BUSINESS INTELLIGENCE IN THE CLOUD                                  ShimaaOuf                                             ...
expensive analytics or reporting. [11]Cloud computing istransforming the economics of BI and opens up theopportunity for s...
Virtualization      and    cloud-based          infrastructure        will                Merger and acquisition frenzy wi...
assembling customer-specific solutions more quickly.                    for reports. It makes it seamless to add new users...
Fig. 2 Business Intelligence Proposed Model                                                                               ...
need to store the staging data and queries are not that                            influences the way business intelligenc...
Upcoming SlideShare
Loading in …5
×

Business intelligence in_the_cloud

571 views
522 views

Published on

Published in: Technology, Business
0 Comments
1 Like
Statistics
Notes
  • Be the first to comment

No Downloads
Views
Total views
571
On SlideShare
0
From Embeds
0
Number of Embeds
0
Actions
Shares
0
Downloads
32
Comments
0
Likes
1
Embeds 0
No embeds

No notes for slide

Business intelligence in_the_cloud

  1. 1. BUSINESS INTELLIGENCE IN THE CLOUD ShimaaOuf Mona Nasr Faculty of Computers & Information Faculty of Computers & Information Helwan University, Egypt. Helwan University, Egypt. shimaaouf@yahoo.com m.nasr@helwan.edu.egAbstract: Business Intelligence (BI) deals with integrated characterized by ability to consume resources as required inapproaches to management support. Currently, there are an elastic manner and scaling the consumption arbitrarily asconstraints to BI adoption and a new era of analytic data required. The advent of infrastructure as a service impliesmanagement for business intelligence these constraints are the that Computational power is available on demand and on aintegrated infrastructures that are subject to BI have become pay-as-user-go basis with similar characteristics applying tocomplex, costly, and inflexible, the effort required storage of data as well. This enables a layer of servicesconsolidating and cleansing enterprise data and Performanceimpact on existing infrastructure 1 inadequate IT sitting on top of this infrastructure to decouple the deliveryinfrastructure. So, in this paper Cloud computing will be used aspect of the services from the core business orientedas a possible remedy for these issues. We will represent a new aspects involved in these services. Related to this, is the factenvironment atmosphere for the business intelligence to make that storage of the underlying data is also decoupled andthe ability to shorten BI implementation windows, reduced segregated from the services. [4]cost for BI programs compared with traditional on-premise BIsoftware, Ability to add environments for testing, proof-of­ II. BUSINESS INTELLIGENCEconcepts and upgrades, offer users the potential for fasterdeployments and increased flexibility. Also, Cloud computing Business intelligence (BI) has been referred to as theenables organizations to analyze terabytes of data faster and process of making better decisions through the use of people,more economically than ever before. Business intelligence (BI)in the cloud can be like a big puzzle. Users can jump in and processes, data and related tools and methodologies. Theput together small pieces of the puzzle but until the whole roots of business intelligence are found in relationalthing is complete the user will lack an overall view of the big databases, data warehouses and data marts that helppicture. In this paper reading each section will fill in a piece of organizing historical information in the hands of businessthe puzzle. analysts to generate reporting that informs executives and senior departmental managers of strategic and tactical trends Keywords: Cloud Computing, business Intelligence, platform and opportunities. [8] In recent years, business intelligenceas a services. has also come to rely on near real-time operational data found in systems including enterprise resource planning I. INTRODUCTION (ERP), customer relationship management (CRM), supply chain, marketing and other databases. "Operational" BI is In the wake of the economic slowdown, organizations meant to provision many more functions in the organizationare increasingly looking for ways to do more with the same with role-specific dashboards and scorecards and isresources; articulate differently - to make every penny, input increasingly tied to the topics of performance managementand contribution count. In such situations, technologies like and business process management. Inherent to any form ofCloud Computing and Business Intelligence (81) are BI is the notion of data quality, consistent and dependablebecoming increasingly important in gaining and maintaining data and the