Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
What’s all the fuss about “Big Data”? Doug Cackett Oracle Enterprise Architecture [email protected] Copyright © 2014, Oracle and/or its affiliates. All rights reserved. | Adapted from Tom Davenport Delivering business benefits Analytics 3.0 Analytics 2.0 • Platform for monetisation Analytics 1.0 • “Competing on Analytics” • Basic reporting • Limited range of tabular data • Batch oriented analysis • Analysis bolted onto limited set of business processes Data as a business cost • Extended analytics to larger and less structured datasets • Fact and exception based management doctrine • Recognition of Data Science • • • • Deeper analysis on more data Faster test-do-learn wider business process coverage Analysts focus on discovery and driving business value Data as a business benefit Copyright © 2014, Oracle and/or its affiliates. All rights reserved. Momentum around Information Management Data scope & operational linkage Changing technology footprint • Blurring of operational and analytical boundaries • Data volumes driving costs and surfacing performance issues • Externally sourced data mashups driven by business users • Polyglot persistence model & Hadoop market momentum • Rationalising existing investments that often date back many years • Expansion in range of possible architecture patterns Delivery and alignment strategy Governance & organisation • Recognition of Bi-Modal delivery pattern for alignment • Shift in power and budgets towards Business • Shift from “truth” to “trust” and the implications on data quality • The critical need for rational Analytical governance • Efficient, everyday, curated • Rapid discovery of business insights • Agile development approach that links the modes together Gartner 2015 BI Summit Copyright © 2014, Oracle and/or its affiliates. All rights reserved. Discovery is a key capability Business Value • Architecture, approach and governance aim to: Commercial Exploitation –Accelerate time to discovery with lower manual effort –Reduce the size of the Governance step –Eliminate growth in technical and architectural debt Governance Understanding of the data Discovery phase Time / Effort Copyright © 2014, Oracle and/or its affiliates. All rights reserved. Innovation through Data Discovery process It’s about “Data Discovery” not IT development What IT thinks it should be Requirement Analysis High Level Design Low Level Design Coding Testing Acceptance Testing Local Access Database Shared Access Server SQL Server Database Oracle Datawarehouse What normally happens Excel Spreadsheet Shared linked spreadsheets What we need to achieve in the Discovery process Discovery & Profile Model Copyright © 2014, Oracle and/or its affiliates. All rights reserved. | Exploit Conceptual architecture for Information Management “At the Internet age, we can not confront the challenges of the future by doing the things as was thought 20 years ago” CaixaBank executive statement for Inspiring the “data pool” and User Experience Initiatives Copyright © 2014, Oracle and/or its affiliates. All rights reserved. | Initial customer deployment patterns Discovery Lab Re-conceptualise IM 1 Operationalise Insights 5 2 Budget Use-cases Sponsorship Timeline Skills 4 3 Big Data Application Big Data technology pilot Copyright © 2014, Oracle and/or its affiliates. All rights reserved. Oracle’s Information Management Reference Architecture Full lifecycle that also balances ingestion and interpretation costs Copyright © 2014, Oracle and/or its affiliates. All rights reserved. | Balancing solution demands and the persistence model Identifying the right technology and design pattern Straight Through Processing (STP) Tooling maturity 5 4 Availability & Business Continuity 3 Ingestion rate 2 1 ACID transactional requirement In-Memory Set or row loading into database or memory Relational Hadoop 0 Cost effectively store low value data Hadoop Relational Security Ingestion (coding) simplicity Variety of data formats Data sparsity API or utility loading into Non-relational stores Access NoSQL Other stores… Copyright © 2014, Oracle and/or its affiliates. All rights reserved. | 9 Big Data SQL – an aggregate approach ETL of aggregate data moved to RDBMS 10’s of Gigabytes of Data WEB_LOGS B B B CUSTOMERS Hadoop Cluster Oracle Database Copyright © 2014, Oracle and/or its affiliates. All rights reserved. | 10 Big Data SQL – a federated approach All the data copied to RDBMS for query processing 10’s of Gigabytes of Data WEB_LOGS B B B CUSTOMERS Hadoop Cluster Oracle Database Copyright © 2014, Oracle and/or its affiliates. All rights reserved. | 11 Big Data SQL – a franchised approach SELECT w.sess_id, c.name FROM web_logs w, customers c WHERE w.source_country = ‘Brazil’ AND w.cust_id = c.customer_id; Relevant SQL runs on BDA nodes Big Data SQL 10’s of Gigabytes of Data WEB_LOGS B B B Only columns and rows needed to answer query are returned Hadoop Cluster CUSTOMERS Oracle Database Copyright © 2014, Oracle and/or its affiliates. All rights reserved. | 12 Apache Zoo Families & Species (Solution Classes & Products) Keeping pace with the rapidly evolving Apache Zoo Non-Persistent Data Stores Data Capture Data & Integration Delivery Data Processing Persistent Data Stores Data Discovery Search Workflow & Orchestration Advanced Analytics Data Query Other Data Movement (Relational) Copyright © 2014, Oracle and/or its affiliates. All rights reserved. Finding out more about the Reference Architecture • Download the architecture white paper here • Attend an Information Masterclass event –Full day workshop with other customers or Partners –Delivered by architect for architect –Whiteboard not PPT –Product agnostic Customer event details here (Oct 15th) Partner event details here ( Nov 24th) Copyright © 2014, Oracle and/or its affiliates. All rights reserved. | 14