Looking to the Clouds for Business Intelligence

Looking to the Clouds for BusinessIntelligence

DAVID GASHAdvisory Services, Sogeti USA, Cincinnati, Ohio, USA

THILINI ARIYACHANDRA and MARK FROLICKWilliams College of Business, Xavier University, Cincinnati, Ohio, USA

As cloud computing becomes more prevalent, information tech-nology (IT) groups large and small are looking for guidance asto how they can leverage this new resource. This article specificallyfocuses in on the field of business intelligence (BI) and works toprovide a framework for evaluating and moving out of atraditional in-house hosted BI environment to one hosted withinthe cloud.

KEYWORDS business intelligence, cloud computing

INTRODUCTION

Executives today are constantly looking for ways to gain insight into theirbusinesses, their industry, and the world around them, but as history tellsus, this longing for knowledge is nothing new. Early peoples across a multi-tude of civilizations have all had this same desire, to be given answers to theirquestions and a direction in their lives. One authoritative source that ancientpeople have turned to was the clouds. Open and accessible to everyone, theclouds provided insight into the most important questions of life. In much thesame way, business leaders today put their faith in business intelligence (BI)to give them a competitive advantage and direction.

However, this competitive advantage is not as accessible as the openskies, but instead comes at a significant cost. Corporations, implementing arobust BI framework, often spend millions of dollars in initial investmentand millions more each year to sustain that framework. Unlike simply lifting

Address correspondence to Mark Frolick, Xavier University, Williams College of Business,3800 Victory Pkwy., Cincinnati, OH 45207, USA. E-mail: [email protected]

Journal of Internet Commerce, 10:261–269, 2011Copyright # Taylor & Francis Group, LLCISSN: 1533-2861 print=1533-287X onlineDOI: 10.1080/15332861.2011.622694

261

your head to gaze up at the clouds, a BI framework requires systems to beput in place to gather and centralize data, custom reports written to interpretthat data, and teams of people to support that framework. All of these costsmake the benefits of BI less accessible than those garnered from the openskies (Thompson 2009).

For the sake of simplicity, we define a BI framework in the most basicterms. As pictured below in figure 1, this framework consists of source sys-tems that create the data; extract, transform and load (ETL) servers that movethe data; data warehousing servers that house the data; reporting servers thatdisplay the data; and end users that consume the data.

For many companies this initial investment into a BI framework is toogreat a risk to put on the limited resources that they have available. Comingfull circle, the ‘‘Clouds’’ offer the potential to make BI more accessible to anycompany. Just as the clouds were an external source of divination for ancientpeoples, free and accessible to all, Cloud Computing offers companies a wayto avoid the significant upfront investment that an internal BI frameworktraditionally requires (White 2008).

The purpose of this article is to explore proposed benefits and risks tousing Cloud Computing for BI and to propose a framework for evaluatingand implementing such a framework. To get to this framework, first theadvantages and disadvantages of a Cloud-based environment is describedand follows the presentation of a specific roadmap to help a companyevaluate and implement BI in the Cloud.

LOOKING TO THE CLOUDS

With the overall BI market growing by 22% in 2008 followed by a 4.2%increase in 2009, to a global market size of over $9.3 billion, companiesare putting money towards BI (Gartner 2009, 2010). As this market growsand as the complexity and costs of these platforms grow, information tech-nology (IT) organizations are looking for new and better ways to moreefficiently manage these systems. To this end, Hewlett Packard’s top 10trends in BI have, for the second year in a row, mentioned a ‘‘growinginterest in cloud computing for BI’’ (Hewlett Packard 2010).

FIGURE 1 Traditional business intelligence framework (color figure available online).

262 D. Gash et al.

Cloud Computing offers businesses a way to overcome the upfrontand much of the ongoing costs normally deterring companies from imple-menting BI.

Upfront Investment Costs

One major cost that companies must address is the upfront capital investmentin hardware and software required to get a BI platform off the ground. Asmentioned previously, there are many pieces that make up a BI framework,and the hardware and software that make up these pieces can quickly add upto millions of dollars. Hosting this environment in a cloud, however, canreduce both the hardware and software footprint that a company needs topurchase. As pictured below in figure 2, ETL servers, data warehousing ser-vers, and reporting servers and the software associated with them can all beshifted into the cloud, leaving only the source systems and the end users inthe company’s domain. This can be achieved in a number of practical andsecure manners, including creating a Web service to expose a company’ssource systems to the cloud provider or by simply establishing a secure virtualprivate network tunnel from the cloud provider into a company’s intranet.