processes involved in its creation anda competitive edge. [5]These technologies when combined maintenance. [12]Business Intelligence involves intelligentenable a variety of new analytic data management projects reporting on top of existing data which helps in prompt andand business possibilities. Cloud computing will change the actionable decision making. These decision making mighteconomics of BI by making available the hardware, involve "geography based investment decision for anetworking, security and software needed to create data multinational company" or even a "buy decision for amarts and data warehouses on demand with a pay-as-user-go product by the consumer". BI has evolved over time but theapproach to usage and licensing. More and more businesses key components still continue to hold true. It is stillare turning to analytic applications to provide critical necessary to be able to aggregate the factual data frombusiness insights. [8]Whether focused on achieving higher various data sources and doing involved transformations.Return On Investment (ROI), better understanding of the This data then either needs to be stored in a data mart orcompetitive landscape, improving product and service warehouse to enable reporting and analysis on top or it couldquality, Business Intelligence is one of the few technologies then be further aggregated into metrics which are thenthat can equip organizations to more effectively prepare for reported. Nevertheless the ability to perform BI involves keytomorrow today. Its no wonder the BI platform is expected aspects related to data management and computationallyto grow by 7.9% through 2012. [11] Cloud computing is978-1-61284-486-2/111$26.00 ©2011 IEEE 650
  2. 2. expensive analytics or reporting. [11]Cloud computing istransforming the economics of BI and opens up theopportunity for smaller enterprises to compete using theinsights that BI provides. Cloud-based analytics will impactBI by:Accelerating HI technology adoption: the cloud becomingthe default platform for evaluating new software.Easier evaluation: the cloud enables software companies tomake new technology available to evaluators on a self­services basis, avoiding the need to download and set upfree software downloads. [16]Increased short-term ad-hoc analysis: avoiding data martsspawned as a result of new business conditions or events.Where short term needs [weeks or months] for BI isrequired, cloud services are ideal. A data mart can create in Fig. 1 Cloud computing [1]a few hours or days, used for the necessary period, and then The cloud computing wave is the most dramatic changethe cloud cluster cancelled, leaving behind no redundant we have observed in the computing industry since the wavehardware or software licenses. The cloud makes short term of the Internet. Cloud computing will significantly changeprojects very economical. [II] data centers and IT organizations as well as the infrastructureIncreased flexibility: due to the avoidance of long term and software vendors business models. In fact, the cloudfinancial commitments, individual business units will have computing wave is not just a wave - it is more like a tsunami.the flexibility to fund more data mart projects. This is ideal What is causing this cloud tsunami? [21]We start with listingfor proof of concept, and ad-hoc analytic data projects on­ 4 of the Gartner top 10 IT predictions for the next three todemand. This agility enables isolated business units to five years for cloud computing, software as service (SaaS),respond to BI needs faster than their competitors and data center power/cooling efficiency and open sourceincrease the quality of their strategy setting and execution. software. All of these predictions indicate that data center[17] efficiency and cost containment will transform the ITGrowth considerations industry over the recent & next years. Key predictions forAs data volumes grow, for analytic cloud projects to the data center for the next 5 years: [20]succeed they will require a database architecture that is • In 2011, early technology adopters will forgo capitaldesigned to function efficiently in elastic, hosted computing expenditures and instead purchase 40 percent of theirenvironments like the cloud. At a minimum, such databases IT infrastructure as a service.must include the following architectural features: • By 2012, at least one-third of business application "Scale-out" shared nothing architecture to handle software spending will be as service subscription changing analytic workloads as elastically as the cloud. instead of as product license. Aggressive data compression - to keep storage costs • In 2009, more than one-third of IT organizations will low. have one or more environmental criteria in their top Automatic grid replication and failover- to provide high six buying criteria