Redefining the Cloud based on Beneficial Service Characteristics
A New Cloud Taxonomy Leads to Economically Reasonable Semi-cloudification
Bastian Kemmler and Dieter Kranzlm
¨
uller
Leibniz Supercomputing Centre (LRZ), Bavarian Academy of Sciences and Humanities,
Boltzmannstr. 1, 85748 Garching n. Munich, Germany
Keywords:
Cloud, Semi-cloud, Service, Cloud Service, Semi-cloud Service, Service Management.
Abstract:
Cloud services promise benefits for customers and providers such as scalability, elasticity and reduced invest-
ment costs. Unfortunately, many of the promised benefits are not fulfilled by today’s cloud offerings and not
every service can be cloudified, e.g. if the service’s intrinsic structure contains unavoidable time-consuming
or manual tasks. A new cloud definition, based on a survey and comparison of existing cloud definitions, but
derived from beneficial cloud characteristics, leads to a service-oriented understanding of clouds and provides
an extension to the usual cloud service types. The characteristics of the given cloud definition uncover the
so-called ”MOUSETRAPS ” of cloud services. The term ”semi-cloudification” for the transformation of ser-
vices towards a cloudified state presents a solid foundation for further discussions on the topic and enables
the improvement of non-cloudifiable services by semi-cloudification. Even services which partly consist of
unavoidable time-consuming or manual tasks qualify for semi-cloudification.
1 INTRODUCTION
Cloud computing influences the way today’s IT re-
lated businesses work. The percentage of compa-
nies which utilize clouds continually increases (Wall-
raf and Pols, 2014, p. 15). According to Gartner the
use of cloud services is growing faster than the over-
all enterprise IT market. (...) Cloud computing is set
to have a considerable impact on business in the fu-
ture. (...) Three key factors (...) will significantly im-
pact enterprise cloud use in the near to midterm fu-
ture (Rivera and Meulen, 2013):
1. Cloud services will be primarily used as a solu-
tion for specific problems with limited scope.
2. Cloud services will have an increased business
impact, while the use of cloud services moves
up the service chain from infrastructure towards
business process services.
3. Cloud solutions will lead to a more diverse so-
lution portfolio, widely varying in timelines, re-
source requirements, benefit profiles, business
criticality and complexity.
Although the migration challenges for the cloud
service customer are debated extensively (Khajeh-
Hosseini et al., 2010a, Khajeh-Hosseini et al., 2011,
Khajeh-Hosseini et al., 2010b, Kaisler and Money,
2011,Ward et al., 2010, Andrikopoulos et al., 2013,
Paulus and Riemann, 2013), there is little discussion
on the service provider challenges for transforming
legacy services into cloud services.
Besides other aspects, the three key factors in-
dicate four major requirements, which will be high-
lighted in this work:
1. Cloud taxonomies need to consider the promised
benefits of the cloud.
2. Services besides the general well-known clas-
sification Infrastructure-as-a-Service (IaaS),
Platform-as-a-Services(PaaS) and Software-as-
a-Service(SaaS) need to be included in a more
generalistic approach to the cloud. Especially, so-
lutions with limited scope, IT-related services of
the service chain towards business processes and
highly complex services should be considered.
3. Cloud taxonomies should enable and improve
discussions about the emerging service provider
challenges, which accompany the transformation
of a legacy service towards a cloud service.
4. Cloud services in the long run have the potential to
substitute the vast majority of legacy IT services
(M
¨
unzl et al., 2009, p. 22). Therefore, the cloud
service definition needs to be based on a solid def-
inition for IT services.
135
Kemmler B. and Kranzlmüller D..
Redefining the Cloud based on Beneficial Service Characteristics - A New Cloud Taxonomy Leads to Economically Reasonable Semi-cloudification.
DOI: 10.5220/0005446401350144
In Proceedings of the 5th International Conference on Cloud Computing and Services Science (CLOSER-2015), pages 135-144
ISBN: 978-989-758-104-5
Copyright
c
 2015 SCITEPRESS (Science and Technology Publications, Lda.)

Based on this list, we extract the following:
• Which benefits are expected from a cloud service?
• Which of the existing cloud definitions support
these benefits?
Unfortunately, the promised benefits of cloud ser-
vices are often not fulfilled. A potential reason for the
lack of fulfilled benefits could stem from a major de-
sign fault in existing cloud definitions. Therefore, we
need to ask the question:
• Which cloud service characteristics need to be ful-
filled for a service to become a cloud service, thus
fulfilling the promised benefits?
In order to evaluate these questions, we also need
to investigate the following:
• Which services are cloudifiable, which are not?
• Which cloud service characteristics can be ful-
filled for non-cloudifiable services?
Following this introduction, Section 2 discusses
cloud benefits, that are promised in relevant literature.
Of these benefits several are still unfulfilled in exist-
ing cloud implementations. Section 3 compares ex-
isting cloud definitions and presents their inherent in-
ability to fulfill the promised benefits, while Section 4
proposes a new cloud taxonomy. By focusing on the
limits of cloudification Section 5 introduces the term
semi-cloud as a midway point for cloudification and
a possible solution for non-cloudifiable IT services.
To give an example for the use of the benefit oriented
approach Section 6 presents a scenario and a possible
solution. Section 7 concludes the paper and provides
an outlook on future work.
