BUSINESS INTELLIGENCE AND CLOUD COMPUTING

Martin Demker 1 Buzzword Cloud Computing 2 Attributes of Cloud Computing 3 Layers of Cloud Computing 4 Cloud Infrastructure 5 Risks of Cloud Computing (1) 6 Risks of Cloud Computing (2) 7 Benefits of Cloud Computing for BI 8 Impact of Cloud Computing on BI 9 BI in the Cloud Today 10 The Future of BI in the Cloud

1 BUZZWORD CLOUD COMPUTING

Cloud Computing is a relatively recent term The technology itself is nothing new Used as a buzzword to generate hype No common definition, distinction to grid computing and virtualization not clear to everyone Hype not over yet according to Gartner
Martin Demker

Image 1 [trnd]

References: [dck]

Image 2 [grtn1]

Page 2/12

2 ATTRIBUTES OF CLOUD COMPUTING

Service-based: consumer concerns abstracted from provider concerns; interfaces hide the implementation details; "ready to use" and "off the shelf Scalable and Elastic: service scale capacity up or down to meet demand Shared: services share a pool or resources to optimize economies of scale; better resource utilization Metered by use: plans based on the amount of service used by the consumers Uses Internet Technologies: based on existing standards (URLs, HTTP, IP, )
References: [grtn2]

Martin Demker

Page 3/12

3 LAYERS OF CLOUD COMPUTING

Application Plattform Infrastructure

Software-as-a-Service (SaaS): customer gets access to ready-made software, everything is managed by the cloud provider (e.g. Salesforce, Gmail, ) Application-Components-as-a-Service (A-CaaS): personalized PaaS components which can be integrated by the user into higher-level applications in the SaaS layer

Platform-as-a-Service (PaaS): customer gets access to a solution platform, develops and manages own application on top of that platform (e.g. Force.com, Google App Engine, Windows Azure, )

Infrastructure-as-a-Service (IaaS): customer gets access to a virtualized hardware system, manages software himself (e.g. Amazon EC2, Amazon S3)

Based on many loosely coupled and distributed computers built with commodity hardware Software used to achieve parallelization and avoid hardware failures

References: [ccj]

Page 4/12

4 CLOUD INFRASTRUCTURE

Infrastructure is built using cheap hardware

Hardware failures will occur Bandwidth of one single node is not very high

Virtualization used to hide the hardware implementation Data-parallel programming model required to allow scalability and fault-tolerance Well known solutions: Google MapReduce / Apache Hadoop

Distributed file system ensures fault-tolerance Processing jobs are split into map and reduce tasks, which can be executed in parallel on different nodes, if a task fails or takes too long it is repeated on another node Used at: Google, Facebook, Amazon, Yahoo!,
References: [brkl]

Martin Demker

Page 5/12

5 RISKS OF CLOUD COMPUTING (1)

Authentication / Privileged user access: danger of outside party access to data Regulatory compliance: requirements for privacy and security of sensitive data Data location: data distributed beyond control, country / jurisdiction not clearly defined Data segregation: data is stored alongside data from other customers, has to be separated and protected from them (e.g. through encryption)

References: [grtn3] [itex]

Martin Demker

Page 6/12

6 RISKS OF CLOUD COMPUTING (2)

Integration with internal systems: some (strategic) systems cant be outsourced, have to interface with the cloud; may lead to huge data transfers System availability: no control over the cloud, uncertainty about availability of system Long-term viability / business continuity: cloud vendor could go bankrupt or be acquired Recovery: restoration ability of cloud provider unclear Investigative support: maybe impossible to investigate errors or illegal activity in the application due to system architecture
References: [grtn3] [itex]

Martin Demker

Page 7/12

7 BENEFITS OF CLOUD COMPUTING FOR BI

Lower costs & usage billing: no huge up-front investment for hardware and software needed; pay only as much as you use Fast deployment, low maintenance: cloud BI platforms can be up and running in a matter of minutes; maintenance is done by the cloud provider Scalable provisioning of resources: BI applications show different work load over time (e.g. analytics during the day, ETL during the night); the cloud dynamically allocates resources where they are needed On-demand resource improvements: when DW capabilities need to be expanded, this can be done without interrupting daily operations References: [mit]
Martin Demker Page 8/12

8 IMPACT OF CLOUD COMPUTING ON BI

Easier evaluation of new technology: allowing faster BI technology adoption Increased short-term ad-hoc analysis: when short term needs arise, they can be easily implemented in the cloud and be canceled after the necessary period (no leftover HW/SW) Increased flexibility: no long-term commitments individual business units can fund more data mart projects Growth considerations: data volume increases database architecture required that can handle large volumes and the changing analytics workloads and still provide high availability; cloud architecture is well suited for that
References: [ijcte]

Martin Demker

Page 9/12

9 BI IN THE CLOUD TODAY

Megavendors (SAP, Oracle, MS, IBM) continue to control majority of BI market share/revenue All of them working on SaaS offerings; product maturity not very high yet; development focus on hybrid solutions (part of the data in traditional systems, part in the cloud) Many emerging vendors trying to compete with diverse product range Customer BI platform SaaS adoption currently very low, but rising

References: [grtn4]

Martin Demker

Page 10/12

10 THE FUTURE OF BI IN THE CLOUD

Cloud BI expected to grow in the coming years Factors that might lead to faster cloud BI adoption:

Rapid growth of data volume; incorporation of diverse data from outside of the corporate data sources (especially social networks) Ease of use becoming more important; intuitive and fun user interfaces required; business users want to explore data without relying on IT (data mashup) Mobile availability of BI solutions; access to data from outside of the company using smartphones and tablets

Cloud Computing and Business Intelligence consistently highranked on Gartners list of top strategic technologies for the last few years; trend likely to continue
References: [ijcte] [grtn4]

Martin Demker

Page 11/12

REFERENCES
[trnd] http://www.google.de/trends/?q=cloud+computing,+grid+computing,+virtualization&ctab=0&geo=all&date=all&sort=0, Last checked: 31.01.2012 [dck] http://www.datacenterknowledge.com/archives/2008/03/25/whats-in-a-name-utility-vs-cloud-vs-grid/, Last checked: 31.01.2012 [grtn1] Gartner, Hype Cycle for Cloud Computing, Stamford, July 2011 [grtn2] Gartner, Gartner Highlights Five Attributes of Cloud Computing, Stamford, 2009 [grtn3] http://www.networkworld.com/news/2008/070208-cloud.html, Last checked: 31.01.2012 [grtn4] Gartner, Magic Quadrant for Business Intelligence Platforms, Stamford, January 2011 [ccj] http://cloudcomputing.sys-con.com/node/1200642, Last checked: 31.01.2012 [itex] http://www.infotex.com/portal_blog/white_papers/risk_landscape_of_cloud_computing_isaca.pdf, Last checked: 31.01.2012 [mit] E. Reyes, A systems Thinking Approach to Business Intelligence Solutions Based on Cloud Computing, Cambridge, June 2010 [brkl] M. Zaharia, Cloud Computing with MapReduce and Hadoop, Berkley, 2010 [ijcte] Ouf S., The Cloud Computing: The Future of BI in the Cloud, International Journal of Computer Theory and Engineering, Vol. 3, No. 6, December 2011

Martin Demker

Page 12/12