for IT-related goods. Initially, the availability in the cloud. [11] motivation will come from the wish to contain costs. Enterprise data centers are struggling to keep pace III. THE CLOUD COMPUTING TSUNAMI with the increasing power requirements of their (EFFICIENCY AND COST CONTROL IN THE IT infrastructures. INDUSTRY) By 2012, 80 percent of all commercial software wi II include elements of open source technology. Cloud computing has become one of the hottest These predictions clearly indicate that data center efficiencybuzzwords in the IT area. Many companies and institutions and cost are forming massive waves for increasing dataare rushing to define clouds and provide cloud solutions in center efficiency with virtualization, better server utilization,various ways. [1]Cloud computing is Internet-based cost reduction with more efficient power/cooling and thecomputing, whereby shared resources, software, and leveraging of open source software. [18]information are provided to computers and other devices on The Cloud Impact on Data Centers:demand, like the electricity grid. Due to the fact it involves • A mix of private and public clouds will become thethe existence of data centers that are able to provide services; norm.the cloud can be seen as a unique access points for allrequests coming from the world wide spread clients. [19] • Many organizations and lines of business will bypass IT to secure cloud-based infrastructure and SaaS applications. 651
  3. 3. Virtualization and cloud-based infrastructure will Merger and acquisition frenzy will become the norm become the norm. for survival.[7,S] Power and cooling efficiency and green data centers IV.THE BENEFITS OF CLOUD COMPUTING FOR will become critical and the norm. BUSINESS INTELLIGENCE A new breed of cloud computing skills will become common in data center operations. The convergence of the Cloud, Open Source, and Private cloud technology, such as cloud storage, will specialized analytic databases has brought a new world of find its way to the IT organizations. possibilities to businesses. No longer are dedicated hardware Service oriented architectures (SOAs) will drive the IT and software required to meet ad hoc or unpredictable BI infrastructure and application architecture. [11] requirements. The combination of these technologiesThe Cloud Impact on IT Organizations: running in the cloud will enable organizations to: [2] Create analytical sandboxes on demand to optimize- We will see a transformation from programming to serviceintegration and customization. response to sudden market changes.- With the cloud and SaaS usage-based pricing, IT budgets Offload expensive processing from overburdened datawill transform from CAPEX to more OPEX, opening the warehouses and avoid costly upgrades.door for immediate IT investments.[16] Prototype new proof-of-concept BI applications without purchasing and implementing new data centerThe Cloud Impact on Cloud Computing Vendors:- Significant market growth and momentum will fuel hyper infrastructure.growth. Run full business intelligence and data warehousing for- Cloud infrastructure utilization and efficiency will become existing cloud applications entirely in the cloud.[12]critical to success. V. CLOUD BUSINESS INTELLIGENCE: STORMY- Power and cooling costs will become enormouslyimportant factors to profitability. FUTURE- Cloud infrastructure and SaaS vendors will become the Much has been said and written about cloud computing.new giants of the industry where the IT operations shop for The cost and scale advantages it can offer public and privateinfrastructure and SaaS applications. computing infrastructures are astounding - providing- Merger and acquisition frenzy will become the norm for genuine hope for more efficient and far-reaching, next­hyper growth.[17,IS, 19] generation computing. In fact, cloud BI holds great potentialThe Cloud Impact on SaaS Vendors: to substantially disrupt the business intelligence market There will be hyper growth in the number of SaaS because of its perfect storm of low-cost, scalability, and applications and vendors. flexibility. With business intelligence software running in Venture spending will grow significantly. the cloud, it is still possible to make comprehensive New requirements and standards for APIs, reporting, integration with back-end systems - both within the user security and service-level agreements (SLAs) will company and in the cloud. [13]Cloud computing emerge. environments mean that organizations no longer need to SaaS vendors will become the main source of expend capital upfront for hardware and software purchases. applications. [1] Nor do they need to suffer through prolonged in-houseThe Cloud Impact on Infrastructure Vendors: implementations. In these two areas, cloud computing and Server, storage and networking customer