2 CLOUD SERVICE BENEFITS
2.1 Promised Benefits
Several research papers and practitioner reports
promise cloud service benefits, which are listed in
this subsection. Particularly we included the research
done by Khajeh-Hosseini et. al., who identified the
benefits and risks of using public IaaS clouds [...] by
reviewing over 50 academic papers and industry re-
ports (Khajeh-Hosseini et al., 2011). The (promised)
benefits of cloud computing depend on the enter-
prises perspective and can be distinguished into cloud
service customer benefits and cloud service provider
benefits.
Promised Cloud Service Customer Benefits
B1 DISTRIBUTED ACCESS
1
: Enhanced mobile and
geographically distributed access enables cus-
tomers (and consumers) to access cloud services
almost anywhere. Cloud services provide (end-
user-)device mobility, which improves collabo-
ration, and geo-distribution of similar cloud ser-
vices, thereby increasing existing backup facili-
ties, availability and continuity capabilities.
B2 INCREASED SCALABILITY
1
: Customers ben-
efit from scale-up/-down capabilities of cloud
services by adjusting their cloud service usage
according to the existing workload. Under-/ and
overprovisioning can be reduced.
B3 ELASTICITY
2
: Elasticity is the degree to which
a system is able to adapt to workload changes
by provisioning and deprovisioning resources in
an autonomic manner, such that at each point
in time the available resources match the cur-
rent demand as closely as possible (Nikolas Ro-
man Herbst et al., 2013).
B4 IMPROVED ORGANIZATIONAL FLEXIBILITY
AND AGILITY
3
: Cloud services can be flexi-
bly adjusted to business needs. A higher IT ab-
straction level leads to (IT-related) business de-
cisions which focus on core business activities
and not on IT details. Organizational changes
are less restricted by the local IT environment.
B5 TERM TRANSFORMATION OF INVEST-
MENT
4
: The use of services transforms capital
expenditures (CAPEX) for investments like
hardware or software into operational expen-
ditures (OPEX)(Etro, 2011, p. 12). Thereby
longer, unmodifiable investment periods, which
can last up to several years, are transformed
into several much shorter terms, which can
be adjusted on short notice. The frequently
occurring charge-per-use or pay-as-you-go
accounting-models for cloud services greatly
encourages this transformation.
B6 REDUCED TIME TO MARKET
1
: Cloud services
enable the faster creation of new or redesigned
services or products. The timeframe between
idea and service/product is reduced.
1
(Khajeh-Hosseini et al., 2011; Carroll et al., 2011;
Phaphoom et al., 2012; Wallraf and Pols, 2014)
2
(Khajeh-Hosseini et al., 2011)
3
(Khajeh-Hosseini et al., 2011; Carroll et al., 2011;
Wallraf and Pols, 2014)
4
(Etro, 2011; Khajeh-Hosseini et al., 2011; Wallraf and
Pols, 2014)
CLOSER2015-5thInternationalConferenceonCloudComputingandServicesScience
136

B7 ENTRY BARRIER REDUCTION
5
: Initial re-
source demand for new or redesigned services
can be decreased. The term transformation of
investment reduces the constraints on entry and
promotes business creation (Etro, 2011, p. 12).
B8 LOWER ADMINISTRATION COSTS
1
: The ex-
penditure for administration, maintenance and
general operation is reduced for the customer.
B9 IMPROVED AVAILABILITY AND PERFOR-
MANCE
1
: Both can be improved due to the huge
resources of cloud service providers. Quality
assurance and control is centralized.
B10 ENHANCED DATA SECURITY
6
: Professional
centralized security management, which is
transferred to the service provider, is often bet-
ter than the security management of customers.
B11 DISASTER RECOVERY
7
: Centralized disaster
recovery and geo-distribution of resources is
beneficial for cloud customers.
Promised Cloud Service Provider Benefits
B12 IMPROVED MANAGEMENT EFFI-
CIENCY
8
: Economies of scale lead to more
efficient management and provider automation.
B13 ENERGY EFFICIENCY
9
: The aggregation of
system components should also have a positive
effect on the level of power consumed on hard-
ware and software(Rajan and Jairath, 2011).
10
B14 ECONOMIES OF SCALE
2
:Because computing,
storage and other service needs are aggregated
at provider level, cloud services make better use
of economies of scale.
B15 IMPROVED CAPACITY MANAGE-
MENT
11
: Providers benefit by an improved
capacity management through using otherwise
idle system components.
2.2 Unfulfilled Benefits
A study based on the contributions of users, develop-
ers, consultants, entrepreneurs and researchers to the
Cloud Computing Google Group indicates, that the
5
(Etro, 2011)
6
(Carroll et al., 2011; Phaphoom et al., 2012; Wallraf
and Pols, 2014)
7
(Khajeh-Hosseini et al., 2011; Phaphoom et al., 2012)
8
(Byung et al., 2013; Khajeh-Hosseini et al., 2011; Car-
roll et al., 2011)
9
(Rajan and Jairath, 2011; Carroll et al., 2011)
10
Although this needs further proof.
11
(Khajeh-Hosseini et al., 2011; Carroll et al., 2011)
promised benefits B8, B9 and B10 have not been fully
achieved, while the results on the benefits B2 and B5
demonstrate their fulfillment. (Phaphoom et al., 2012)
Moreover, KPMG and BITKOM state, that the ini-
tial goals B1, B2 and B4 have been achieved by the
majority of enterprise cloud users. The promised ben-
efits B6, B8, B9, B10 and reduced IT investment costs
still lack fulfillment.(Wallraf and Pols, 2014, p. 26)
Obviously the benefits B13, B14 and B15 could
not be achieved, because if otherwise, IT investment
costs could have been reduced. Additionally, the fail-
ure to achieve benefit B6 also indicates the failure in
achieving benefit B7.
