Oracle OpenWorld 2016. Big Data references
0 likes349 views
The document outlines an agenda for a customer panel session on data warehousing and big data hosted by Oracle. The agenda includes presentations from five customer panelists on their data and analytics architectures, followed by a question and answer session. Each panelist presentation provides details on the customer's business and technical needs, data environments, transformation efforts, and analytics use cases.
Oracle OpenWorld 2016. Big Data references
- 10. Copyright © 2016, Oracle and/or its affiliates. All rights reserved. | Functional Use • Landing Zone for Raw Data • Data Transformation to RDBMS • Long Term Data Archival • Data Science Discovery & Advanced Analytics Primary Technologies Used • HDFS, HIVE, Pig, Impala, Parquet, Sqoop, Oozie, HUE, Spark, Cloudera Manager • R, Python, Scala • Oracle Big Data Connectors Oracle - Big Data Appliance
- 11. Copyright © 2016, Oracle and/or its affiliates. All rights reserved. | 6 Big Data Key Learnings 1. Deliver quick wins for the business 2. Prototype! There are 5 ways to do one thing 3. Fail fast and move on 4. Think about governance & security early 5. Not a silver bullet, Bad Data is sRll Bad Data 6. Talent is criRcal for success
- 26. Copyright © 2016, Oracle and/or its affiliates. All rights reserved. | Pre TransformaRon AnalyRcal Architecture Oracle ConfidenRal – Internal/Restricted/Highly Restricted 28 C1 NBVKinetiq B1MDMS SAP ERP Market N M M M MM F C&I C&I n/w payment (manual) accruals C C&I C C&I SAP BW F Operational Analytical RDW Intellect HDS CIDS EMDMEDW RGMSM CBF C C C&I M C B F M M C C&I C C&I F C B C&I BF Quote Digital NMI discovery M Bill payment history B C C&I C B enriched - Aspect - CentrePoint - Infra S B B BB PMDS MADB & PFDB PARS SAP BPC HRIS Maximo F F F F M W Info- server W W W Campaign Tool C
- 27. Copyright © 2016, Oracle and/or its affiliates. All rights reserved. | TransformaRon EnergyAustralia’s InformaRon Management (IM) Augmented Data Warehouse Architecture Source Store Structured Data Big Data Store (Distributed File Clusters) Real-timeEventCapture Master Data Management Data Quality Metadata Master Data Store Meter Data (Usage) SDS SDS SDS Customer Analytical datasets Feature vectors Meter Market Other SDS . . Consume Traditional Business Intelligence Data Discovery Self Service Legal / Compliance Descriptive Analytics Management Reports Pattern Detection Data Visualisation Downstream Feeds Operational Systems Legacy Warehouses Other Financial Reporting and Analysis Risk Management Financial planning and budgeting Financial modelling and forecasts Operational Analytics Channels Web Multichannel Campaign Management Real Time Decision Engine Call Centre ShopEmail Digital Messaging Platform SMS/ MMS Mobile Apps 3rd Party Interfaces Other Advanced Analytics Prescriptive Analytics Descriptive Analytics Predictive Analytics Procurement Performance Management Ingest Structured C1 B1 Kinetiq External Retail Meter Data Wholesale Meter Data Wholesale Market Data 3rd Party Data Semi-structured EA Websites EA Mobile Apps Social Media EA Call Center Discovery Data Store (Data Labs) Big Data Operational Data Store Analytical Data StoreSource Data Store Operational Systems …etc BI Governance (acquisiRon, integraRon, transformaRon, needs-aligned delivery, innovaRon) IM Governance (end-to-end, mulR-domain, strategic alignment, enterprise context) Governance FoundaRon Technical FoundaRon Business AnalyRcal CapabiliRes Key Accelerate technical foundaRon development Establish informaRon governance Setup Customer master store TransiRon Meter data to future state Establish Source Data Stores Establish data discovery for advanced analyRcs using the Hadoop Ecosystem Develop new business capabiliRes • Customer and Meter AnalyRcs • ProducRvity and Cost AnalyRcs • Energy Asset management Migrate exisRng data stores Data quality – customer and meter
- 29. Copyright © 2016, Oracle and/or its affiliates. All rights reserved. | Technical SpecificaRons 2 x (¼ + ⅛) Rack X5-2 & X6-2 (8TB spindle) Exadata (Primary + Secondary) - PCI Flash: 19.2TB - Disk Capacity: 54TB - Cores: 18 - Memory: 512GB - 2000 Staff connecRons and 2.5 Million Customer connecRons - 18 Billion Records + 2 x Starter Rack X5-2 Big Data Appliance - 6 x Compute Nodes
- 31. Copyright © 2016, Oracle and/or its affiliates. All rights reserved. | • 179 Years old this year • Countries of operation: ~70 • Countries where are brands are sold: ~180 • Consumers served by our brands: 5 billion (approximate) • Last FY Revenue: $70.7 billion About Procter and Gamble… Business Description: The Procter & Gamble Company (P&G) is a leading manufacturer and marketer of branded consumer packaged goods across the world. P&G manufactures, markets and distributes premium quality, branded beauty, health, fabric, home, baby, family, and personal care products.
- 32. Copyright © 2016, Oracle and/or its affiliates. All rights reserved. | Source & Data ApplicaRon Data AcquisiRon Base Data Storage AnalyRcs Reports & Conclusion Data Data InformaRon Knowledge Wisdom Retailer Data Shopper Data Market Data & more Internal Data External Data Big Universe On PREM Big Universe Exernal Cloud Domain Expert Analyst End User
- 33. Copyright © 2016, Oracle and/or its affiliates. All rights reserved. | OpenSource 3 PLATFORM LINUX OPERATING SYSTEM CLOUDERA MID LAYER CUSTOM OPEN SOURCE EDGE NODES CLUSTER e.g…. DATA SOURCES DATA SOURCES DW OR DW SAP EXT CLUSTERs CLUSTER server for storage and processing ORACLE ORACLE ORACLE EDGE (UTILITY) NODES ON PREM CLOUD A CLOUD B Dedicated server to access cluster services BI SERVER Dedicated server for BI applicaRons CLOUD A CLOUD B ON PREM BI TOOL LAYER
- 34. Copyright © 2016, Oracle and/or its affiliates. All rights reserved. | AnalyRcal Use Cases • Selling Opera8ons: IntegraRng store, product & consumer data to develop models & simulaRons which generate revenue opportuniRes in assortment, pricing*, promoRon and shopping behavior. • Consumer Engagement: OpRmizing consumer engagement, purchase intent and conversion by modeling and simulaRng their on and off-line behavior driving up and cross selling opportuniRes. • Supply Chain Opera8ons: Transforming our supply network, and operaRons through big data analyRcs to save inventory costs, idenRfy drivers of missed cases and increase overall agility and responsiveness. • Business Management: ForecasRng results of new iniRaRves days into launch, assessing our ability win target consumers, and driving a significant increase in trade and media efficiency. * pricing is at the sole discreRon of retailers