s ele ance o business p ocess
suppo is an in angible isk d i e ha needs o be con-
side ed. Ambigui ies in he dependency o he business
p ocess on a componen a e he e o e aken in o accoun as
a isk ac o and a e included in he model as a secu i y
pe cep ion pa ame e . The se ice componen s also ea u e
down ime simula ion and mon hly a ailabili y agg ega ion
capabili ies. Finally, he e a e di e en ypes o penal y
unc ions ha can be modeled o each componen .
Se ice componen cha ac e is ics:
oi: ype o subsc ip ion owne ship o he se ice com-
ponen (i)
ci: mon hly cos s o he se ice componen (i)
si: in he SLA gua an eed a ailabili y o he se ice
componen (i) pe mon h
pai: penal y amoun o he se ice componen (i) o
missing he SLA
ki: expec ed alue and a iance o he Poisson a i al o
he down ime o a se ice componen (i)
li: expec ed alue o he down ime du a ion o se ice
componen (i)
i: s anda d de ia ion o he down ime du a ion o
se ice componen (i)
l¼ln l2
i
ffiffiffiffiffiffiffiffiffiffi
l2
iþ 2
i
p

:lpa ame e o he logno mal dis i-
bu ion
2¼ln 1 þ 2
i
l2
i

:
2
pa ame e o he logno mal
dis ibu ion
idi¼0;i i dispensable
1;else
: indispensabili y o he se -
ice componen (i) o he business p ocess
ii2½0,1: isk ac o ega ding he indispensabili y o
he se ice componen (i) o he business p ocess
ei¼idið1 iiÞ: secu i y pe cep ion pa ame e o he
indispensabili y o he se ice componen (i) o he
business p ocess
dai;n¼ ðkiÞ: simula ed down ime a i al o he se ice
componen (i) pe hou (n)
di;n¼ ðdai;l; Þ: simula ed down ime du a ion o he
se ice componen (i) pe hou (n)
ai;m¼ ðdi;mÞ: calcula ed a ailabili y o he se ice
componen (i) in pe cen pe mon h (m)
pi;m¼ ðai;m;si;paiÞ: penal y unc ion o he se ice
componen (i) o missing he SLA pe mon h (m)
The se ice composi ion has speci ic cha ac e is ics, such
as he mon hly cos o he cus ome , which is equi alen o
he composi ion e enue o he p o ide . The se ice
composi ion i sel has a de ined a ailabili y in he so-called
op-le el SLA, wi h which he agg ega ion o he a ailabil-
i ies o he componen s in ol ed is la e compa ed o mea-
su e SLA compliance. Analogous o he componen s, he
composi ion has a penal y amoun and a co esponding
penal y unc ion.
Ano he essen ial dimension o he se ice composi ion
cha ac e is ics is he modeling o business isk (and ela ed
cos s). Acco ding o ou concep o ‘‘sha ing o isk’’, he
isk o penal ies on he p o ide ’s side mus be balanced
wi h he isk o business p ocess ou ages on he cus ome ’s
side, bo h o which a e connec ed ia he se ice compo-
si ion and co esponding penal y and cos unc ions. The
123
500 M. Sei e , S. Kuehnel: HySOR: A Simula ion Model o he Sha ing o Risk..., Bus In Sys Eng 67(4):495–510 (2025)
business p ocess ou age cos unc ion ypes o composi-
ions a e ei he cons an , linea , o quad a ic, as desc ibed
in Sec . 2.5.
Se ice composi ion cha ac e is ics:
cs: mon hly cos s o he se ice composi ion (s)
ss: in he ( op-le el) SLA gua an eed a ailabili y o he
se ice composi ion (s) pe mon h
pas: penal y amoun o he se ice composi ion (s) o
missing he SLA
ds;n¼ ðdiÞ: agg ega ed down ime du a ion o he se -
ice composi ion (s) pe hou (n)
as;m¼ ðdsÞ: calcula ed a ailabili y o he se ice com-
posi ion (s) in pe cen pe mon h (m)
ps;m¼ ðas;m;ss;pasÞ: penal y amoun o se ice com-
posi ion (s) o missing he SLA pe mon h (m)
b ype
n¼ ðds;eiÞ: business p ocess ou age cos unc ion
ype, depending on he una ailabili y o he se ice
composi ion (s) pe hou (n)
bm¼ ðbnÞ: business p ocess ou age cos s caused by he
una ailabili y o he se ice composi ion (s) pe mon h
(m)
3.2 Calcula ion and Simula ion P ocedu e
As al eady men ioned abo e, we assume a Poisson a i al
p ocess as he basis o he a ailabili y simula ion o se -
ice componen s, in line wi h Wang and F anke (2020).
The down ime du a ion is simula ed sepa a ely o each
se ice componen using a logno mal dis ibu ion.
Rema k 1 We assume ha se ice componen s ha ing a
dedica ed SLA also ha e independen ou ages, as hey each
ha e di e en ia ed a chi ec u es and esilience p ocedu es
(co esponding o he SLA o e ed).
Down ime a i al pe se ice componen pe hou :
dai;n¼daki¼ nðÞ¼ðkÞn
n!ek
;
8n2 1;2;3;...;720 Mg
Down ime du a ion pe se ice componen pe hou :
di;n¼ da
i;n