Although benefit B5 can be considered fulfilled,
because of the general construction of IT services,
there is no real proof that management efficiency
could be improved (B12) by providing cloud services.
For the benefit of existing cloud service implementa-
tion it can be assumed that B11 is fulfilled, due to
the existing experience with legacy IT services. Un-
fortunately, there is no proof for the (un)fulfillment
of benefit B3. Even though the argumentation in
this section mainly follows a paper which summarizes
contributions of the Google Cloud Computing Group
(Phaphoom et al., 2012) and a report which describes
the situation in Germany (Wallraf and Pols, 2014)
it can be assumed that many of the presented unful-
filled benefits are also unfulfilled worldwide. Other-
wise, globalisation and the ubiquitous nature of the
cloud would ruin any cloud service provider in the US
(Google Cloud Computing Group) and in Germany
shortly. An overview of the fulfilled and unfulfilled
benefits is given in Table 1.
Nevertheless, eight out of ten enterprise users tes-
Table 1: Fulfilled and unfulfilled cloud benefits.
Fulfilled B1 Distributed Access
Cloud B2 Increased Scalability
Benefits B4 Improved Organizational
Flexibility and Agility
B5 Term Transformation of
Investment
B11 Disaster Recovery
Unfulfilled B6 Reduced Time to Market
Cloud B7 Entry Barrier Reduction
Benefits B8 Lower Administration Costs
B9 Improved Availability and
Performance
B10 Enhanced Data Security
B13 Energy Efficiency
B14 Economies of Scale
B15 Improved Capacity
Management
Undecided B3 Elasticity
B12 Improved Management
Efficiency
RedefiningtheCloudbasedonBeneficialServiceCharacteristics-ANewCloudTaxonomyLeadstoEconomically
ReasonableSemi-cloudification
137

tify that the utilization of a cloud is beneficial for
them.(Wallraf and Pols, 2014, p. 24)
3 EXISTING CLOUD
TAXONOMY
Almost any big player in the business of software and
computer services
12
and additionally several com-
panies in the business of technology hardware and
equipment
13
use the term ”cloud ” as an integral part
of their advertising language. Despite widely used
for several IT environments, the terms ”cloud ” and
”cloud computing” are still not used with a commonly
accepted understanding.
14
3.1 Characteristics of Existing Cloud
Definitions
The ongoing research on cloud computing resides on
several incomplete and in parts contradictory defini-
tions, like the definitions of NIST(Mell and Grance,
2011), Forrester(Staten, 2008), Gartner(Cearley,
2010), European Commission(EC)(European Com-
mission, 2010, p. 8), BITKOM(M
¨
unzl et al., 2009,
p. 16) and BSI(Federal Office for Information Secu-
rity, 2011, p. 13). These definitions describe the cloud
with specific characteristics like access, service type,
customer type, business benefits and provisioning fea-
tures, which are listed in detail in the following.
Cloud Access Characteristics
15
AC1 ubiquitous
AC2 on-demand
AC3 self-service
AC4 over a private
network
AC5 over a public
network
AC6 over a broad
network only
AC7 API
12
like Microsoft, Google, IBM, Oracle and SAP, listed
in the Financial Times global top 500(Financial Times,
2014),which provides an annual snapshot of the world’s
largest companies (Dullforce, 2014)
13
like Apple, Intel, Cisco and EMC, listed in the Finan-
cial Times global top 500(Financial Times, 2014)
14
The iCloud connects you and your Apple device[,]
safely store[s] all your presentations, spreadsheets, PDFs,
images, and other kinds of documents (Apple, 2014), finds
your Apple device and protects it against theft; IBM is
building a smarter planet with IBM SmartCloud cloud com-
puting (IBM, 2014) and Microsoft brings higher education
into the cloud [Microsoft Deutschland GmbH, 2012-10-11].
15
Convenient access is also a possible characteristic
which is required by the NIST definition. But since con-
venience is only measurable subjectively and depends very
much on the users point of view, it does not qualify as a
cloud characteristic.
Cloud Service Type Characteristics
SC1 IT service
SC2 SaaS, PaaS, IaaS
SC3 XaaS
16
SC4 configurable service
parameters
Cloud Customer Characteristics
CC1 public
CC2 private
CC3 community
CC4 hybrid
Cloud Business Characteristics
BC1 measured (priced) service
BC2 specified level of quality
BC3 tailored to a market need
BC4 CAPEX ⇒ OPEX
Cloud Provisioning Characteristics
PC1 elastic
PC2 scalable
PC3 rapid
PC4 real-time
PC5 pooling
PC6 no management
effort or provider in-
teraction
17
3.2 Comparison of Existing Cloud
Definitions
By using the extracted definition characteristics to
take a closer look at these cloud definitions, the dif-
ferences become visible (see Table 2).
Especially if the introduced expected benefits (see
Section 2) of the cloud are also considered, the
deficits of existing cloud definitions are obvious. A
mapping between the expected benefits of the cloud
and their counterparts (characteristics, see Table 3)
applied to the definitions of the NIST, Forrester,
Gartner, EC, BITKOM and BSI clearly shows those
deficits (see Table 4). Therefore, a clean definition set
for the terms cloud, cloud computing, cloud service,
cloud service provider and cloudification is essential
for further scientific work regarding the limits of the
cloud and non-cloudifiable services.