influence will SaaS models both share similar benefits. [1]Cloud BI decrease. represents a way for reporting and analysis solutions to be Server, storage and networking vendors will be selling developed, installed, and consumed more easily due to its to cloud vendors. lower cost and easier deployment. Ideally, a cloud-based Infrastructure vendors will be fighting for mind-share business intelligence platform makes use of infrastructure­ as-a-service (IaaS), complements and extends todays with both cloud and SaaS vendors. platform-as-a-service (PAAS), utilizes an on-demand, Infrastructure vendors will lose contact with many virtualized, elastic software and hardware environment, and enterprise customers as they flock to cloud delivers application-level functionality as a service infrastructure and SaaS. (commonly referred to as software-as-a-service). [2] Merger and acquisition frenzy will become the norm for survival. Cloud-based HI platform is used to solve one of three Infrastructure vendors will experience a dramatic primary customer needs: change of business model. [15] 1. As a horizontal BI tool to deliver standalone,The Cloud Impact on Application Software Vendors: internally facing reporting and analysis applications Application software vendors will have to adopt the probably using a traditional relational database (or data SaaS model to survive. mart) as the primary source data system. They will lose business to SaaS companies. 2. As an application framework or pre-built reporting and Software licensing will dramatically change. analysis template for systems integrators to use for 652
  4. 4. assembling customer-specific solutions more quickly. for reports. It makes it seamless to add new users as well These solutions are probably function or domain allow user. specific and contain reusable components and As more and more applications and data sets move to application logic (but are assembled uniquely for each the cloud, BI services need to adapt to look at Cloud as customer). [3] the data source. [14] 3. As a development platform that enables embeddable, externally-facing applications that solve a function­ VI. THE PROPOSED MODEL specific data analysis problem (for example, CRM analytics, financial analytics, or supply chain analytics). The future will be very bright for the use of BI in the In this case, an ISV (or an enterprise IT team with cloud, both because of the advantages that underpin this new appropriate skills) would probably use the BI platform computing paradigm as well as the explosion of digital data to deliver reporting and analytics as a well-defined and that grows each day. "BI in the Cloud" architecture is only well-featured layer within its larger application. The going to be feasible when most of users source data lives in result is an analytic application that solves a customer the cloud already, possibly in something like SQL Server Data Services or Amazon Simple DB or Google BigTable; or problem with minimal customization and that is ideally possibly in a hosted app. like Salesforce.com. Cloud BI is the delivered using a software-as-a-service architecture on new way to do Business Intelligence instead of top of a cloud infrastructure. [4] implementing expensive and complex software on-site, the Also cloud computing represents several benefits to the BI BI software runs in the Cloud. It is accessible via any web such as: browser in a so-called software-as-a-service model. There is On-Demand: Immediately available with no infrastructure no need to install software, or to buy any hardware. And to deploy. when users are computing needs grow, the system will Elastic: Can scale up or down quickly with changing automatically assign more resources. This elastic scale is requirements. what makes Cloud BI so powerful users pay for what they Affordable: No large upfront costs, pay as user go. use as opposed to always paying to provision for peak load. Flexibility: to scale computing resources with few barriers. Starting at the back, the first objection raised to a purely BI Geographic scalability. [7] in the cloud architecture is that user has got to upload his Deploying BI in the cloud can help programs become more data to it somehow. Users can use the tools and applications flexible, scalable and agile. theyre familiar with to work from anywhere. With the It can be challenging to configure databases and BI tools to choice and flexibility of cloud computing, users businesses run in the cloud.