¼1
dai ffiffiffiffiffiffi
2p
pexp ln dai
ðÞlðÞ
2
2 2
!
;
o dai;n[0;
8n2 1;2;3;...;720 Mg
The agg ega ion o he se ice componen s in o a se -
ice composi ion wi h ega d o he down ime du a ion is
de e mined by he componen wi h he maximum down-
ime pe hou (ac oss all componen s). As pa o he la e
a ailabili y agg ega ion, he alues a e hen summed up o
he mon h (wi h 720 hou s pe mon h).
Rema k 2 We assume ha each mon h consis s o
30 days o 24 hou s each and ha a s anda d con ac
yea comp ises 12 mon hs. This is based on he assump ion
ha business p ocess ou ages ha e he same inancial
impac each mon h, as he associa ed cos s a e independen
o he occu ence o down ime du ing he con ac pe iod.
Down ime du a ion agg ega ion pe se ice composi ion
pe hou :
ds;n¼ d
i;n

¼max ðdi;nÞ;
8i2 1;2;3;...;qg;
8n2 1;2;3;...;720 Mg
The agg ega ion o he down ime du a ion pe hou is
also c ucial o de e mining he amoun o he business
p ocess ou age cos s. This is because wo o mo e com-
ponen s can ail simul aneously, esul ing in he same
se ice composi ion ou age.
Rema k 3 We ha e a limi a ion in he case whe e a
down ime a i al coincides wi h he down ime du a ion o
an ea lie down ime a i al. Due o a mode a e a i al
a e pe hou , we assume his o be negligible.
The business p ocess ou age cos s a e calcula ed pe
hou o he con ac du a ion and subsequen ly agg ega ed
o each mon h. The penal ies a e calcula ed o each
componen and o he composi ion based on he espec i e
a ailabili y agg ega ion. This pape illus a es he calcula-
ion o he mon hly penal y.
A ailabili y agg ega ion pe se ice componen pe
mon h:
ai;m¼ d
i;n

¼720 Pdi;n
720 ;
o 0 n 721 !m¼1; o 720 n 1441 !m
¼2;...;!m¼M
A ailabili y agg ega ion pe se ice composi ion pe
mon h:
as;m¼ d
s;n

¼720 Pds;n
720 ;
o 0 n 721 !m¼1; o 720 n 1441 !m¼
2;...;!m¼M
Business p ocess ou age cos s pe hou :
123
M. Sei e , S. Kuehnel: HySOR: A Simula ion Model o he Sha ing o Risk..., Bus In Sys Eng 67(4):495–510 (2025) 501
bcons an
da ¼ d
s
ðÞ¼ba;e
i;i ds;n[0;
blinea
da ¼ d
s
ðÞ¼ds;nba;e
i;
bquad a ic
da ¼ d
s
ðÞ¼ds;n2ba;e
i;
8n2 1;2;3;...;720 Mg
Business p ocess ou age cos agg ega ion pe mon h:
bm¼ b
da
ðÞ¼
Xbda;
o 0 n 721 !m¼1; o 720 n 1441 !m
¼2;...;!m¼M
Penal y calcula ions pe mon h:
p ix
k;m¼ a
k;m;sk