4 CLOUD TAXONOMY
PROPOSAL
As discussed above, cloud taxonomy (see Figure 1)
should be build on top of the definition for the generic
16
besides S/P/IaaS
17
Some definitions use terms like with minimal manage-
ment effort or minimal provider interaction. From a theoret-
ical approach this is synonymous with no management ef-
fort or no provider interaction, because minimal is nothing
if there are no constraints and with unspecified constraints
minimal is meaningless.
CLOSER2015-5thInternationalConferenceonCloudComputingandServicesScience
138

Table 2: Comparison of cloud definitions.
Dimension Characteristic NIST Forrester Gartner EC BITKOM BSI
Access
AC1 4
AC2 4
AC4 4 6 4
AC5 4 4 4 4 4
AC6 4
AC3 4 4
AC7 4
Service
Type
SC1 4 4 4 4 4
SC2 4 (4) 4 4
SC3 6 6 4
SC4 4 4
Customer
CC1 4 4 4
CC2 4 6
CC3 4 6
CC4 4 6
Business
BC1 4 4 4 4
BC2 4
BC3 4 4
BC4 4
Provisioning
PC1 4 4 4 (4) 4
PC2 4
PC3 4
PC6 4 6
PC4 4
PC5 4
4: cloud definition contains characteristic
6: cloud definition contradicts characteristic
Figure 1: Cloud taxonomy.
scope of this paper
Service
IT Service (XaaS)
Cloud
Service
Cloud
Service
Provider
Cloud
Cloud Customer
Cloud Provider
Cloud Consumer
Cloud Computing
Figure 1: Cloud taxonomy.
service.
18
In this paper only the terms ”service”,
”IT service”, ”cloud service” and ”cloud ” will be de-
fined. All other terms listed in Figure 1 can be derived
18
Byung et al.(Byung et al., 2013) also state that the man-
agement of some cloud services is more efficient than the
management of standard IT services.
Table 3: Benefits targeted by cloud definition characteris-
tics.
Benefit
Essential Nice-to-have
Definition Definition
Characteristics Characteristics
Customer
B1 AC1 AC5
B2 AC7, PC2, PC3 AC2
B3 PC1 AC2, AC7, SC4, PC2
B4 AC2, PC2 CC3, CC4
B5 BC4 BC1
B6 AC2, AC3, AC7 BC2, BC3
B7 AC3, BC4 BC2, AC7, BC3
B8 AC7, BC3
B9 AC1, AC2, BC2, AC4, AC5, AC6
PC1, PC2
B10 BC2 CC2
B11 AC1
Provider
B12 BC3, PC6
B13 PC5
B14 AC3, AC7, BC1, PC5
PC2, PC6
B15 BC2, PC2 PC5
Table 4: Benefits targeted by essential cloud definition char-
acteristics.
Benefit NIST Forrester Gartner EC BITKOM BSI
Customer
B1 4
B2 m m
B3 4 4 4 (4) 4
B4 m m
B5 4
B6 m m m
B7 m m m
B8 m 4
B9 m m m m m
B10 4
B11 4
Provider
B12 m m m
B13 4
B14 m m m m m m
B15 m m m m
4: targeted by at all essential cloud def. characteristics
m: targeted by at least one essential cloud def. char.
easily and are omitted due to space limitations.
4.1 IT Service
According to ITIL (Iqbal and Nieves, 2007, page 16)
a service is defined as follows
19
:
Definition 1 (Service). A service is a means of de-
livering value to customers by facilitating outcomes
19
See also ISO/IEC 20000-1:2011; a service is a means
of delivering value to customers by facilitating outcomes
customers want to achieve (ISO/IEC, 2011, 3.26).
RedefiningtheCloudbasedonBeneficialServiceCharacteristics-ANewCloudTaxonomyLeadstoEconomically
ReasonableSemi-cloudification
139

customers want to achieve without the ownership of
specific costs and risks.
Definition 2 (IT Service). An IT service is a service,
which consists at least partly of IT-related aspects.
Generally spoken, an IT service can be created by
a specific business unit utilizing assets of the IT orga-
nization to add value to the customers business. These
assets A can be segmented into sets like management
A
m
, organization A
o
, processes A
pr
, knowledge A
kn
,
people A
ppl
, information A
in f
, applications A
app
, in-
frastructure A
i
and financial capital A
f c
(see (Iqbal
and Nieves, 2007, page 39)) and their subsets like
A
app
i
, i ∈ N, which hold the following:
Σ
type
:= {m, o, pr, kn, ppl, in f , app, i, f c} (1)
A =
[
k∈Σ
type
A
k
(2)
∀k ∈ Σ
type
∃n ∈ N : A
k
=
n
[
i=1
A
k
i
(3)
Definition 3 (XaaS). The value of an IT service is
generated by providing X-as-a-Service (XaaS), which
(as of today) can be segmented into the following ser-
vice types:
• Management-as-a-Service (MaaS), like interim
management
• Organization-aaS (OaaS), like franchising
• Processes-aaS (PRaaS), like franchising
• Knowledge-aaS (KNaaS), which splits into
Education-aaS (EDUaaS) and consulting
• People-aaS (PPLaaS), like temporary work
• Information-aaS (INFaaS), like Reuters or
Bloomberg
• Application-aaS (APPaaS), which splits into
Software-aaS (SaaS) and Platform-aaS (PaaS)
• Infrastructure-aaS (IaaS), which splits into
Hardware-aaS (HaaS), virtual-Infrastructure-aaS
(vIaaS) and Desktop-aaS (DaaS)
• and Capital-aaS (FCaaS), like a call loan or a
credit facility
Mixtures of these service types are also XaaS.
