a taxonomy for describing bi cloud servicesd.researchbib.com/f/1nagv2bghhptez.pdfa taxonomy for...

A Taxonomy for Describing BI Cloud Services

Oliver Norkus Jurgen SauerUniversity of Oldenburg, Department for Computing Science

Escherweg 2, 26121 Oldenuburg, Germanyoliver.norkus@uni-oldenburg, [email protected]

ABSTRACT

Cloud computing is still a driver of many innova-tions. The cloud comes with features nowadays re-quired in the area of Business Intelligence (BI). BICloud offerings are proactive, integrated tools forreporting that are mobile accessible, flexible andscalable. BI in the Cloud is becoming increasinglypopular, but a comprehensive reference architecturefor describtion, comparison and evaluation of BICloud services does not exist yet.A universal model to describe Cloud services ingeneral exists. This, however, is too generic to beapplicable to BI Cloud services. Even several con-ceptual frameworks to describe and characterize ex-plicit BI Cloud services exist. But, as they are de-veloped by different groups and organizations, theydiffer in their intention, level of description, type offormulation, and in terms of which key issues of BICloud service evaluation are addressed.To help overcome the prevalent skepticism of en-terprises regarding BI Cloud services, we surveyeddifferent frameworks and approaches for describeand compare BI Cloud services. Based on this sur-vey and the universal model for Cloud service of-ferings we propose a consolidated taxonomy as anextension of the existing general taxonomy with theaim to uniformly describe, compare and evaluate BICloud services.

KEYWORDS

Cloud service, Business Intelligence in the Cloud,BI as a Service, taxonomy, standardization.

1 INTRODUCTION

The term Business Intelligence means the inte-grated IT-based management support. BI aimsto improve the process of decision making.The focus is to provide relevant information tothe business users (i.e., decision makers) at theright time at any place [1].

In particular, strategic, intercompany and ex-plorative missions of BI have become the focusof discussion. These involve greater attentionto poly-structured data sources, a stronger inte-gration of more sophisticated methods of anal-ysis (advanced, visual and predictive analyt-ics), as well as a stronger orientation towardsagility (e.g., via self-service BI). Together withthe parallel conducted process integration aswell as an addition of new internal and externaluser groups there is a need to revise establishedarchitecture.Especially by regarding that previously used BIsystems are mostly rigid, complex and costly[2], cloud computing (CC) is more often usedin the field of BI to improve flexibility, scala-bility and agility [3, 4].CC is defined as ”a model for enabling ubiq-uitous, convenient, on-demand network accessto a shared pool of configurable computing re-sources (e. g., networks, servers, storage, ap-plications, and services) that can be rapidlyprovisioned and released with minimal man-agement effort or service provider interaction”[5].The combination of BI and CC is used more in-tensively. BI in the Cloud yet offers new tech-nology combinations to provide individual anddifferentiated configurable, scalable and flexi-ble analytical services. BI Cloud services havefive main properties [4, 6]:

• The infrastructure is installed and main-tained by the service provider. A BI Cloudservice therefore makes available the nec-essary infrastructure for BI capabilities.So the business users do not have to man-age it on their own.

• The infrastructure is scalable, so it can beexpanded dynamically when required. Inthis way, it can be ensured that the de-

Proceedings of the International Conference on Semantic Web Business and Innovation (SWBI2015), Sierre, Switzerland, 2015

ISBN: 978-1-941968-19-2 ©2015 SDIWC 1

sired analysis can be performed more per-formant, even with multiple, simultaneousaccess.

• The access to the BI Cloud service is donevia a browser, and is generally supportedby all common mobile and stationary ter-minals by using standards (e.g., protocols,interfaces). By this, the business users canalways access to the BI Cloud service.

• The use is flexible and dynamic, and canbe performed at any time without the ser-vice customer to interact with the serviceprovider to agree a time of use.

• The users pay for their measured con-sumption on the base of the utilization(pay-per-use). This can be based on thenumber of visits or based on the usagetime. So, cost transparency and pre-dictability is enabled.

On the base of the novelty of BI Cloud ser-vices, there are not much field reports yet. Theabsence of standards and transparency espe-cially for description and comparison of BICloud services promotes skepticism and in-comprehension.There are several approaches to describe, char-acterize and compare BI Cloud service offer-ings. To help overcome the prevalent skepti-cism of enterprises regarding BI Cloud offer-ings, we surveyed different frameworks for BICloud service offerings. However, as they aredefined and elaborated by different groups andorganizations, they differ in their intention andscope, in their level of abstraction and the typeof formalization.In addition, a model for universally describ-ing Cloud services exists [7, 8]. But, this is sogeneric and universal, that it is not adequate ap-plicable in the field of BI. Based on the surveyand also taking into account the existing modelfor describing Cloud services, we conceptual-ize a consolidated taxonomy to uniformly de-scribe, characterize and compare BI Cloud ser-vice offerings, regardless whether an organiza-tions own offerings or the offerings of exter-nal providers are evaluated. In addition, we