Personnel Costs

Before putting anything in place, however, one of the first costs a companyincurs in implementing a BI solution is staffing teams to design, implement,and support the platform (Pineau 2009). By implementing a Cloud-based sol-ution, much of these staffing needs can be reduced or eliminated. From adesign standpoint, a Cloud provider has a pre-designed solution for a cus-tomer, therefore, reducing any design effort associated with setting up theinfrastructure and integrations between the different pieces within a BI plat-form. Support costs are also reduced, as a company is no longer required to

FIGURE 2 Cloud-based business intelligence framework (color figure available online).

Looking to the Clouds for Business Intelligence 263

handle the support of the hardware and software that make up the majorityof the framework, as these are shifted into the Cloud.

Growth and Expansion Costs

Even if a company can get through the initial capital investment of bringingthe infrastructure online required for a BI platform, this usually tends to be afirst phase. Additional hardware expansions are usually required as BIbecomes adopted throughout a company, to account for increased usage.With Cloud Computing, however, the sky is the limit in terms of expandedcapacity. Cloud-based platforms offer a ‘‘pay as you grow’’ model, allowinga company to expand and contract the size of its environment without theneed for additional capital investments and without being locked into aspecific infrastructure size (Kennedy 2009).

Depending on the amount of data and the location of users, oneadditional cost that a company may incur is the cost of distributing the BIinfrastructure across geographies. One way to resolve this is to replicate dataacross different datacenters across the globe. This greatly increases the hard-ware, software, and personnel investment required for a BI platform.However, with Cloud Computing, the Cloud provider can distribute anenvironment across many locations for a consistent user experience withouta significant investment.

Cloud Computing also offers a company built-in redundancy for disasterrecovery scenarios. Setting up dedicated disaster recovery failover environ-ments and backup procedures can quickly add up to a significant part of acompany’s overall BI budget. Cloud Computing takes the burden off of acompany by managing this fully within the Cloud. While Cloud Computingoffers many benefits to BI implementations, Cloud Computing does notcome without specific concerns.

DARKENING CLOUDS, CONCERNS WITH CLOUD COMPUTING

There are several concerns that must be weighed against the proposed ben-efits of the cloud for BI. Specifically, companies evaluating a Cloud-basedsolution should consider security, data ownership, legal implications,and concerns around the technical direction of a third party managing thisframework (Stamper 2008).

Security Concerns

One of the first concerns that many companies express is around security,including opening source systems, data traveling from inside a companieswalls to a third party, and trusting that third party to manage a secure

264 D. Gash et al.

environment (Kavis 2009). In order for a company to implement a cloud sol-ution, they must open their source systems to a cloud provider. This opensthe door for hackers and or malicious competitors to piggy back into a com-pany’s environment and gain access into those systems. Another securityconcern is around data traveling over the Internet. At the beginning whenoperational data is sent, and at the end, when reports are hosted in the cloud,data is vulnerable whether encrypted or not. Finally, once that data is housedwithin the cloud provider’s environment, a company must trust that the thirdparty provider is doing everything it can to maintain a secure environment(Shipley 2010).

Data Ownership Concerns

Companies must also consider implications of data ownership. In a tra-ditional BI environment, data ownership concerns are kept in-house andare typically resolved by management within the company. In a Cloud-basedsolution, however, a third-party vendor is involved, making these concernsslightly more complicated. Although the data resides within the cloud provi-der’s domain, a company still wants to have a say in who has access to thatdata and how that data is shared. Though this issue may be more legal thantechnical, a company must still weigh this when considering a Cloudsolution.

One special case concerning data ownership within the cloud deals withthe legal ramifications of export-controlled data. Many governments acrossthe globe put certain restrictions on specific data. When audited, a companymust prove that they are adhering to the restrictions placed on this data.However, in a Cloud environment, a company must trust its Cloud providerto adhere to restrictions down to the citizenship of support personnel hand-ling systems containing export-controlled data. As governments can levyheavy fines on companies failing to adhere to these restrictions, companiesmust judge the benefits of a Cloud offering in light of these risks (MacVittie2009).