In accordance to the definition of XaaS (Defini-
tion 3) an IT service can be characterized by service
type and the needed input assets iA
k
and output assets
oA
k
, which can be split up into provider assets PiA
k
and PoA
k
and customer assets CiA
k
and CoA
k
.
Furthermore, an IT service can also be created by
a specific business unit, additionally utilizing avail-
able XaaS of the IT organization to add value to
the customers’ business. As already discussed, those
XaaS S can also be segmented into sets like manage-
ment S
m
, organization S
o
, processes S
pr
, knowledge
S
kn
, people S
ppl
, information S
in f
, applications S
app
,
infrastructure S
i
and financial capital S
f c
and their
subsets like S
app
i
, i ∈ N
Altogether, these assets and services are the build-
ing blocks of a generic IT service. They serve as input
parameters to the service function
f
S
: CiA ×CiS ×PiA ×PiS → CoA ×CoS × PoA ×PoS
(4)
which creates the service according to a service level
agreement (SLA) or an operational level agreement
(OLA). These agreements specify the the level, scope
and quality of the service in detail.
Definition 4 (Service Level Agreement (SLA)). SLAs
are the documents agreed with the customers that
specify the level, scope and quality of service to be
provided (Lloyd et al., 2007, page 24)).
Definition 5 (Operational Level Agreement (OLA)).
OLA are any underpinning agreements necessary to
deliver the quality of service agreed within the SLA
(Lloyd et al., 2007, page 24)).
Additionally the SLA/OLA may contain specifi-
cations about occurring customer obligations. The
provided service itself is an entity of the set CoS.
Therefore the SLA can be defined as the function
f
SLA
: CiA
|
SLA
× CiS
|
SLA
→ CoS
|
SLA
(5)
Figure 2: Generic service model.
PiA PoA
PiS PoS
CiA CoA
CiS CoS
f
S
+
+
+
SLA
SLA
SLA
Figure 2: Generic service model.
Finally, the service may contain a configurable
service parameter vector p, which leads to
f
S
p
: CiA ×CiS × PiA ×PiS
→ CoA ×CoS × PoA ×PoS (6)
and f
SLA
p
: CiA
|
SLA
p
× CiS
|
SLA
p
→ CoS
|
SLA
p
(7)
4.2 Cloud Service
Based on the discussions above a new cloud definition
can be derived, which, on the one hand gives, a solid
foundation to fulfill customers/providers expectancy
CLOSER2015-5thInternationalConferenceonCloudComputingandServicesScience
140

Table 5: Benefits targeted by cloud definition characteristics
without service benefits and characteristics.
Benefit
Essential Nice-to-have
Definition Definition
Characteristics Characteristics
Customer
B1 AC1 AC5
B2 AC7, PC2, PC3 AC2
B3 PC1 AC2, AC7, PC2
B4 AC2, PC2 CC3, CC4
B6 AC2, AC3, AC7 BC2
B7 AC3 BC2, AC7
B8 AC7
B9 AC1, AC2, BC2, AC4, AC5, AC6
PC1, PC2
B10 BC2 CC2
B11 AC1
Provider
B12 PC6
B13 PC5
B14 AC3, AC7, BC1, PC5
PC2, PC6
B15 BC2, PC2 PC5
and, on the other hand, considers the already listed
commonly accepted cloud characteristics. The most
important cloud service characteristics can be sum-
marized by the acronym ”MOUSETRAPS”.
Definition 6 (Cloud Service). A cloud service is a
M easured, O n-demand, U biquitous, S calable
and E lastic IT service T ailored to a market need
with a specified level of quality (SLA/OLA). It is provi-
sioned, deprovisioned or reconfigured R apidly with-
out provider interaction, by using an A PI, P ooling
mechanisms and a S elf-service. The service should
be accessible over a broad network.
By utilizing the previously given definition for an
IT service (see Definition 1 and 2) benefit B5 and the
cloud characteristics SC1, SC2, SC3, SC4, BC3 and
BC4 already have been fully included. Thus Table 3
can be significantly reduced as shown in Table 5. Ad-
ditionally, the inclusion of SC3 already disqualifies
the definitions of NIST and BITKOM (see Table 2).
Particularly, the mandatory SLA/OLA of cloud
services enable customers to conduct an appropriate
risk-analysis for the cloud operation and migration of
legacy customer systems. This enables customers to
migrate even mission-critical systems into the cloud.
Definition 7 (Cloud). A cloud is a multitenant elec-
tronic marketplace, accessible by a broad network,
with strictly defined interfaces for customers and
provider(s), where cloud services can be traded
20
.
Access to the cloud can be restricted to a private,
community, public or hybrid audience.
20
General linguistic usage: Services do not ”leave” the
cloud. They ”live” in the cloud.
Table 6: Comparison of cloud definitions without service
characteristics.
Dimension Charact. NIST Forrester Gartner EC BITKOM BSI Def. 7 Def. 10
Access
AC1 4 4 4
AC2 4 4 4
AC4 4 6 4 4
AC5 4 4 4 4 4 4
AC6 4 4
AC3 4 4 4 4
AC7 4 4 4
Customer
CC1 4 4 4 4 4
CC2 4 4 4
CC3 4 4 4
CC4 4 4 4
Business
BC1 4 4 4 4 4 4
BC2 4 4 4
Provisioning
PC1 4 4 4 4 4 4
PC2 4 4 4
PC3 4 4
PC6 4 6 4
PC4 4
PC5 4 4 4
Table 7: Benefits targeted by all essential and nice-to-have
cloud definition characteristics.