present the Business Intelligence Cloud Ser-vice Navigator (BI-CSN) as a suitable visual-ization technique to describe visually and com-pare BI Cloud service offerings.The contribution of this paper is a survey ofdifferent frameworks to describe BI Cloud ser-vice offerings. With the aim to abolish theskepticism of enterprises regarding BI Cloudservices we identified their key factors and de-signing a defined and elaborated a taxonomy todescribe and compare BI Cloud services.The rest of this paper is structured as fol-lows: We present an overview on the surveyedBI Cloud service frameworks and models insection 2, within a description of the surveymethod. Based on the survey we proposed theconsolidated feature models to provide a com-mon foundation for BI Cloud service descrip-tion and comparison for all roles, even for cus-tomers with a lack of basic understanding ofthe combination of BI and CC. We discuss thefeature model and the individual factors of BICloud service offerings in section 3. We alsopresent the BI-CSN as an appropriate visual-ization technique for the visual description andcomparison of BI Cloud service offerings andrequests in section 4. Within the application,we also have an effect on the evaluation. Fi-nally, in section 5, we conclude with a shortsummary and remarks on future work.

2 SUBJECT OF RESEARCH

This section contains an overview on the sur-veyed BI Cloud service frameworks, and a de-scription of the surveyed method. We thensummarized the differences between the in-dividual frameworks and reference models,which represents the motivation for the con-solidated taxonomy for BI Cloud services wepropose in section 3.As of this writing, several models and concep-tual frameworks to describe and characterizeBI Cloud service offerings exist. To help clearaway the skepticism of enterprises regardingBI as a service (BIaaS), we conducted a surveyof these different frameworks. The surveyedapproaches are:


ISBN: 978-1-941968-19-2 ©2015 SDIWC 2

• Government and associations: BITKOM[9, 10, 11], Gartner [12], BARC [13, 14].

• Industry: SAP [15, 16, 17], Microsoft [18,19], QlikTech [20], Oracle [21, 22], IBM[23, 24], MicroStrategy [25], TATA [26].

• Academia: Adabi [27], Ouf and Nasr[28], Chadha and Iyer [29], Tamer etal. [30], Grivas et al. [31], Demirkanand Delen [32], Baars and Kemper [2],Gurhar and Rathore [33], Gash et al. [34],Thompson and van der Walt [35], Hasel-mann and Vossen [36], Seufert and Bern-hardt [37], Bernhardt and Balluch [38],Weinhardt et al. [39], Torkashvan andHaghighi [40, 41], Leinmeister et al. [42],Schirm et al. [43], Chang [44] and Erethand Dahl [45].

2.1 Survey Completeness

As shown, the surveyed models represent ap-proaches from different types of organizations;models that were developed by academia, byindustry and by governmental structures orother associations.Regarding the industry models, we consideredthe models provided by SAP, Microsoft, Qlik-Tech, IBM and Oracle. These are representa-tive IT companies and show significant activi-ties in the area of BI in the Cloud [3, 38].Regarding academia, we surveyed frameworksthat represent individual models for BI Cloudand approaches that discuss individual topicssuch as technical or organizational capabili-ties. Finally, regarding models by governmen-tal structures of other associations, we consid-ered studies from Gartner, BITKOM, the Ger-man Federal Association for Information Tech-nology, Telecommunication and New Mediaand BARC the German Business ApplicationResearch Center.

2.2 Survey Method

The survey method is based on the methodol-ogy used by defining and elaborating the ap-proach for description and comparison general

Cloud services. Thus, the survey was con-ducted in several steps. We began by conduct-ing a breathfirst search in order to identify rel-evant works; with the results described abovein subsection 2.1. Next we described the sur-veyed works using feature models [46, 47, 48].Feature models provide three benefits for thesurvey [8]:

• Accessibility: Like mind maps, featuremodels are simple and can be clearlyarranged, represented, and compared toeach other.

• Formality: Feature models are formalmodels that can be validated. Furhter-more, they allow modeling exclusions,multiple selections, and cardinality.

• Structure recognition: Feature models donot impose an organizational structure be-forehand and even allow to model incom-plete concepts. This is in contrast to ta-bles, for example, wherein columns androws typically have to be named.

Figure 1 shows an example of a feature model.The root element is called Feature Model. Theelement has two sub features, one is manda-tory and the other is optional. Again, both fea-tures have subfeatures that illustrate inclusion(and) and exclusion (or). Feature models al-low an arbitrary model depth, but for the sakeof understandability and readability, the mod-eling depths of different branches should be asuniform as possible [46, 8].

FeatureModel

MandatoryFeatureXor1SubFeature

Xor2SubFeature

OptionalFeatureAlt1SubFeature

Alt2SubFeature

Legend:

MandatoryOptionalOrAlternativeAbstractConcrete

Figure 1. An example of a feature model based on [8].

We then compared the surveyed BI Cloud ser-vice frameworks by analyzing the feature mod-els, see subsection 2.3. Finally we synthesized


ISBN: 978-1-941968-19-2 ©2015 SDIWC 3

a consolidated feature model based on the anal-ysis, to uniformly describe and evaluate Cloudservice offerings, see section 3.