¼ak;m sk!pai;else !0;
ppe cen age
k;m¼ a
k;m;sk;ck

¼max 2ckdskak;me;ck

;
pnone
k;m¼ a
k;m;sk;ck

¼0;
k2 i;sg;
8m2 1,2,3;...;Mg
Following he p e ious calcula ions a he le els o
componen s, composi ions, and business p ocesses, he
inancial impac o he cus ome ( o alcos scus Þand he
p o ide ( o alcos sp o Þo he se ice composi ion can be
de e mined. This is done by dis inguishing be ween ixed
cos s, which a e incu ed on a mon hly basis ega dless o
he simula ed a ailabili y o he componen s, and a iable
cos s, which a e composed o business p ocess cos s and/o
he penal ies o he se ice componen s and composi ion.
Fixed cos s pe mon h:
ixcos scus ;m¼ c
s;ci
ðÞ¼csþX
n
i¼0
ci;
8oi2 cus ome public cloud subsc ip iong;
ixcos sp o ;m¼ c
s;ci
ðÞ¼csþXn
i¼0ci;
8oi2 p o ide public cloud subsc ip ion;p o ide p i a e
cloud componen g;
8m2 1,2,3;...;Mg
Rema k 4 Fo easons o comp ehensibili y, he ixed
cos s o he composi ion ðcsÞ ha a cus ome pays o he
p o ide we e modeled in he p o ide ’s calcula ion wi h
he same a iable, bu wi h he ma hema ical sign e e sed,
i.e. as nega i e cos s ðcsÞ
Va iable cos s pe mon h:
a cos scus ;m¼ p
i;m;ps;m;bm

¼pi;mps;mþbm;
8oi2 cus ome public cloud subsc ip ion;p o ide
p i a e cloud componen g
a cos sp o ;m¼ p
i;m;ps;m