Benefit NIST Forrester Gartner EC BITKOM BSI Def. 7 Def. 10
Customer
B1 4 4
B2 4
B3 4
B4 4 4 4
B6 4 4
B7 4 4
B8 4 4 4
B9 4
B10 4 4
B11 4 4
Provider
B12 4
B13 4 4 4
B14 4
B15 4 4
This newly defined cloud definition fulfills all
needed characteristics to achieve the promised ben-
efits (see Tables 6 and 7), and therefore gives a bet-
ter understanding of the cloud. Although the char-
acteristic PC4 could be dropped, because according
to Table 6 it is not beneficial to any of the promised
benefits, it is included in Definition 6. Obviously, the
dependence between AC2 and PC4 cannot be fully
denied.
Consequently ”cloudification” can be defined as:
RedefiningtheCloudbasedonBeneficialServiceCharacteristics-ANewCloudTaxonomyLeadstoEconomically
ReasonableSemi-cloudification
141

Definition 8 (Cloudification). Cloudification is the
transformation by which an IT service becomes a
cloud service which is provisioned using a cloud.
5 LIMITS OF CLOUDIFICATION
Unfortunately, not every IT service can be fully cloud-
ified. IT services like the hardware maintenance of
desktop computers or laptops, full service copier lease
with ink refill service, education, consulting or other
services which at least partly consist of physical work
in not properly standardized environments, do not
qualify for cloudification, because of their unavoid-
able manual (and time consuming) work load. Ad-
ditionally, full service cloudification would be very
costly and contradicts benefits which target the over-
all reduction of costs. By dropping the cloud charac-
teristics AC4, AC5, AC6, PC3 and PC6 an econom-
ically more reasonable definition can be found. The
new terms, ”semi-cloud service”, ”semi-cloud ” and
”semi-cloudification” should be used as terms to dis-
cuss those ”nearly” cloudified services. The most im-
portant semi-cloud service characteristics can again
be summarized by the acronym ”MOUSETrAPS ”.
Definition 9 (Semi-Cloud Service). A semi-cloud
service is a M easured, O n-demand, U
biquitous
and S calable IT service T ailored to a market
need with a specified level of quality (SLA/OLA). It
is provisioned, deprovisioned or reconfigured with
E conomically reasonable provider interaction, by
using an A PI, P ooling mechanisms and a S elf-
service.
Definition 10 (Semi-Cloud). A semi-cloud is a multi-
tenant electronic marketplace with strictly defined in-
terfaces for customers and provider(s), where semi-
cloud services can be traded. Access to the semi-
cloud can be restricted to a private, community, pub-
lic or hybrid audience.
Definition 11 (Semi-Cloudification). Semi-
cloudification is the transformation by which an
IT service becomes a semi-cloud service which is
provisioned using a semi-cloud.
As can be seen in Table 7, only the benefits B1,
B2, B9, B12 and B14 are not fully targeted by the
semi-cloud as defined in Definition 10. Further in-
vestigation shows, that these benefits are only slightly
missed:
B1. All essential definition characteristics fulfilled
B2. Only PC3 unfulfilled
B9. All essential definition characteristics fulfilled
B12. Benefit B12 is based on characteristic PC6.
21
B14. Only PC6 unfulfilled
Altogether, the shortcomings of the proposed defini-
tion for the semi-cloud focus around the missing char-
acteristics PC3 and PC6, which were omitted on pur-
pose to encourage the development of designs which
balance investment in the semi-cloud systems with
economical benefits. Additionally, non-cloudified
services which at least partly consist of unavoidable
manual tasks, can be semi-cloudified. Therefore the
term semi-cloud can be considered a substantial ex-
tension to the cloud taxonomy.
6 APPLICATION SCENARIO
To show some real-life aspects of the given approach
towards (semi-)cloudification, we introduce the fol-
lowing scenario and observe possible solutions. By
the nature of a scenario, the given problems and the
solution can not touch the full extend of this work,
but give an impression on the change of perspective
set by the new cloud taxonomy.
6.1 Scenario
Consider a midsize IT company which started busi-
ness activities in 1965 by selling copying machines
and fulfilling corresponding maintenance contracts.
In 1985 the CEO decided to include office print-
ers for personal computers into the product range.
Later on the company began selling PCs, servers and
network equipment always together with their corre-
sponding maintenance contracts until in 2005 the new
CEO introduced a new service oriented strategy. To-
day the enterprise has around 200 employees, 5 VIP-
Customers (three of those since 1965) and around 400
SME customers total. The company mainly provides
four services: printing (on-site and off-site), printing
and mailing, desktop management and virtual root
servers, which are located at a company owned fa-
cility. Last year an ISO/IEC 20000 certification was
achieved. Most of the employees are booked to capac-
ity and there is little time left for innovative projects.
In the last two years the CEO heard about the cloud
and its benefits but deferred a project to enable cloud
computing for his customers, because of the resources
involved with the organizational transformation to-
wards ISO/IEC 20000. Although the budget is tight
the CEO expects his new CIO to ”cloudify” his com-
pany. He privately assumes that the company will es-
pecially benefit from the resulting management effi-
21
which is dropped deliberately
CLOSER2015-5thInternationalConferenceonCloudComputingandServicesScience
142

ciency, energy efficiency and the improved capacity
management of the cloud. By this, he believes, his
company will reach the turning point towards making
profit again. With great commitment the CIO initiates
the cloud project.