2.3 Survey Results

We compared the surveyed BI Cloud servicemodels by analyzing their feature models. Forthe sake of brevity, this section represents ashort summary of the main results. We firstsummarize the models and then present ananalysis. The analysis represents the motiva-tion for the consolidated BI Cloud service fea-ture model that we propose in section 3.

2.3.1 Models from Government and Asso-ciations

The models created by governmental structuresor other associations are intended to be vendor-neutral models. These models describe the de-gree and intensity of utilisation and capabili-ties of BIaaS. Thereby a distinction is madebetween the focus on the deployment model(private, public, and hybrid) and for the archi-tectural level (infrastructure, platform and soft-ware).

2.3.2 Industry models

The models created by the industry representstheir specific BI Cloud service offering. Thereare descriptions in the fields of architecture,examples of use cases, visualization and con-figuration functions, service level agreements,data security, deployment model and account-ing. For example, SAP [15, 16, 17] settle ac-count on the base of usage hours and Microsoft[18, 19] according the resource using.

2.3.3 Models from academia

Adabi [27] and Hasselmann and Vossen [36]deals with data management in the cloud. Oufand Nars [28] as well as Baars and Kemper [2]present a layer architecture for BI Cloud ser-vices. Thompson and van der Walt [35] re-port from a survey primary in the sector of ap-plication of BI Cloud services and offer firststatements to consolidation. Demirkan and De-len [32] provide a service-oriented architecture

for putting BI in the cloud. Chadha and Iyer[29] show an overview of an architecture foreasy implementation first steps to gain quickwins. Tamer et al. [30], Gurhar and Rathore[33] and Gash et al. [34] also show first archi-tectural models, but consider merely aspects.Chadha and Iyer [29] go further and offer ameta model. Seufert and Bernhardt [37] of-fer a submodel for the aspect of service type.Bernhardt and Balluch [38] use this submodeland so classify some BI Cloud service offer-ings. Weinhardt et al. [39] present a frame-work for business models in Clouds. Torkash-van and Haghighi [40] present a framework forcloud service level agreement management anda general service oriented framework for cloudcomputing [41]. Leinmeister et al. [42] intes-tigate the stakeholder in the business perspec-tive of CC. Schirm et al. [43] are dealing withsucess factors of BI Cloud solutions. Chang[44] analyse BI in the cloud in the applica-tion in the finance industry. Ereth and Dahl[45] present a first approach for a service-basedevaluation concept.

2.3.4 Analysis

There are several models, offerings and ap-proaches of BIaaS. Differences in the offeringsexists in the field of architecture model, visual-ization functions and userfriendliness as wellas the billing model. From an architecturalpoint of view, these offerings must be viewedas a black box, because the available informa-tion does not go far enough beyond marketing-driven product descriptions. The models fromgovernmental structures and other associationsprovide primary market observations regardingthe use of various offerings. In the academiaarea, there are already first conceptual con-siderations and approaches, which mainly fo-cus on aspects and partly only include generalstatements.Table 1 gives an overview of our findings. Asthis table show, the surveyed references differin their level of abstraction, the type of formal-ization (based on [8]):

• Level of abstraction: Does the respectivereference describe one or more concrete


ISBN: 978-1-941968-19-2 ©2015 SDIWC 4

BI service offerings (instance), does itprovide a generic description of BI Cloud(meta model), or does it discuss certainaspects (aspects)?

• Type of formalization: How is the respec-tive framework described? Typically ex-amples are taxonomies, block diagrams,ontologies, lists and plain text.

Table I illustrates that the surveyed referencesare mostly informal, which hinders compari-son and leaves space for unwanted interpreta-tion. The lack of a formal yet comprehensibleBI Cloud reference model that represents thecurated aggregation of the works of many rep-resents the motivation for the BI Cloud servicefeature model we propose in the next section.

3 FACTORS OF BI CLOUD SERVICEOFFERINGS

As a recent work in the research project FutureBusiness Clouds (FBC) [7], different frame-works and approaches for Cloud service offer-ings description in general were surveyed andbased on this a consolidated reference architec-ture was proposed [8]. Thus, there is a compre-hensive and complete [7, 8], accepted and used[49, 50, 51] approach for universal descrip-tion of Cloud service offerings. As noted evenby publishing this, the approach needs furtherextension and specification [8]. The existingframework with factors of Cloud service of-ferings in general represent the base enhancedhere. In the following, the identified BI spe-cific factors are presented as an extension ofthe existing generic framework. For the sakeof brevity, in this paper, only the added fea-tures are displayed. For a first overview here isreferred to Fig. 7, wherein the added featuresare marked by coloring them green.Based on the analysis in subsubsection 2.3 andas an extension of the existing approach, wedesigned a consolidated feature model to uni-formly describe and evaluate BI Cloud serviceofferings. In this section we describe the fac-tors of BI Cloud service offerings and in fol-lowing section 4 we present the BI-CSN as a

suitable visualization technique to visually de-scribe and compare BI Cloud services. Withthis we will have an effect on the applicationand evaluation.The feature model, as shown in Fig. 2, con-sists of nine main features for service descrip-tion: the description of the service type, de-ployment, pricing, roles, service integration as-pects, sourcing options, service level agree-ments (SLAs), organization specific aspects,and Cloud service characteristics. Additionshave resulted in the following sub-models: ser-vice type, service level agreement, organiza-tional specific aspects and pricing model.