¼pi;mþps;m;
8oi2 p o ide public cloud subsc ip iong;
8m2 1,2,3;...;Mg
Rema k 5 Fo easons o comp ehensibili y, penal y
paymen s o he se ice composi ion ðps;mÞ ha a p o ide
pays o he cus ome o missing he SLA we e modeled in
he cus ome ’s calcula ion wi h he same a iable, bu wi h
he ma hema ical sign e e sed, i.e. as nega i e a iable
cos s ðps;mÞ:
Financial impac calcula ions pe mon h:
o alcos scus ;m¼ ixcos scus ;mþ a cos scus ;m;
o alcos sp o ;m¼ ixcos sp o ;mþ a cos sp o ;m;
8m2 1,2,3;...;Mg
The inancial impac pe mon h is he sum o he
mon hly a iable and ixed cos s. The implici connec ion
be ween he agg ega ed a ailabili y o he se ice compo-
si ion and he inancial impac on he cus ome and he
p o ide se es as a basis o decision suppo ega ding
he sha ing o isk. Using he i ed o al cos g aph o he
cus ome and he p o ide , in e es ing conside a ions a ise
conce ning he decision-making and nego ia ion op ions o
bo h pa ies, as he ollowing case s udy shows.
4 De elopmen and Demons a ion
HySOR was implemen ed using he in eg a ed de elop-
men en i onmen RS udio/2023.03.1, wi h which an
R-based applica ion was de eloped ha builds on he
lib a y ‘‘Shiny’’. The modeling o he Poisson a i al o
se ice ou ages was ealized by he R unc ion pois, he
logno mal dis ibu ion o he down ime du a ion by lno m.
The inpu ields a e dis ibu ed o e h ee a eas o he
use in e ace, as shown in Fig. 2. In he i s , he ‘‘Se ice
Composi ion’’ panel (see Fig. 2[A]), he cha ac e is ics o
he se ice composi ion a e ep esen ed by business p o-
cess a iables and SLA pa ame e s a he op le el. In he
second a ea, he ‘‘ Type [Componen ’’ panel, he
cha ac e is ics o each se ice componen in ol ed a e
pa ame e ized, as shown in Fig. 2[B] using he example o
a p i a e cloud componen . The ou ‘‘Add Componen ’’
bu ons (see Fig. 2[C]) a e used o add an addi ional
componen o he composi ion as a hi d inpu a ea. The
addi ion o a ypical public cloud componen was a
equi emen om an ea ly es phase o he p o o ype and
was implemen ed o s eng hen he demons abili y o he
model. SAP, Mic oso , and Sales o ce we e chosen as
sui able examples o he expec ed case s udies due o he
123
502 M. Sei e , S. Kuehnel: HySOR: A Simula ion Model o he Sha ing o Risk..., Bus In Sys Eng 67(4):495–510 (2025)
desc ibe he planned una ailabili y o cloud se ices. In
addi ion, he expe in e iews e ealed – con a y o
ema k 2 – ha business p ocess ou age cos s in p ac ice
a y g ea ly depending on he ime ame. This is imme-
dia ely appa en a he weekends, du ing which many
companies’ business p ocesses do no incu any ou age
cos s. A mapping o di e en ou age cos unc ions in
ela ion o ime ames can lead o u he in e es ing
esul s, especially when conside ing scheduled main e-
nance pe iods.
A second in e es ing di ec ion o u he esea ch is he
de elopmen o sui able es ima ion me hods and, in con-
nec ion wi h his, he examina ion o his o ical da a o
modeling ne wo k, in e ace, o o he unknown isks which
could in luence he a ailabili y o se ice composi ions
(e.g., depending on he numbe o se ice p o ide s and
se ice componen s in ol ed). This could be ac o ed in o
he simula ion model wi hou he expe s ha ing o use
hei in ui ion o make es ima es.
8 Conclusion
This pape p esen s a isk calcula ion and simula ion model
ha was de eloped conside ing a i ac s o he exis ing
knowledge base in a igo ous DSR p ocess i e a ion.
The co e elemen s o se ice componen s, se ice
composi ions, associa ed SLAs/OLAs as well as hei
owne s (in di e en oles), suppo ed business p ocesses,
and a ious isks we e de i ed om he li e a u e, de ined,
and concep ualized. The a eas o se ice commi men and
se ice c edi , which we e iden i ied as inancially ele an
in connec ion wi h he SLAs, complemen his o answe
he i s esea ch ques ion (RQ1). Using he se ice com-
ponen and se ice composi ion cha ac e is ics o HySOR,