[6] can deploy services on-premises, in the cloud, or a blend of Cheap Processing Power. The parallelization of cloud both. And, our solutions are all built on a unified computing makes Relational Online Analytical Processing. productivity platform thats not only cost-effective, but (ROLAP)-based analytics possible, since queries can be gives user the agility to respond as business needs evolve. spread across multiple CPUs simultaneously. Elastic Scale. BI is very sensitive to unpredictable and high Component of the proposed model (Business Intelligence peak loads. Users don t have to build for this, and can in the Cloud elastically scale to meet demand when it happens. [7] The proposed model represents a new environment Massive Multi-Tenancy. Users run a single instance of their atmosphere for the business intelligence to make the ability BI platform across 1000s of customers, meaning the marginal cost to provision, service and upgrade each user is to shorten BI implementation windows, reduced cost for BI extremely low. programs, Ability to add environments for testing, proof-of­ Service-Orientation. Given the transient nature of hardware concepts and upgrades .This environment represented as follows: nodes. [5]Additionally, Cloud computing provides key enablers for BI in 1- Cloud computingthese specific areas: Computational resources can be consumed for heavy The cloud provides a virtually unlimited pool of computing calculations that could be involved in predictive analytics as power, storage and memory. However, these resources are an example. delivered in discrete modules. Each node consists of Extremely heavy data loads could be stored for cheaper "standard" units of processing power, storage space and prices in storage resources in the cloud. [7] memory. While the amounts may vary (by service provider, Reporting and visualization are naturally fit for offering in a price point, etc.) and they may increase over time, the software as a service (SaaS) model, this would enable newer Cloud s pool of resources is a large grid of interchangeable, Consumption behavior for these specific BI components, industry-standard computing resources. Achieving true also given that web is the most common delivery method scalability requires a database architecture that can fully maximize this pool of resources. 653
  5. 5. Fig. 2 Business Intelligence Proposed Model 3.1 Database Platform: Business intelligence can be moved to cloud using PlatformA shared nothing, massively parallel database architecture is as services. It is a cloud-based relational database serviceparticularly designed to take advantage of multiple units of built on database technologies. It provides a highlycomputing resources. In the proposed model we use available, scalable, multi-tenant database service hosted inMicrosoft windows azure and we describe How cloud the cloud. Platform as a services helps to ease provisioningcomputing can be affected on the business intelligence? and deployment of multiple databases. Users do not have to 2. AppFabric: install setup and patch or manage any software. HighHelps connect applications and services in the cloud or on­ availability and fault tolerance is built-in and no physicalpremise, for example applications running on Windows administration is Database Platform of their existing on­Azure, Windows Server and a number of other platforms premises databases. Database Platform delivers scale toincluding Java, Ruby, PHP and others. meet the needs of the entire organization and provides IT 3. Platform as services: with flexibility to respond quickly to the evolving needs ofPlatform as services is a cloud services operating system the business. Organizations can bring great new experiencesthat serves as the development, service hosting and service and empowerment to their end users on a familiar ITmanagement environment. It is a flexible cloud-computing infrastructure thats more manageable and cost effective.offering that lets user focus on solving business problems Database Platform includes features to help user manageand addressing customer needs. No need to invest upfront critical data assets company-wide and across diverseon expensive infrastructure. Pay only for what user use, systems, helping to ensure integrity of information. Mostlyscale up when he need capacity and pull it back when he master data is maintained into the permanent stagingdon t. Users handle all the patches and maintenance all in database and synchronized with delta from extracts and thisa secure environment. Cloud computing as a platform piece of data is not that huge too. With a proper design andsupports multiple languages and integrates with existing on­ dissecting permanent staging area into two parts, by movingpremises environment. In addition, it supports popular master tables to the cloud and delta records to a temporarystandards, protocols and languages. staging area, the intermediate need of a staging server can be eliminated. Database Platform is a fit for this, as we just 651