BIaaS_Feature_Model

ServiceType

Deployment

Pricing

Role

Integration

Sourcing

SLA

Characteristics

Organization

Legend:

MandatoryAbstractConcrete

Figure 2. Main features of the taxonomy based on [8]

3.1 Service Type

The generic framework considers the servicetypes Software as a Service (SaaS), Platformas a Service (PaaS), Infrastructure as a Service(IaaS) and Mashup as a Service with the sub-feature Business Process as a Service (BPaaS).To describe BI Cloud service offerings this fea-ture model was extended by the subtypes Visu-alization as a Service (VaaS), Model as a Ser-vice (MaaS), Datawarehouse System (DWS)and Data as a Service (DaaS), see Fig. 3.Firstly, the service types VaaS and MaaS areordered beneath SaaS because they both de-scribe Software that follows the SaaS ap-proach. VaaS represents all display and presen-tation aspects whereover analytical functions


ISBN: 978-1-941968-19-2 ©2015 SDIWC 5

can be used. This includes for example Dash-boards, Scorecards and Reports. MaaS coversthe aspects of flexibility in the form of self-service BI (e.g. modifying reports, redefin-ing scores) and offers several analytical func-tions by providing On-Line Analytical Process(OLAP) and Data Mining functionalities.Secondly, the PaaS feature was extended by thesubfeature Datawarehouse System as a plat-form service. DWS covers the managementof the Cloud BI system. This includes e. g.development of analytical functions, the con-figuration of the extract, transform and load(ETL) process, the control of data flows, theuser management and the creation of multidi-mensional data models for the use in reports.Thirdly, the service type DaaS is ordered be-neath IaaS, cause the focus is on storage. IaaSis devided in the infrastructure features storage(DaaS) and computing. DaaS includes the pro-vision of infrastructure in the form of storageexplicitly for Data Warehouses, data marts, BIrepositories and staging areas for the ETL pro-cess.

ServiceType

SaaS

VaaS

Dashboard

Scorecard

Report

MaaS

OLAP

DataMining

SelfServiceBI

PaaS DWSDevelopment

Configuration

IaaS

DaaS

StagingArea

Repository

DataMarts

DataWareHouse

Computing

MashupAsAService BPaaS

Legend:

MandatoryOptionalAbstractConcrete

Figure 3. Service Type

3.2 Service Level Agreement

Service Level Agreements (SLAs) are of cen-tral importance for the successful cooperationbetween the participants in a BI Cloud envi-ronment. Figure 4 shows the extended SLAfeature model. The model considers functionaland non-functional service qualities, agreedgovernance aspects such as key performance

indicators (KPIs), as well as agreed changeprocesses. Both, the functional and non-functional got extended by service qualitiesthat are important to describe BI Cloud serviceofferings. The functional qualities for examplenow contain the feature Availability, Backupsand Data Recovery. The non-functional on theother hand got extended by the qualities DataProtection, Access Authorization and Laws.

SLA

Qualities

Functional

DataIO

Usability

ResponseTime

ErrorBehavior

Availability

Backups

DataRecovery

NonFunctional

Locality

Performance

UsagePrice

Integrability

Safety

DataProtection

AccessAuthorization

Laws

Processes

Measures

Modalities

Termination

Incentives

Penalties

KPIs

Legend:


Figure 4. Service Level Agreements

3.3 Organizational specific aspects

To establish trust among the participants in aBI Cloud environment, organizations must beable to state their reputation and to assess theircapabilities. The model considers the organi-zations reputation as well as the organizationscapabilities, see Fig. 5. Regarding the reputa-tion the model consider certificates, referenceprojects, KPIs, benchmarks and reports, andfurther information such as websites. On theother hand, an organizations capabilities de-pend on the given resources, employee knowl-edge, technical skills and business skill. Fur-ther the capabilities or trust were extended bythe feature Support. This feature describes if


ISBN: 978-1-941968-19-2 ©2015 SDIWC 6

the cloud service provider offers support ser-vices inclusive or if the customer has to payfor it separately.

Company

Trust

BusinessSkills

TechnicalSkills

KnowledgeBase

Resources

Support

Reputation

Information

ExpertAssesments

Benchmarks

KPIs

References

Certificates

Legend:


Figure 5. Organizational specific aspects

3.4 Pricing model

As Fig. 6 shows, the pricing model is dis-tinguished between Free, Pay per Usage andAbonnement. Free means free use, for ex-ample, to attract new customers. Pay per Us-age covers payment based on measured usage.Abonnement stands for periodic payments fora recurring use.Filed under Pay per Usage are pay-per-time-of-use, pay-per-usage and pay-per-unit. Withinpay-per-time-of-use is mentioned a paymentfor certain period of use. Pay-per-usage coversthe payment for the number of considerationrespective views of a reports (e. g. pay-per-report). A resource-based billing, e.g. paymentfor calculation power units or storage require-ments, so used hardware resources, is summa-rized with pay-per-unit.