as well as he co esponding unc ions, a ailabili y simu-
la ion and agg ega ion, business p ocess ou age calcula ion
and agg ega ion, and penal y calcula ion and agg ega ion,
he impac o IT down imes o componen s was imple-
men ed in he simula ion o cus ome and p o ide cos s o
answe he second esea ch ques ion (RQ2). The hi d
esea ch ques ion (RQ3) is answe ed on he basis o he
esul s o simula ing and calcula ing he agg ega ed a ail-
abili y o he se ice composi ion and he esul ing inan-
cial impac o he cus ome and he p o ide in he bes ,
mean, and wo s case. Finally, he isualiza ion o he
inancial impac depending on he op-le el a ailabili y in
he o m o ba cha s, supplemen ed by linea end lines,
was conside ed help ul by he leading expe s su eyed.
The esul s o his wo k illus a e he ele ance o he
opic and con i m he di e en pe cep ions o anspa ency
in hyb id cloud a chi ec u es. HySOR p o ides use ul
impulses o he nego ia ion o hyb id cloud se ice com-
posi ions in he con ex o he sha ing o isk.
The con ibu ion o his pape is wo old. On he
empi ical side, we add ess a p ac ically impo an opic
wi h an e ec i e a i ac implemen ed. The e idence om
he e alua ion o leading expe s p o ides p omising
di ec ions o u he esea ch. On he concep ual side, he
desc ip ion o he calcula ion and simula ion elemen s o
he HySOR model/a i ac can se e as an e alua ed
ounda ion o o he esea che s.
Funding Open Access unding enabled and o ganized by P ojek
DEAL.
Re e ences
Ahle M, S a
¨ e KF (2016) Re ining Rai a – aspi a ion adap a ion
wi hin he zone o possible ag. Ge Econ Re 17:298–315.
h ps://doi.o g/10.1111/gee .12096
Al-Ghuwai i A-R, Khala MN, Al-Yasen L, Salah Z, Alsa han A,
Baa ah AH (2016) A dynamic model o au oma ic upda ing
cloud compu ing SLA (DSLA). In: P oceedings o he in e na-
ional con e ence on in e ne o hings and cloud compu ing.
ACM, New Yo k, pp 1–7. h ps://doi.o g/10.1145/2896387.
2896442
Aljoumah E, Al-Mousawi F, Ahmad I, Al-Shamm i M, Al-Jady Z
(2015) SLA in cloud compu ing a chi ec u es: a comp ehensi e
s udy. In J G id Dis ib Compu 8:7–32. h ps://doi.o g/10.
14257/ijgdc.2015.8.5.02
Base SA (2012) Cloud SLAs: p esen and u u e. SIGOPS Ope Sys
Re 46:57–66. h ps://doi.o g/10.1145/2331576.2331586
Benlian A, Kou a is M, Hess T (2011) Se ice quali y in so wa e-as-
a-se ice: de eloping he SaaS-Qual measu e and examining i s
ole in usage con inuance. J Manag In Sys 28:85–126. h ps://
doi.o g/10.2753/MIS0742-1222280303
B ei e G, Naik VK (2013) A amewo k o con olling and
managing hyb id cloud se ice in eg a ion. In: 2013 IEEE
in e na ional con e ence on cloud enginee ing, pp 217–224.
h ps://doi.o g/10.1109/IC2E.2013.48
Comuzzi M, Ko sokalis C, Ra h elde C, Theilmann W, Winkle U,
Zacco G (2009) A amewo k o mul i-le el SLA managemen .
In: Dan A, e al (eds): Se ice-o ien ed compu ing. Sp inge ,
Heidelbe g, pp 187–196. h ps://doi.o g/10.1007/978-3-642-
16132-2_18
Comuzzi M, Jacobs G, G e en P (2013) Clea ing he sky: unde -
s anding SLA elemen s in cloud compu ing. BETA publica ie:
wo king pape s ol. 412, Eindho en Uni e si y o Technology,
Eindho en, pp 1–25. h ps:// esea ch. ue.nl/en/publica ions/clea
ing- he-sky-unde s anding-sla-elemen s-in-cloud-compu ing.
Accessed 20 Jul 2024
Faniyi F, Bahsoon R, Theodo opoulos G (2012) A dynamic da a-
d i en simula ion app oach o p e en ing se ice le el ag ee-
men iola ions in cloud ede a ion. P ocedia Compu Sci
9:1167–1176. h ps://doi.o g/10.1016/j.p ocs.2012.04.126
F anke U, Johnson P, Ko
¨nig J (2013) An a chi ec u e amewo k o
en e p ise IT se ice a ailabili y analysis. So w Sys Model
13:1417–1445. h ps://doi.o g/10.1007/s10270-012-0307-3
Goo J, Kisho e R, Rao HR, Nam K (2009) The ole o se ice le el
ag eemen s in ela ional managemen o in o ma ion echnology
123
M. Sei e , S. Kuehnel: HySOR: A Simula ion Model o he Sha ing o Risk..., Bus In Sys Eng 67(4):495–510 (2025) 509