6.2 Problems
In this setting, besides many other, several problems
arise which were discussed in this paper.
1) Each of the company’s employees has a slightly
different understanding of the cloud:
• The customer relationship manager looks for-
ward to the new cloud service, which will be
added to the company’s service lines and is
highly demanded by the company’s customers.
• The infrastructure department head expects
benefits by accommodating desperately needed
resource demands with an external cloud
provider.
• The hardware maintenance group thinks, that
the cloud is a topic for the software guys.
• The virtualisation specialists state, that cloud
computing is already their daily business.
• Another approach is given by the head of re-
search and development, who proposes to in-
stall a software called FreeStack.
2) The CEO wants to fully cloudify the company.
But the CIO thinks, that many services of the
company can not be cloudified. Installing the
proposed IaaS software FreeStack for virtual ma-
chines provisioning, which also provides inter-
faces to external cloud providers, could be a so-
lution for the VM service. But who is going to
manage the emerging new cloud service, if most
of the employees are booked to capacity?
3) The CIO’s project budget is limited. Therefore
she would like to focus on the most important ben-
efits and implement an economically reasonable
cloud solution. But which are these and which
one of the several available cloud definitions sup-
port these benefits?
4) Especially the VIP customers might stick to their
legacy services and refuse the new cloud services.
6.3 Solution with the Given Approach
By introducing the employees to the benefit oriented
cloud taxonomy of Section 4, a commonly accepted
understanding of the cloud can be reached (Prob-
lem 1). Especially because of the benefit oriented ap-
proach, the characteristics in the given definitions can
be directly connected to the CEOs desires.
Additionally, a structural analysis and proper de-
composition of the company’s services according
to Subsection 4.1 reveals that the services printing
(on-site and off-site), printing and mailing, desk-
top management can not be cloudified, but semi-
cloudified because of their unavoidable manual work-
load. Moreover, existing services should be trans-
formed into semi-cloud services with at least equal
service features one-by-one to free employees from
working on legacy services (Problem 2).
An Interview with the CEO based on the given list
of benefits in Section 2 reveals, that he is especially
interested in gaining the benefits B12, B13 and B15.
Semi-Cloudification of the company’s services with
a special focus on characteristics PC5, BC2 and PC2
will specifically address benefits B13 and B15 (see
Table 5), while characteristic PC6 and therefore bene-
fit B12 contradicts the budget constraints(Problem 3).
A moderate, cost-conscious automation of service
production steps seems to be advisable.
By the transformation of existing services VIP
customers can be migrated without loss of the existing
service features (Problem 4).
7 SUMMARY AND OUTLOOK
As indicated by our analysis, most of the promised
cloud benefits are not addressed by existing cloud def-
initions. This presents a possible root cause for the
unfulfillment of most of those benefits. Based on an
analysis of the generic IT service and by identifying
beneficial cloud service characteristics, the so-called
MOUSETRAPS, a new cloud definition has been de-
rived, which focusses intensely on the desired ben-
efits of the cloud. The terms ”semi-cloud”, ”semi-
cloud service” and ”semi-cloudification” present a
solid base terminology for further discussions of the
topic. On the one hand, semi-cloudification can be
understood as a midway point towards full cloudifi-
cation and, on the other hand, many of the existing
non-cloudifiable services can be semi-cloudified.
Further studies are needed to analyze the trans-
formation process between the steps non-cloudified,
semi-cloudified and cloudified to improve the cloud-
ification of legacy services. Generally cloud re-
search should shift towards a more benefit-oriented
approach, especially, when it comes to subjects re-
garding the service provider and their motives for im-
plementing cloud solutions. Service providers have
to realize that the impact of (semi-)cloudification will
not only enhance their own services, but also services
of their competitors, resulting in a tighter, but better
focused service portfolio for each service provider.
RedefiningtheCloudbasedonBeneficialServiceCharacteristics-ANewCloudTaxonomyLeadstoEconomically
ReasonableSemi-cloudification
143

REFERENCES
Andrikopoulos, V., Binz, T., Leymann, F., and Strauch, S.
(2013). How to Adapt Applications for the Cloud En-
vironment. Computing, 95(6):493–535.
Apple (2014). iCloud. http://www.apple.com/icloud/.
04.12.2014.
Byung, C. T., Urgaonkar, B., and Sivasubramaniam, A.
(2013). Cloudy with a Chance of Cost Savings. Par-
allel and Distributed Systems, IEEE Transactions on,
24(6):1223–1233.
Carroll, M., van der Merwe, A., and Kotze, P. (2011). Se-
cure Cloud Computing: Benefits, Risks and Controls.
In 2011 Information Security for South Africa (ISSA),
pages 1–9.
Cearley, D. W. (2010). Cloud Computing: Key Initiative
Overview. http://www.gartner.com/it/initiatives/pdf/
KeyInitiativeOverview CloudComputing.pdf.
21.11.2014.
Dullforce, A.-B. (2014). FT 500 2014 Introduction
and Methodology. http://www.ft.com/intl/cms/
s/0/6fdb9d70-fdf5-11e3-bd0e-00144feab7de.html.
08.07.2014.
Etro, F. (2011). The Economics of Cloud Computing. IUP
Journal of Managerial Economics, 9(2):7–22.