Pricing

Free

PayperUse

PayperUsage

PayperUnit

PayperTimeofUse

Abonnement

Legend:

AlternativeAbstractConcrete

Figure 6. Pricing model

4 BUSINESS INTELLIGENCE CLOUDSERVICE NAVIGATOR

Regarding the description of a concrete BICloud service, a feature of the BI Cloud fea-ture model presented in section III can eitherbe said to be present or not. This presence orabsence can be represented visually by color-ing these two states of each BI Cloud servicedescription feature in a given graphical repre-sentation of the feature hierarchy.

Figure 7. BI-CSN: CSN extended by BI specific fea-tured (colored: green).

Several visual representations to describe andcharacterize BI Cloud services exist. For ex-ample, Grivas et al. [31] and Seufert and Bern-hardt [37] uses informal boxes and lines dia-grams, Chadha and Iyer [29] uses tables andother like Thompson and van der Walt [35] justuse text. We developed BI Cloud Service Nav-igators (BI-CSN) as enlargement of the exist-ing Cloud Service Navigator [8] for describinggeneric Cloud services. Thereby, BI-CSNs arebased on Sunburst diagrams. The goal of thisvisualization is to support the fast perceptionof service description.In this section, we first present our BI-CSN insubsection 4.1. Next to that, we discuss the


ISBN: 978-1-941968-19-2 ©2015 SDIWC 7

application in section 4.2 and the evaluation insubsection 4.3.

4.1 BI-CSN

The CSN was extended by the mentioned BIspecifics to be able to describe BI Cloud ser-vice offerings. The new BI specifics were ar-ranged in the BI-CSN shown in Fig. 7. Theroot of the feature hierarchy is placed at thecenter of the BI-CSN. The child features arethen concentric placed in surrounding rings.Features on the same level of hierarchy areplaced on the same ring as siblings, while subfeatures are placed in the outer sections thattheir ancestors cut from the pie. Present ser-vice descriptions are colored and absent fea-tures would be transparent.The green colored features in Fig. 7 are thenew BI specifics, which are added based onfactors of BI Cloud service offerings as resultof the survey and the consolidation. By usingthe BI-CSN to describe existing BI Cloud ser-vice the coloring represents the service proper-ties. Present features will be colored and ab-sent feature will be transparent. The applica-tion of the taxonomy by using the BI-CSN isfollowing in the next section.

4.2 Application

In total, the BI-CSN was applied to eight BICloud service offerings. Here, to not over ex-tend this paper, are shown two cases the BICloud service Power BI from Microsoft andthe BI Cloud service SAP HANA Cloud fromSAP. In the BI-CSN (Power BI in Fig. 8 andSAP HANA Cloud in Fig. 9) the blue coloredfeatures means that these features are presentand the transparent (white) features are absent.Figure 8 shows the specified BI-CSN for theBI Cloud offer Microsoft Power BI. MicrosoftOffice 365 with Microsoft Excel and supple-mented with Microsoft Sharepoint Online rep-resents the main components of Power BI.Power BI follows the SaaS approach and of-fers the possibilities to visualize, to model andto analyze data for example by using OLAPfunctions and Data Mining. Furthermore, the

analyzed data can be visualized through dif-ferent types of diagrams. Thereby, Power BIfollows the MaaS and VaaS approach. PowerBI is based on the Microsoft Platform Win-dows Azure which is hosted in a Virtual-PrivateCloud in Microsofts own data centers. Thescale of the demand can be performed in selfmanagement and the accounting is based onthe resources used. Microsoft offers detailedSLA with assurances e.g. in security, mainte-nance, response times. Support has to be adi-tionally orded and payed.

Figure 8. BI-CSN for Microsoft Power BI

An other BI Cloud service offering that wasused for the evaluation was the SAP HANACloud Platform, see Fig. 9. SAP HANA Cloudis offered as a Platform as a Service in thesense of a Data Warehouse System as platformservice (DWSaaS). On this plattform the VaaSand MaaS offerings SAP BI OnDemand andLumira Cloud are deployed. Further consid-ered here, is the platform offering. The SAPHANA Cloud is hosted by Amazon, so that athird party is involved. Just as Microsoft, SAPoffers detailed SLA aspects.


