BI4MASSES · Business Intelligence for the Masses

Business Intelligence for the Masses

The objective of Business Intelligence for the Masses (BI4MASSES) project is to build practical and theoretical know-how on the design and implementation of Cloud Computing (CC)-based large-scale data processing and analytics technologies that have a high potential for wide-scale shared use and social impact. Whether public or private, institutes that can process big data faster and turn it into useful information quickly become global leaders in their field. There may be hundreds of supercomputers and millions of personal computers in the world, but the analysis of both stored and streaming data keeps lagging due to the lack of online, easy-to-use, and semi-automated data processing and analytics software tools. It is crucial for Europe to boost investments in novel CC-based Knowledge and Data Engineering technologies just like USA, Russia, China and others do. Respectively, BI4MASSES project had a leading role in the dissemination of cloud computing and big data processing concepts as well as research-development and commercialization efforts in Turkey. The proposed project consisted of four major phases each taking about 1 year. PHASE1 was the design and implementation of a scalable distributed cloud infrastructure. The computing infrastructure was established inside the data center of Ozyegin University and several distributed systems software such as Hadoop and MPI were installed on top of the infrastructure as big data processing and high-performance computing (HPC) platforms. PHASE2 consisted of the design and implementation of the data mining and reporting service over the established platforms. Several open-source tools including Weka and Mahout were tested in this phase as candidate platforms. These studies and prototypes led to several publications, but the developed prototypes could not be scaled to support millions of users as intended inside the university, since this kind of support needed a sustainable business model. A startup company called Havooz was established by the P.I. and one of his M.S. students. The company is still serving in several big telecom and oil - gas companies in Turkey as well as their customers measured in millions. Therefore, we can state that the main goal of the project to reach masses has also been achieved through technology transfer. In PHASE3, the researcher project focused on developing the real-time data stream and complex event processing (CEP) service. This phase was also successfully completed with several critical publications and a CEP prototype with a high potential for productization. Specifically, the developed CEP engine has real-time rule mining, data validation and spatio-temporal indexing capabilities. Finally, in PHASE4 the researcher aim was to will develop and demonstrate the intelligent applications using the platform services developed so far. Big data processing and stream mining applications were developed and delivered to two telecom companies, one bank, and one oil and gas company. The mobile telecom applications consisted of analyzing wireless access protocol (WAP) logs and finding top visited URLs and network log analyses for failure reasons. The banking application was for stock portfolio analysis. The oil and gas sector application was for real-time sensor data validation and reconciliation. Overall 5 M.S. thesis were completed under the supervision of the P.I. at Ozyegin University with full or partial support from this project. 3 Ph.D. students are still continuing their studies. Several other researchers also directly or indirectly benefited from the grant through collaboration with the P.I. About 20 publications, several invited talks, and media appearances were made during the project for general dissemination of the results obtained. Cloud Computing Research Group (CCRG) For more information please visit: • http://faculty.ozyegin.edu.tr/ismailari • http://cloud.ozyegin.edu.tr/ Earlier version of this summary: http://cordis.europa.eu/projects/rcn/95421_en.html

Текст от CORDIS, на английски · Данни: CORDIS, © Европейски съюз