  6. 6. need to store the staging data and queries are not that influences the way business intelligence software projectscomplex. And just for storing of this staging data, we do not are managed which it provide a virtually unlimited pool ofneed an enterprise level database and access to this data is computing power, storage space and memory for thenot that frequent too. business intelligence infrastructure, so our proposed model2. BI infrastructure(consists of three layers) represents a new environment atmosphere for the businessThe Data Layer: The Data Layer is responsible for storing intelligence that help in shortening BI implementation windows, reduced ion of cost for BI programs, enabling tostructured and unstructured data for management support. add environments for testing, proof-of-concepts andRegarding structured data, the central component is the data upgrades. Business Intelligence in the cloud has beenwarehouse (DWH). A DWH is commonly defined as a developed in order to enhance the efficiency and"subject-oriented, integrated, time-variant, and non volatile productivity of business intelligence and increase thecollection of data in support of management s decision­ performance of BI software. In the future we are aiming tomaking process". Many current realizations of DWHs are develop business intelligence by using web 3.0 technologies.based on so called Core DWHs .Core DWHs are usually not In the future we are aiming to develop business intelligenceused as a direct source for analysis systems, but rather by using web 3.0 technologies.distribute data to individual Data Marts. Data Marts keepexcerpts of application specific data. More recently; there REFERENCEShas been a shift towards DWH infrastructures that are [I] Pocatilu, P., F. Alecu, et al. "Measuring the EfIiciency of Cloudintegrated with operational systems. This is usually Computing for E-Ieaming Systems", Romania IJanuary 2010.achieved by the introduction of an Operational Data Store [2] Baars, Henning and Kemper, et ai, "Business Intelligence in the Cloud?",(2010). PACIS 2010 Proceedings.(ODS) that is designed to keep real time data on a [3] Withee, Microsoft Business Intelligence for Dummies, Wileytransactional level for time critical tasks. ODS/DWH Publishing, Inc.architectures allow to build Closed-loop and Active Data [4] Relan, p.andSibbingz, "Business intelligence for the cloud softwareWarehousing solutions. To feed the various data storages, stack helps gaming sites quickly adapt to evolving customer needs",ETL (Extract-Transform-Load) tools are needed. An ETL 2009.tool supports the extraction and transformation of data from [5] Dialogic Corporation. "Introduction to cloud computing", whiteheterogeneous source systems. The transformation includes paper.filtering out syntactical and semantic errors, harmonizing [6] JCios, K, W Pedrycz, et ai, "Data Mining A Knowledge Discovery Approach", 2007, Springer.data from different sources, as well as aggregating and [7] www.wikipedia.comenriching it. For the storage and administration of [8] www.microsoft.comunstructured data, Content Management Systems (CMS) [9] http: //www.information-management.comand Document Management Systems (DMS) are inserted [10] www.thebusinessintelligenceguide.cominto the data layer. [II] http: //www.ctoedge.comThe Logic Layer: The Logic Layer provides functionality to [12] www.thebusinessintelligenceguide.comanalyze structured data or unstructured content and supports [13] http: //www.slideshare.netthe distribution of relevant knowledge among different [14] http: //technet.microsoft.comusers. The most salient tools in BI environments are [15] http/www.salesforce.comreporting, data mining, and OLAF tools: Reporting tools [16] http: //aws.amazon.com, Amazon Elastic Compute Cloudpresent quantitative data in a report-oriented format that (Amazon EC2).might include numbers, charts, or business graphics. OLAP [17] http: //www.gc-solutions.net. Cloud Computing & Howdenotes a concept for interactive and multidimensional It Impacts e-Learning June 24th, 2010 by Gupta, M.analysis of aggregated quantitative business facts. Data [18] http://code.google.com, Google App Enginemining tools support the identification of hidden patterns in [19] http: //content.dell.com, Dell and Cloud Computing.large volumes of structured data based on statistical [20] https: //www.microsoft.com. Windows Azure Platfonnmethods like aSSOCIatIon analysis, classification, or [21] http: //www.ibm.comclustering. Data mining and similar model based tools arealso referred to by the term Advanced Analytics.The Access Layer: The Access Layer allows the user toconveniently use all relevant functions of the Logic Layer inan integrated fashion within the confines of defined userroles and user rights. VII. CONCLOSION AND FUTURE WORK The development of business Intelligence field cannotignore the cloud computing trends. There are many benefitsfrom using the cloud computing for business intelligence. It 655

×