ISBN: 978-1-941968-19-2 ©2015 SDIWC 8

Figure 9. BI-CSN for SAP HANA Cloud

4.3 Evaluation

As part of a comprehensive literature review(see section 2), many approaches in the areaof BI Cloud have been discovered. We iden-tified and removed dublicates, unrelevant ref-erences and reviews and editorials. We identi-fies approaches of different groups - that’s whyour approach is based on a certain wide opin-ion. On the base of the survey and their analy-sis, the conceptualization of the taxonomy wasperformed as part of an interative design sci-ence process with repetetive incremental steps.The individual factors of the references weresuccessively compared, considered and consol-idated. This ensures that our architecture ison the current state of the status of scientific,industry and governmental and other associa-tions.The application is importance to demonstratethe feseability. So we took eight BI Cloudofferings and used our taxonomy to describethem (due to the brevity, the application is dis-played on two BI Cloud services, see section4.2). In the application it was found out, thatour architecture is completely inasmuch as allfeatures of the observed offerings were cov-ered.

Within expert dialogues on this subject, wepresented the taxonomy and the application toseveral offerings and discussed it. The com-pleteness and feasibility has been confirmedthere. By visualizing the consideration dealswith the BI-CSN, the comparability showedthe visually simple as possible. A comparisonof existing factors of a BI Cloud services is eas-ily possible by visually obervation of the spe-cific BI-CSN. Also, with the BI-CSN the re-quirements of customers can be described andthese can simply be compared with existing BICloud services to find the most appropriate ser-vice.

5 CONCLUSION

In this article, a taxonomy for the descriptionand comparison of BI Cloud services was in-troduced. This is a tool for orientation and un-derstanding of the subject area BI in the cloud.The taxonomy can also be used to identify andevaluate business needs and to support the se-lection of suitable offerings.To help overcome the prevalent skepticism ofenterprises regarding BI Cloud services, wesurveyed different frameworks for BI Cloudservice description. Based on this survey, weproposed a consolidated taxonomy in the formof a feature model to uniformly describe, com-pare and evaluate BI Cloud services, regardlessof whether an organizations own offers or theoffers of external providers are evaluated.In addition, we present the BI-CSN as a suit-able visualization technique to describe visu-ally and compare BI Cloud service offeringsand requests. Based on the application of theBI-CSN on existing services we evaluated anddemonstrated the feasibility and completeness.However, there are several open issues thatwe consider as future work: The proposedframework should be continuously adapted tothe developments in order not to lose validityand soundness by the appearance of new BIcloud service facets. An automatic selectionof BI Cloud offerings can be realized by us-ing the taxonomy for describing business re-quirements and to compare these with servicedescriptions to find a suitable service for the


ISBN: 978-1-941968-19-2 ©2015 SDIWC 9

considered scenario. Here in particular, exist-ing approaches from the service-oriented field(e. g. service oriented architecture (SOA), uni-fied service description language (USDL) etc.)could be used.For the future development of BI Cloud sys-tems and services this new domain shouldbe standardized. A standardized architecturecould provide transparency, reduce uncertaintyand help to raise the potentials. Some ques-tions of enterprise, system and software archi-tecture are not still clear and transparent clari-fied. Therefore many offerings are in an earlymaturity and heterogeneous.With this contribution is a first step in the di-rection of standardization is done. Regardingthe overall goal of the contribution, we con-sider our work as an important step towardsreducing the skepticism in BI Cloud services,since it provides a consolidated view on theirfeatures. We delivered a taxonomy that can beused for the description, comparison, selectionof BI Cloud services.

REFERENCES

[1] H. Chen, R. Chiang and V. C. Storey, “BusinessIntelligence and Analytics: From Big Data to BigImpact”, MIS quarterly, vol. 36(4), pp. 11651188,2012.

[2] H. Baars and H. G. Kemper, “Business intelligencein the cloud?”, PACIS 2010 Proceedings, Paper145, 2010.

[3] O. Norkus and H.-J. Appelrath, “Towards a Busi-ness Intelligence Cloud”, Proceedings of the ThirdInternational Conference on Informatics Engineer-ing and Information Science (ICIEIS2014) LodzUniversity of Technology, Lodz, Poland, pp 5566,September 2014.

[4] O. Norkus et al., “An Approach for a Cloud-basedContribution Margin Dashboard in the Field ofElectricity Trading”’, In: Douglas Cunningham,Petra Hofstedt, Klaus Meer, Ingo Schmitt (Hrsg.):INFORMATIK 2015, Lecture Notes in Informat-ics (LNI), Gesellschaft fr Informatik, Bonn 2015,Bonner Kllen Verlag, 9/2015.

[5] P. Mell and T. Grance, “The NIST Denition ofCloud Computing”, National Institute of Standardsand Technology, Information Technology Labora-

tory, Gaithersburg, NIST, Special Publication 800-145, September 2011.

[6] O. Norkus, B. Friedrich, F. Merkel, I. Schweer,“Margin Control as a Cloud Service” (Deckungs-beitragsrechnung als Cloud-Dienst), In: NorbertGronau (Ed.): ERP Management, 15(2), GITOVerlag, pp. 27-29, vol 5, 2015.

[7] H.-J. Appelrath, H. Kagermann and H. Krcmar,“Future Business Clouds: A contribution to theproject of the future Internet-based services for theeconomy” (Future Business Clouds: Ein Beitragzum Zukunftsprojekt Internetbasierte Dienste frdie Wirtschaft), Herbert Utz Verlag, 2014

[8] S. Gudenkauf, M. Josefiok, A. Goring andO. Norkus, “A Reference Architecture for CloudService Offers”, Software Engineering for Busi-ness Information Systems, 2013.