ou sou cing: an empi ical s udy. MIS Q 33:119–145. h ps://doi.
o g/10.2307/20650281
Gue
´ ou T, Medjiah S, Da Cos a G, Mon eil T (2014) Quali y o
se ice modeling o g een scheduling in clouds. Sus ain Compu
In o m Sys 4:225–240. h ps://doi.o g/10.1016/j.suscom.2014.
08.006
Gulia P, Sood S (2013) Compa a i e analysis o p esen day clouds
using se ice le el ag eemen s. In J Compu Appl 71:1–8.
h ps://doi.o g/10.5120/12335-8603
Hussain O, Dong H, Singh J (2010) Seman ic simila i y model o isk
assessmen in o ming cloud compu ing SLAs. In: Mee sman R
e al (eds) On he mo e o meaning ul in e ne sys ems, OTM
2010. Sp inge , Heidelbe g, pp 843–860. h ps://doi.o g/10.
1007/978-3-642-16949-6_12
Ii a i J, Hansen M, Haj-Bolou i A (2021) A p oposal o minimum
eusabili y e alua ion o design p inciples. Eu J In Sys
30:286–303. h ps://doi.o g/10.1080/0960085X.2020.1793697
Ga ne Inc. (2024) Ga ne Fo ecas s Wo ldwide Public Cloud End-
Use Spending o Su pass $675 Billion in 2024. In: Ga ne
News oom, In o ma ion Technology, P ess Release, h ps://
www.ga ne .com/en/news oom/p ess- eleases/2024-05-20-ga
ne - o ecas s-wo ldwide-public-cloud-end-use -spending- o-su
pass-675-billion-in-2024. Accessed July 20, 2024.
Jiang M, By ne J, Molka K, A ms ong D, Djemame K, Ki kham T
(2013) Cos and isk awa e suppo o cloud SLAs. 2184–5042.
h ps://doi.o g/10.5220/0004377302070212
Johnson P, Ullbe g J, Buschle M, F anke U, Shahzad K (2014) An
a chi ec u e modeling amewo k o p obabilis ic p edic ion. In
Sys E-Bus Manag 12:595–622. h ps://doi.o g/10.1007/s10257-
014-0241-8
K i ikos K, Plexousakis D, Plebani P (2016) Seman ic SLAs o
se ices wi h Q-SLA. P ocedia Compu Sci 97:24–33. h ps://
doi.o g/10.1016/j.p ocs.2016.08.277
Labidi T, M ibaa A, B ab a H (2016) CSLAOn o: a comp ehensi e
on ological SLA model in cloud compu ing. J Da a Seman
5:179–193. h ps://doi.o g/10.1007/s13740-016-0070-7
Ma eo-Fo nes J, Solsona-Tehas F, Vilaplana-Mayo al J, Teixido-
To elles I, Rius-To en o J (2019) CART, a decision SLA
model o SaaS p o ide s o keep QoS ega ding a ailabili y and
pe o mance. IEEE Access 7:38195–38204. h ps://doi.o g/10.
1109/ACCESS.2019.2905870
Maz ekaj A, Shabani I, Sejdiu B (2016) P icing schemes in cloud
compu ing: an o e iew. In e na ional Jou nal Ad ance Com-
pu e Science Applica ions 7. h ps://doi.o g/10.14569/IJACSA.
2016.070211
Pan W, Mi chell G (2015) So wa e as a se ice (SaaS) quali y
managemen and se ice le el ag eemen . In u u e 26:225–234.
h ps://doi.o g/10.17234/INFUTURE.2015.26
Paque e S, Jaege PT, Wilson SC (2010) Iden i ying he secu i y isks
associa ed wi h go e nmen al use o cloud compu ing. Go In Q
27:245–253. h ps://doi.o g/10.1016/j.giq.2010.01.002
Pe e s K, Tuunanen T, Ro henbe ge MA, Cha e jee S (2007) A
design science esea ch me hodology o in o ma ion sys ems
esea ch. J Manag In Sys 24:45–77. h ps://doi.o g/10.2753/
MIS0742-1222240302
Rai a H (1982) The A and Science o Nego ia ion: How o esol e
con lic s and ge he bes ou o ba gaining. Ha a d Uni . P ess,