Software Control Concerns

Another major concern with the use of a Cloud-based environment is the riskassociated with being locked in to a set of technologies dictated by a Cloudprovider. With an in-house BI platform, the technologies used can gothrough a rigorous screening process. Products being evaluated can bejudged on a company’s specific usability needs, security concerns, and costrestriction. However, in the cloud, a single company’s specific needs maynot be taken into consideration or heavily weighed (Baker 2010). Therefore,a Cloud provider, based on the provider’s desire to meet a wider audience of

Looking to the Clouds for Business Intelligence 265

customers, may not offer the best tools or software available for a company(Agrawal 2010).

Data Volume Concerns

With the increased adoption of BI across the business community, the sheervolume of data being transformed, housed, and analyzed is rapidly increas-ing. Performance within a local environment is a concern today for many BIplatforms, with data remaining within the walls of the company. One poten-tial risk for a company moving into the Cloud is the volume of data beingsent between the company and their Cloud provider. This covers both duringan initial load of data from source systems housed within a company’s intra-net as well as the size and volume of reports being pulled down by endusers. The size and timing of data transfers to and from the Cloud provideris a major concern for any company looking to move into the Cloud.

SETTING OUT THE RIGHT PATH TO CLOUD COMPUTING

As discussed above, there are many proposed risks and benefits to aCloud-based BI platform. Once a company has made the decision to lookin that direction, however, the path to getting there may not be clear. Belowis a proposed roadmap to guide companies through this process. As picturedbelow in figure 3, this is a cyclical process of evaluation, design, build, sup-port, and reevaluation. It should be stressed that this is not a static solution,

FIGURE 3 Roadmap to a cloud-based business intelligence framework (color figure availableonline).

266 D. Gash et al.

but one geared toward continual improvement and growth in the Cloud(Agrawal 2010).

The first step for a company to take is around evaluating their currentenvironment. First, a company must conduct an analysis from a data perspec-tive, determining where their source data resides, how often that data isupdated, what that data is used for, and if there are any restrictions aroundaccess to that data. Next, a company should look at what tools are in placetoday, both formal and informal, that constitutes a BI framework. A companyshould then evaluate their existing end users, what roles do they currentlyfill, where they are the located, and how do they use the data and tools thatare in place today. Last, a company needs to evaluate any special require-ments around this system, such as is there a need to increased securityaround certain data? Is there a need for access from other parts of the globe?Once a company has done the groundwork of understanding what is in placetoday, as detailed in table 1 below, they can move forward to design asolution within the Cloud.

The second step in this roadmap is for a company to work through thedesign of a Cloud-based solution. In a Cloud-based solution, the design effortfocuses around the two ends of the BI framework. First, a company mustdesign a solution for exposing their data to the Cloud provider, eitherthrough a Web service or secure tunnel. With this, as laid out in table 1below, considerations have to be made for the amount of data being trans-ferred, speed at which that data must be transferred, and security concernsaround accessing that data. Next, a company must work with the Cloudprovider to design the appropriate reporting structure and specific

TABLE 1 Roadmap to the Cloud

Evaluate Design

. Analysis around the data.� Where does it reside?� How often is it updated?� What is it used for?

. Expose data to the Cloud.� Amount of data� Speed of data� Security concerns

. What tools are in use today?� Formal and informal

. Design reporting framework.� What are the customer needs?

. Any special needs?� Security requirements� Global access

Build Support

. Create integrations.� Who will support these?

. Day-to-day support of environment.� Are procedures in place to work

with the Cloud Provider?. Write reports.

� Work with the CloudProvider and business users.

. Monitor usage for growth.� Response times� User feedback

Looking to the Clouds for Business Intelligence 267

reports required by the end user. The design must also ensure that thecompany meets client side requirements in order to access the reports.After the design is complete, the roadmap next takes a company on toimplementation.

In the build phase of the road map, a company must not onlyimplement the designed solution, but also put in place the appropriate sup-port staff to maintain that solution. Working off the design created in the pre-vious step, companies must work collaboratively with the cloud provider toimplement the integrations required to expose the data to the Cloud pro-vider. Once the data is in the Cloud, a company must work with experts fromthe Cloud vendor to create reports to meet the needs uncovered in the evalu-ation phase of the roadmap, referenced in table 1. Lastly, a companymust build out the appropriate support structures to keep the frameworkfunctioning properly.