[9] M. Weber et al. “Cloud Computing Monitor 2015”,BITKOM KPMG , 2015.

[10] M. Weber et al. “Cloud Computing Monitor 2013”,BITKOMKPMG, 2013.

[11] M. Weber et al., “Cloud Computing - Evoluton inder Technik, Revolution im Business”, BITKOM-Leitfaden, BITKOM, 2009.

[12] R. Sallam et al., “Magic Quadrant for BusinessIntelligence and Analytics Platforms”, Gartner,2015.

[13] C. Bange, T. Grosser, N. Janoschek, “Big Data UseCases - Getting real on data monetization”, BARC,2015.

[14] C. Bange, S. Roggers, “Cloud Business Intelli-gence and Data Management as a Service - AGlobal Survey on Adoption, Challenes and Out-look”’, BARC, 2011.

[15] SAP http://www.sap.com/germany/pc/analytics/business-intelligence/software/data-visualization/cloud.html, last visit: 09.05.2015.

[16] SAP, http://go.sap.com/product/analytics/lumira/cloud.html,last visit:09.05.2015.

[17] SAP, https://websmp207.sap-ag.de/˜sapidp/011000358700001269622010E.pdf,last visit: 09.05.2015.


ISBN: 978-1-941968-19-2 ©2015 SDIWC 10

[18] Microsoft, http://www.microsoft.com/de-de/server-cloud/cloud-os/data-insights.aspx, last visit: 09.05.2015.

[19] Microsoft, https://www.microsoft.com/de-de/cloud/glossar/bi.aspx, lastvisit: 09.05.2015.

[20] Qlik, www.qlik.com/us/explore/products/, last visit: 09.05.2015.

[21] Oracle, http://www.oracle.com/us/solutions/business-intelligence/cloud-ready-oracle-bi-177505.pdf,last visit: 09.05.2015.

[22] S. Datta and D. Gupta, “Oracle Enterprise Man-ager Cloud Control 12c: Complete, Integratedand Business-Driven Cloud Management”, Ora-cle Corporation, 500 Oracle Parkway, RedwoodShores, CA 94065, U.S.A., Oracle White Paper,October 2011.

[23] M. Behrendt et al., “Introduction and ArchitectureOverview IBM Cloud Computing Reference Ar-chitecture 2.0”, IBM, Tech. Rep., 2011.

[24] IBM, http://www-01.ibm.com/software/analytics/cloud/, last visit:09.05.2015.

[25] MicroStrategy, “Architecture for enterprise busi-ness intelligence”, MicroStrategy, WhitePaper.

[26] L. Menon, B. Rehani “Business Intelligence onthe Cloud: Overview and Use Cases”, WhitePaper,TATA Consultancy Services, 2012.

[27] D. J. Abadi, “Data Management in the Cloud:Limitations and Opportunities”, IEEE Data Engi-neering Bull, 2009, 32 ed, pp. 312.

[28] S. Ouf and M. Nasr, “The cloud computing: thefuture of BI in the cloud”, International Journal ofComputer Theory and Engineering, 2011, 3rd ed,pp. 750754.

[29] B. Chadha and M. Iyer, “BI in a Cloud: Definingthe Architecture for Quick Wins”, SETLabs Brief-ing, 2010, 8 ed, pp. 3944.

[30] C. Tamer, M. Kiley, N. Ashrafi, J. Kuilbar, “Riskand benefits of business intelligence in the cloud”,In Northeast Decision Sciences Institute AnnualMeeting Proceedings (pp. 86-95).

[31] S. G. Grivas, T. U. Kumar, and H. Wache, “Cloudbroker: Bringing intelligence into the cloud”,

IEEE 3rd International Conference on Cloud Com-puting (CLOUD), pp. 544-545, 2010.

[32] H. Demirkan and D. Delen, “Leveraging the capa-bilities of service-orientied decision support sys-tems: Putting analytics and big data in cloud”, De-cision Support Systems 55, Journal, Elsevier B.B.,pp. 412-421, 2013.

[33] Y. S. Gurjar and V. S. Rathore. “Cloud businessintelligenceis what business need today”, Interna-tional Journal of Recent Technology and Engineer-ing, vol. 1.6, pp 81-86, 2013.

[34] D. Gash, T. Ariyachandra and M. Frolick, “Look-ing to the clouds for business intelligence”, Jour-nal of Internet Commerce vol. 10.4, pp. 261-269,2012.

[35] W. J. Thompson und J. S. van der Walt, “Businessintelligence in the cloud”, SA Journal of Informa-tion Management, vol. 12(1), pp. 1-5, 2010.

[36] T. Haselmann and G. Vossen, “Database-as-a-Service for small and mediem enterprises”(Database-as-a-Service fr kleine und mittlereUnternehmen), Working Paper, Institut furWirtschaftsinformatik, Westfalische Wilhelms-Universitat Munster, Munster, vol. 3, 2010.