European Commission (25.01.2010). The Future Of Cloud
Computing. http://cordis.europa.eu/fp7/ict/ssai/docs/
cloud-report-final.pdf. 02.01.2013.
Federal Office for Information Security (2011). Security
Recommendations for Cloud Computing Providers:
Minimum information security requirements.
https://www.bsi.bund.de/SharedDocs/Downloads/
EN/BSI/Publications/Minimum information/
Security Recommendations Cloud Computing
Providers.html. 2015-01-05.
Financial Times (2014). Global 500 Companies
Ranked by Sector. http://im.ft-static.com/content/
images/70710ff2-fded-11e3-bd0e-00144feab7de.xls.
08.07.2014.
IBM (2014). http://www-07.ibm.com/hk/cloud/cloud-
computing/. 04.12.2014.
Iqbal, M. and Nieves, M. (2007). Service Strategy. ITIL.
TSO (The Stationery Office), London, 2 edition.
ISO/IEC (2011). ISO/IEC 20000-1:2011 - Information
Technology - Service Management, Part 1: Service
Management system requirements.
Kaisler, S. and Money, W. (2011). Service Migration in
a Cloud Architecture. In System Sciences (HICSS),
2011 44th Hawaii International Conference on, pages
1–10.
Khajeh-Hosseini, A., Greenwood, D., and Sommerville, I.
(2010a). Cloud Migration: A Case Study of Migrating
an Enterprise IT System to IaaS. In 2010 IEEE Inter-
national Conference on Cloud Computing (CLOUD),
pages 450–457.
Khajeh-Hosseini, A., Sommerville, I., Bogaerts, J., and
Teregowda, P. (2011). Decision Support Tools for
Cloud Migration in the Enterprise. In 2011 IEEE
4th International Conference on Cloud Computing
(CLOUD), pages 541–548.
Khajeh-Hosseini, A., Sommerville, I., and Sriram, I.
(2010b). Research Challenges for Enterprise Cloud
Computing. CoRR, abs/1001.3257.
Lloyd, V., Rudd, C., and Taylor, S. (2007). ITIL - Service
Design: [IT service management practices; ITIL v3
core publications]. TSO, London.
Mell, P. and Grance, T. (2011). The NIST Definition of
Cloud Computing: Recommendations of the National
Institute of Standards and Technology: Special Pub-
lication 800-145. Computer Security Division, Infor-
mation Technology Laboratory, National Institute of
Standards and Technology, Gaithersburg, MD.
Microsoft Deutschland GmbH (2012-10-11). Hochschulen
in der Wolke: Immer mehr Bildungsein-
richtungen setzen auf Cloud Computing.
http://www.microsoft.com/germany/newsroom/
pressemitteilung.mspx?id=533628. 05.12.2014.
M
¨
unzl, G., Przywara, B., Reti, M., Sch
¨
afer, J., Son-
dermann, K., Weber, M., and Wilker, A. (2009).
Cloud Computing: Evolution in der Technik,
Revolution im Business: BITKOM-Leitfaden.
http://www.bitkom.org/files/documents/BITKOM-
Leitfaden-CloudComputing web.pdf. 02.01.2013.
Nikolas Roman Herbst, Samuel Kounev, and Ralf Reussner
(2013). Elasticity in Cloud Computing: What It Is,
and What It Is Not. In Proceedings of the 10th Inter-
national Conference on Autonomic Computing (ICAC
13), pages 23–27, San Jose, CA. USENIX.
Paulus, S. and Riemann, U. (2013). An Approach for a
Business-Driven Cloud Compliance Analysis Cover-
ing Public Sector Process Improvement requirements.
CoRR, abs/1310.2832.
Phaphoom, N., Oza, N., Wang, X., and Abrahamsson, P.
(2012). Does Cloud Computing Deliver the Promised
Benefits for IT Industry? In M
¨
annist
¨
o, T., editor, the
WICSA/ECSA 2012 Companion Volume, page 45.
Rajan, S. and Jairath, A. (2011). Cloud Computing: The
Fifth Generation of Computing. In Communication
Systems and Network Technologies (CSNT), 2011 In-
ternational Conference on, pages 665–667.
Rivera, J. and Meulen, R. v. d. (2013). Gartner Says
the Road to Increased Enterprise Cloud Usage Will
Largely Run Through Tactical Business Solutions
Addressing Specific Issues. http://www.gartner.com/
newsroom/id/2581315. 05.12.2014.
Staten, J. (2008). Is Cloud Computing Ready
For The Enterprise? A Client Choice Re-
port: Not Yet, But This Disruptive Innovation
Is Maturing Fast. https://www.forrester.com/
Is+Cloud+Computing+Ready+For+The+Enterprise/
fulltext/-/E-RES44229? 05.12.2014.
Wallraf, B. and Pols, A. (2014). Cloud-Monitor 2014:
Cloud-Computing in Deutschland – Status quo
und Perspektiven. http://www.kpmg.com/DE/de/
Documents/cloudmonitor-2014-kpmg.pdf.
05.12.2014.
Ward, C., Aravamudan, N., Bhattacharya, K., Cheng, K.,
Filepp, R., Kearney, R., Peterson, B., Shwartz, L.,
and Young, C. (2010). Workload Migration into
Clouds Challenges, Experiences, Opportunities. In
2010 IEEE International Conference on Cloud Com-
puting (CLOUD), pages 164–171.
CLOSER2015-5thInternationalConferenceonCloudComputingandServicesScience
144