Moving to the Support phase of the roadmap, a company must not onlysupport the day-to-day operations of the Cloud-based environment, but alsomonitor its usage. Much of what a company does in this phase falls within therealm of a traditional support structure, from troubleshooting user accessissues, to dealing with more complex outages. However, as listed below intable 1, a company must also monitor system usage, measure response timesfrom different user locations, solicit user feedback, and recognize areas ofgrowth. Companies must actively monitor their system as a whole and gatherthe right data to take them back to the first phase of evaluation.

Coming full circle, a company within the Cloud environment mustrecognize the need for continual improvement in order to realize the fulldynamic potential of the Cloud. Though this roadmap, detailed above intable 1, does have a first step, it intentionally does not have a last step. Itsintention is to provide a roadmap for the continual improvement of a com-pany’s Cloud-based BI platform. By moving through the phases of Evalu-ation, Design, Build, and Support on a cyclical basis, a company cancontinually refine their Cloud model to better meet the BI needs of its users.

CONCLUSION

With all of the proposed benefits of BI, companies must consider certaincosts associated with a traditional framework. Using a Cloud-based BI plat-form can help a company reduce many of the costs traditionally associatedwith a BI implementation. Proposed above is a basic framework companiescan follow to approach a migration into the Cloud. As more and more com-panies will soon discover, when implemented properly, a Cloud-based plat-form makes the benefits of BI more accessible to companies of any size. Justas early peoples were able to gaze into the heavens for guidance, CloudComputing unlocks the benefits of BI for all companies.

268 D. Gash et al.

As with most areas of IT, however, BI in the Cloud may be approacheddifferently depending on the size of a company. Most mid- to large-size com-panies already have in place a robust BI framework. In which case, the Cloudcan be looked to for a quick and inexpensive development environment orto introduce new technologies to users without having to fully commit to aproduct. With smaller companies, on the other hand, mature BI environ-ments are not common, if in place at all. In this case, the Cloud is a perfectway for a smaller company to test out BI without all the overhead associatedwith a traditional implementation.

REFERENCES

Agrawal, P. 2010. What you should know before moving into the cloud. http://www.ebizq.net/topics/cloud_computing/features/12252.html (accessed March28, 2010).

Baker, A. 2010. SaaS rolls on despite poor governance. http://www.information-management.com/news/saas_software_as_a_service_project_management-10017790-1.html (accessed May 4, 2010).

Gartner Incorporated. 2009. Gartner says worldwide business intelligence, analyticsand performance management grew 22 percent in 2008. http://www.gartner.com/it/page.jsp?id=1017812 (accessed April 26, 2010).

Gartner Incorporated. 2010. Gartner says worldwide business intelligence, analyticsand performance management software market grew 4 percent in 2009. http://www.gartner.com/it/page.jsp?id=1357514 (accessed April 26, 2010).

Hewlett-Packard Development Corporation. 2010. Top 10 trends in business intel-ligence for 2010. http://h20195.www2.hp.com/v2/GetPDF.aspx/4AA0-6420ENW.pdf (accessed April 26, 2010).

Kavis, M. 2009. The real problem with cloud computing . . .people. http://socialcomputingjournal.com/viewcolumn.cfm?colid=823 (accessed March 22, 2010).

Kennedy, G. 2009. Cloud computing and business intelligence. http://cloudcomputing.sys-con.com/node/1086696 (accessed March 22, 2010).

MacVittie, L. 2009. The Business Intelligence A Cloud Paradox. http://cloudcomputing.sys-con.com/node/1071172 (accessed March 22, 2010).

Pineau, J. 2009. SaaS & Business Intelligence at Dreamforce. http://cloudcomputing.sys-con.com/node/1202343 (accessed March 22, 2010).

Shipley, G. 2010. Cloud computing risks. Information Week 1262:20–24.Stamper, J. 2008. Is the enterprise ready for cloud computing? http://bi.cbronline.

com/comment/is_the_enterprise_ready_for_cloud_computing_291008 (accessedMarch 22, 2010).

Thompson, J. K. 2009. Business intelligence in a SaaS environment. Business Intelli-gence Journal 14 (4): 50.

White, C. 2008. Business intelligence in the cloud: Sorting out the terminology.http://www.b-eye-network.com/~ben/view/8122 (accessed March 22, 2010).

Looking to the Clouds for Business Intelligence 269

Copyright of Journal of Internet Commerce is the property of Taylor & Francis Ltd and its content may not be

copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written

permission. However, users may print, download, or email articles for individual use.