Camb idge
Rockmann R, Weege A, Gewald H (2014) Iden i ying o ganiza ional
capabili ies o he en e p ise-wide usage o cloud compu ing. In:
PACIS 2014 P oceedings. h p://aisel.aisne .o g/pacis2014/355
Sei e M, Kuehnel S, Sackmann S (2023) Hyb id clouds a ising om
so wa e as a se ice adop ion: challenges, solu ions, and u u e
esea ch di ec ions. ACM Compu . Su . Vol. 55, No. 11. A icle
228:1–35. h ps://doi.o g/10.1145/3570156
Sei e M (2021) Analysis o public cloud se ice le el ag eemen s -
an e alua ion o leading so wa e as a se ice p o ide . In:
Ku
¨hnel S, Sackmann S, T ang S (eds): P oceedings o he Fi s
In e na ional Wo kshop on Cu en Compliance Issues in
In o ma ion Sys ems Resea ch (CIISR’21), Co-loca ed wi h he
16 h In e na ional Con e ence on Wi scha sin o ma ik (WI’21),
Online (ini ially loca ed in Duisbu g-Essen, Ge many), Ma ch
9 h, 2021. CEUR Wo kshop P oceedings 2966, pp 22–35.
h ps://ceu -ws.o g/Vol-2966/pape 2.pd
Sei e M, Kuehnel S (2021) ‘‘HySLAC’’ - a concep ual model o
se ice le el ag eemen compliance in hyb id cloud a chi ec-
u es. In: Reussne RH, Koziolek A, Hein ich R (eds):
INFORMATIK 2020, Lec u e No es in In o ma ics (LNI),
Gesellscha u
¨ In o ma ik, Bonn 2021 205–218. h ps://doi.
o g/10.18420/in 2020_19
Suakan o S, Supangka SH, Suha di, Sa agih R (2012) Pe o mance
measu emen o cloud compu ing se ices. In e na ional Jou nal
Cloud Compu ing Se ices A chi ec u e 2 2 9 20. h ps://doi.o g/
10.5121/ijccsa.2012.2202
Sun W, Zhang X, Guo CJ, Sun P, Su H (2008) So wa e as a se ice:
con igu a ion and cus omiza ion pe spec i es. In: 2008 IEEE
cong ess on se ices pa II, pp 18–25. h ps://doi.o g/10.1109/
SERVICES-2.2008.29
Theilmann W, Happe J, Ko sokalis C, Edmonds A, Kea ney K,
Lambea J (2010) A e e ence a chi ec u e o mul i-le el SLA
managemen . J In e ne Eng 4:289–298. h ps://doi.o g/10.
21256/zhaw-1757
Venable J, P ies-Heje J, Baske ille R (2016) FEDS: a amewo k o
e alua ion in design science esea ch. Eu J In Sys 25:77–89.
h ps://doi.o g/10.1057/ejis.2014.36
Wang SS, F anke U (2020) En e p ise IT se ice down ime cos and
isk ans e in a supply chain. Ope Manag Res 13:94–108.
h ps://doi.o g/10.1007/s12063-020-00148-x
Yuan X, Li Y, Jia T, Liu T, Wu Z (2015) An analysis on a ailabili y
commi men and penal y in cloud SLA. In: 2015 IEEE 39 h
annual compu e so wa e and applica ions con e ence,
pp 914–919. h ps://doi.o g/10.1109/COMPSAC.2015.39
Zhang L-J, Zhou Q (2009) CCOA: cloud compu ing open a chi ec-
u e. In: 2009 IEEE in e na ional con e ence on web se ices,
pp 607–616. h ps://doi.o g/10.1109/ICWS.2009.144
Sp inge Na u e o i s licenso (e.g. a socie y o o he pa ne ) holds
exclusi e igh s o his a icle unde a publishing ag eemen wi h he
au ho (s) o o he igh sholde (s); au ho sel -a chi ing o he
accep ed manusc ip e sion o his a icle is solely go e ned by he
e ms o such publishing ag eemen and applicable law.
123
510 M. Sei e , S. Kuehnel: HySOR: A Simula ion Model o he Sha ing o Risk..., Bus In Sys Eng 67(4):495–510 (2025)