Gradiant - Technology Offer in Business Analytics

Gradiant - Technology Offer in Business Analytics

• 1.
  Research in BusinessAnalytics February 2013 Contacts: Arnaud Quirin <[email protected]> Héctor Cerezo <[email protected]> GALICIAN RESEARCH AND DEVELOPMENT CENTER IN ADVANCED TELECOMMUNICATIONS
• 2.
  Business Analytics in Gradiant:an Overview Key technologies • Data/Graph mining • Big Data / Business Intelligence • Data network characterization/analysis HW and SW Resources • HPC resources (304 cores, 1.22 TBs RAM, including 8 virtualization and 2 GPU servers) Project participation • National projects: AIBOT, CELTIC, SmartED • International projects: LIFTGATE IP & Technology Transfer • Real-time data compression algorithm for in-memory databases (US pat. pend. 61566689) • Scalable analytical frameworks (in-memory databases, Hadoop, Cascading, HBase) 25+ publications in international journals, conferences, ... CENTRO TECNOLÓXICO DE TELECOMUNICACIÓNS DE GALICIA
• 3.
  Expertise in Data/Graph Mining Data/Graph mining • This technology achieves useful insights in large and complex data collections through the combination of pre-processing, transformation, modeling and analysis techniques. Application cases in Gradiant • Satellite image classification of marine seaweeds for ecological conservation, prediction of wind farms electrical supply, prediction and simulation models in emergency departments. • Discovering the relationships among scientific and industrial data for technological surveys; user profiling, relationship and influence mining using Social Media content. http://www.mapofscience. com/ CENTRO TECNOLÓXICO DE TELECOMUNICACIÓNS DE GALICIA
• 4.
  Expertise in ScalableBusiness Intelligence Big Data/NoSQL • Big Data solutions address the problem of analyzing large datasets with efficient scalable frameworks. This includes distributed NoSQL databases. Business Intelligence (BI) • BI solutions address the problem of processing past records of digital companies to achieve insightful analytics. • Gradiant has experience in main memory analytics, as well as in technologies to visualize interactive analytics of aggregated business-generated content. Application cases in Gradiant • Gradiant participates in projects requiring fast-processing of Social Media content (~4 million records/day). Storage in HBase and analysis with Hadoop and Cascading. • Gradiant owns in-memory algorithms for analytical databases (US pat. pend. 61566689), comparable to QlikView solutions. CENTRO TECNOLÓXICO DE TELECOMUNICACIÓNS DE GALICIA
• 5.
  Expertise in DataNetwork Analysis Data network characterization/analysis • Data network analytic solutions combining monitoring and characterization of network operator traffic using simulation and mathematical models. • Gradiant has experience in projects with Vodafone. Application cases in Gradiant • Monitoring and traffic characterization of xDSL- connected users in operator networks, leased link performance analysis and dimensioning. • Detection of saturation points in leased links using indirect measures. • Large Network simulation models (experience with ns3 and OMNeT++). • Visualization of statistical data (d3.js and processing.js). CENTRO TECNOLÓXICO DE TELECOMUNICACIÓNS DE GALICIA
• 6.
  HW and SWResources Hardware • Storage (in cabinet), total 62 TBs - HP M6612 2TB 6G SAS 7.2K 3.5in MDL HDD x 24 - HP M6625 600GB 6G SAS 10K 2.5in HDD x 24 • Servers, total 304 cores and 1.22 TBs RAM - 1 x HP DL980 G7 Intel E7 2860 (80 cores, 512 GBs RAM) - 2 x HP SL390 G Intel X5675 (24 cores, 192 GBs) - 4 x HP DL380G7 Intel Performance (48 cores, 192 GBs) - 2 x HP DL380G7 Intel Efficiency (24 cores, 96 GBs) - 2 x HP DL385G7 AMD Performance (64 cores, 128 GBs) - 2 x HP DL385G7 AMD Efficiency (64 cores, 128GBs) • Cryptoprocessors - 2 x Utimaco SE10 PCIe boards Software • Big Data platform Hadoop/Hbase/Cascading CENTRO TECNOLÓXICO DE TELECOMUNICACIÓNS DE GALICIA
• 7.
  Current Project Cases(I) CELTIC: Strategical knowledge gathered from business intelligence technologies • Spanish consortium, supported by Innterconecta tech funds. • Under contract, 2012-2014 • Partners: INDRA (leader), Elogia, SaecData, Imaxin Software BA Goals • User profiling for marketing decision support, design of customer influence metrics relying on scalable efficient frameworks. CENTRO TECNOLÓXICO DE TELECOMUNICACIÓNS DE GALICIA
• 8.
  Current Project Cases(II) SmartED: Intelligent management for emergency health services • Spanish consortium, supported by CDTI tech fund. • Under contract, 2011-2013 • Partners: Everis (leader), Arantia, Balidea BA Goals • Simulation of hospital emergency departments, neural net-based prediction models of patient arrivals. CENTRO TECNOLÓXICO DE TELECOMUNICACIÓNS DE GALICIA
• 9.
  Current Project Cases(III) AIBOT: Framework for the integration of BSS/OSS systems designed to market non- telecommunication services through telecommunication operators • Spanish consortium, supported by Innterconecta tech funds • Under contract, 2012-2015 • Partners: Sivsa (leader), Discalis, Optare BA Goals • Root Case Analysis (RCA) algorithms in the cloud to track incidence causes in complex telco systems CENTRO TECNOLÓXICO DE TELECOMUNICACIÓNS DE GALICIA
• 10.
  IP & TechnologyTransfer Patent • Real-time data compression algorithm for in-memory databases (US pat. pend. 61566689) Non patented IP • Distributed framework to process large amounts of data • Scalable framework for user profiling and relationship analysis of social media data • Prediction algorithms for hospital emergencies • Indirect techniques to predict saturation in leased network lines • Root Case Analysis (RCA) prediction algorithms to track incidences in complex systems CENTRO TECNOLÓXICO DE TELECOMUNICACIÓNS DE GALICIA
• 11.
  Dissemination • P. Sendin-Raña, E. Rodriguez-Fernandez, F.J. Gonzalez-Castaño et al, “Web-oriented business intelligence solution based on Associative Query Logic”, Software, Practice and Experience, 40(9): 779-796 (2010). • P. Sendín Raña, F. J. González Castaño, E. Pérez Barros, P. S. Rodríguez Hernández, F. Gil Castiñeira, J, M. Pousada Carballo, "Improving the performance and functionality of Mondrian open-source OLAP systems," Software, Practice & Experience, 39(3): 279-298 (2009). • A. Quirin, O. Cordon, V. P. Guerrero-Bote, B. Vargas-Quesada, F. Moya-Anegon; A Quick MST-based Algorithm to Obtain Pathfinder Networks; Journal of the American Society for Information Science and Technology, 59(12): 1912-1924 (2008). • A. Quirin, O. Cordon, B. Vargas-Quesada, F. de Moya-Anegon; Graph-based Data Mining: A New Tool for the Analysis and Comparison of Scientific Domains Represented as Scientograms; Informetrics, 4(3): 291-312 (2010). • E. Serrano, A. Quirin, J. Botia, O. Cordon; Debugging Complex Software Systems by Means of Pathfinder Networks; Information Sciences, 180(5): 561-583 (2010). CENTRO TECNOLÓXICO DE TELECOMUNICACIÓNS DE GALICIA