[37] A. Seufert and N. Bernhardt, “Business Intelli-gence and Cloud Computing” (Business Intelli-gence und Cloud Computing), HMD Praxis derWirtschaftsinformatik, vol. 47(5), pp. 3441, 2010.

[38] N. Bernhardt and K. Balluch, “Self-service Busi-ness Intelligence in the Cloud - requirements, se-curity aspects, concepts” (Self-Service BusinessIntelligence in der Cloud - Anforderungen, Sicher-heitsaspekte, Konzepte), BI Spektrum, vol. 1, pp.32-37, 2014.

[39] C. Weinhardt et al., “Cloud Computing - A Clas-sification, Business Models and Reserach Direc-tions”, Business & Information System Engineer-ing, pp. 391–399, vol 5, 2009.

[40] M. Torkashvan, H. Haghighi, “CSLAM: A frame-work for cloud service level agreement manage-ment based on WSLA”, Sixth International Sym-posium on Telecommunications (IST), IEEE, pp.577–585, 2012.

[41] M. Torkashvan, H. Haghighi, “A service orientedframework for cloud computing”, Proceedings ofthe 3rd International Conference on Informationand Communication Systems, ACM, paper no. 25,2012.


ISBN: 978-1-941968-19-2 ©2015 SDIWC 11

[42] S. Leinmeister, M. Bohm, C. Riedl, Christoph, H.Krcmar, “The business perspective of cloud com-puting: actors, roles and value networks”, ECIS2010 Proceedings, Paper 56, 2010.

[43] N. Schirm, T. Frank, M. Henkel, F. Bensberg,“Success factors for cloud-based cusiness intel-ligence solutions” (Erfolgsfaktoren cloudbasierterBusiness Intelligence Losungen), Proceedings der12. Internationalen Tagung Wirtschaftsinformatik(WI2015), 2015.

[44] V. Chang, “The business intelligence as a servicein the cloud”, Future Generation Computer Sys-tems, Volume 37, pp. 512-534, 2014.

[45] J. Ereth, D. Dahl, “Business intelligence in thecloud: fundamentals for a service-based evaluationconcept”, Tagungsband des 5. Workshops ”Busi-ness Intelligence” der GI-Fachgruppe Business In-telligence , Paper No. 1, 2013.

[46] D. Batory, “Feature Models, Grammars and Propo-sitional Formulas”, Proceedings of the 9th inter-national conference on Software Product Lines,ser. SPLC05. Berlin, Heidelberg: Springer-Verlag,2005, pp. 720.

[47] K. Czarnecki and U. W. Eisenecker, “GenerativeProgramming: Methods, Tool, and Applications”,6th ed, Addison Wesley, April 2005.

[48] K. Czarnecki and S. Helsen, “Feature-based sur-vey of model transformation approaches”, IBMSystems Journal, vol. 45, no. 3, pp. 621646, 2006.

[49] S. Kolb and G. Wirtz, “Towards application porta-bility in platform as a service”, Service OrientedSystem Engineering (SOSE), 2014 IEEE 8th In-ternational Symposium on. IEEE, 2014.

[50] A. Goring, S. Gudenkauf, M. Josefiok andO. Norkus, “A taxonomy for describing cloudservice offers” (Eine Taxonomie zur Beschreibungvon Cloud-Dienstangeboten), VDE-Kongress2014, VDE VERLAG, 2014.

[51] V. Andrikopoulos, A. Darsow, D. Karastoy-anova, F. Leymann, “CloudDSFThe Cloud Deci-sion Support Framework for Application Migra-tion”, Service-Oriented and Cloud Computing, pp.1-16, Springer, Berlin Heidelberg 2014.

Table 1. Results of Literature Analysis

Reference Level of FormalityAbstraction

met

am

odel

aspe

ct

inst

ance

info

rmal

inpa

rtfo

rmal

form

al

BITKOM [9, 10, 11] X XGartner [12] X XBARC [13, 14] X XSAP [15, 16, 17] X XMicrosoft [18, 19] X XQlikTech [20] X XOracle [21] X XIBM [23, 24] X XMicroStrategy [25] X XTATA [26] X XAdabi [27] X XOuf and Nasr [28] X XChadhaand Iyer [29] X XTamer et al. [30] X XGrivas et al. [31] X XDemirkanand Delen [32] X XBaars andKemper [2] X XGurhar andRathore [33] X XGash et al. [34] X XThompson andvan der Walt [35] X XHaselmannand Vossen [36] X XSeufert andBernhardt [37] X XBernhardtand Balluch [38] X XWeinhardt et al. [39] X XTorkashvan and X XHaghighi [40, 41]Leinmeister X Xet al. [42]Schirm et al. [43] X XChang [44] X XEreth and Dahl [45] X X


ISBN: 978-1-941968-19-2 ©2015 SDIWC 12

a taxonomy for describing bi cloud servicesd.researchbib.com/f/1nagv2bghhptez.pdfa taxonomy for...

Documents