BIG DATA & ANALYTICS Because the speed of business is money Héctor Colmenares, IBM SW Core Database Competitive Sales Leader, IM SPGI hector.colmenares@es.ibm.com Tendencias del mercado Demanda de respuestas en real time es una necesidad. • Vivimos en la generación del “NOW”. • Los datos crecen constantemente, y los usuarios esperan respuestas más rápidas. • Movilidad y el uso “democrático” de la información y la analítica hacen de la tecnologia in-memory y mensajería distribuida un requerimiento obligado. La velocidad cambia el negocio. • Nuevas y modernas arquitecturas de datawarehouse dinámicas reemplazarán los modelos tradicionales de datos por la demanda de datos en real-time. 2 © 2014 IBM Corporation Big Data y analítica del negocio La Revolución de los datos 30 billion RFID 12+ TBs tags today (1.3B in 2005) of tweet data every day 4.6 billion camera phones world wide data every day ? TBs of IT Logs 80% De los datos mundiales NO ESTRUCTURADOS log data every day 76 million smart © 2014 IBM Corporation devices sold annually 2+ billion 25+ TBs of 3 100s of millions of GPS enabled meters in 2009… 200M by 2014 people on the Web by end 2011 Big Data: todo son datos Paradigma para extraer valor Transaccional & Datos Aplicativos 4 Contenido Empresarial Dato Social Data Sensores • Volumen • Variedad • Variedad • Velocidad • Estructurado • No estructurado • No estructurado • Estructurado • Entrada / Salida • Volumen • Veracidad • Ingestión © 2014 IBM Corporation Big Data: todo son datos Gestión de los datos: No es única E-commerce 5 Mobile Storefront Sales Analysis Demand Analysis Key 1 JSON doc 1 Meter 1 Data series 1 Key 2 JSON doc 2 Meter 2 Data series 2 Transactional Database JSON Database Analytics Data Warehouse Transaction Processing Mobile Data Serving Reporting and Analytics © 2014 IBM Corporation Real Time Fraud Detection Operational Data Warehouse Operational Analytics Time Series Database Sensor Data Analysis Analítica dá la clave para incrementar la competitividad Compañías que realizan analíticas sofisticadas superan a su competencia 2.6x 1.6x mas rendimiento que sus iguales del sector Mas ingresos 260% estar entre los mejores del sector 2.5x Valorización del precio del stock Source: The New Intelligent Enterprise, a joint MIT Sloan Management Review and IBM Institute of Business Value analytics research partnership. Copyright © Massachusetts Institute of Technology 2011. Outperforming in a data-rich, hyper-connected world, IBM Center for Applied Insights study conducted in cooperation with the Economist Intelligence Unit and the IBM Institute of Business Value. 2012 6 © 2014 IBM Corporation Cuantificar valor de la analítica Speed equals 39% faster payment* 59 reduced days to payment business growth days to payment increased cash flow 36 2011 2012 2013 *Based on analysis done by Xero, a SaaS company specialising in accounting software, 2014. Link to blog & infographic: HERE Average closing of accounts Source: SAP value engineering study 7 © 2014 IBM Corporation Speed of Business Process, Is Money La lógica del Data Warehouse Contemplar componentes, propósitos, zonas Vertical Industry Accelerators Advanced Application Capabilities Machine and sensor data Image and video Actionable insight Real-time processing & analytics Data types Operational systems Deep analytics & modeling Exploration, landing and archive Predictive analytics and modeling Trusted data Enterprise content Reporting & interactive analysis Transaction and application data Decision management Reporting, analysis, content analytics Social data Third-party data Logical Data Warehouse Information Integration & Governance 8 © 2014 IBM Corporation Discovery and exploration La lógica del Data Warehouse Ejemplo de la Solución de IBM Vertical Industry Accelerators BigSQL and SQL based applications Advanced Application Capabilities Federation and In-memory Real-time processing & analyticsfederated cube Data types Machine and sensor data Image and video Operational systems Deep analytics & modeling Exploration, landing and archive Trusted data ORACLE MicrosoftReporting & interactive Teradata analysis Enterprise content Transaction and application data Actionable insight Decision management Predictive analytics and modeling Reporting, analysis, content analytics Social data BigMatch Third-party data Logical Data Warehouse Streams 9 © 2014 IBM Corporation Information Integration & Governance Discovery and exploration ¿Qué hace diferente BLU Acceleration? Innovacionees de IBM Research & Developments Labs. Next Generation In-Memory Analyze Compressed Data In-memory columnar processing with dynamic movement of data from storage Patented compression technique that preserves order so data can be used without decompressing C1 C2 C3 C4 C5 C6 C7 C8 Encoded CPU Acceleration Data Skipping Multi-core and SIMD parallelism (Single Instruction Multiple Data) Instructions 10 © 2014 IBM Corporation Results Skips unnecessary processing of irrelevant data Data BLU Shadow Tables Dedicated analytics and reporting Operational analytics Mixed workload analytics with OLTP OLTP Indexes Analytical Indexes Traditional row-based tables, with indexes for, for tables dedicated to OLTP or Operational Analytics + Simple BLU tables ( columnar ) for tables dedicated to analytics and reporting workloads OLTP Indexes + + Single Server All 3 scenarios in a single database 11 © 2014 IBM Corporation Traditional row-based tables, with indexes and BLU Shadow Tables for tables with mixed workloads • Power of BLU • Faster analytics and reporting • Faster OLTP • Simpler environment Oportunidad: Big Data y Analytics 30 billion RFID tags today (1.3B in 2005) TBs 12+ PROBLEMAS DE RENDIMIENTO of tweet data every day data every day ? TBs of IT Logs SAP BW CARGAS ANALÍTICAS COMPETENCIA: 80% SAP con HANA ORACLE con EXADATA TERADATA Of world’s data is unstructured MS-SQL 25+ TBs of log data every day PREMISA: AHORRO DE COSTES 76 million smart 12 © 2014 IBM Corporation meters in 2009… 200M by 2014 4.6 billion camera phones world wide 100s of millions of GPS enabled devices sold annually 2+ billion people on the Web by end 2011 DB2 with BLU vs Microsoft SQLServer Query Response Time: (In Seconds, Less is Better) DB2 (extrapolated to 8 cores, 80% scalability) v/MS SQL (8 cores) 90 80 70 Query time (s) 60 50 40 30 20 10 0 1 2 3 4 5 6 7 8 9 Query Runs DB2 BLU Factored 13 © 2014 IBM Corporation MS SQL 7X-8X Better Performance with equal cores 10 DB2 with BLU vs Microsoft SQLServer Query Response Time: (In Seconds, Less is Better) BestOffer Less Cores & Licenses but Much more Performance => Better SLA DB2 BLU (2 cores) v/MS SQL (8 cores) 90 80 Query Time (s) 70 60 50 40 30 20 10 0 1 2 3 4 5 6 7 8 9 Query Runs DB2 BLU 14 © 2014 IBM Corporation MS SQL “56% better performance, with 25% of the cores ! Wow, that’s great !” 10 Estimated HW Infrastructure for Production – Year 1 and Year 5 assumption yearly 20% growth Source Oracle database 8 TB on BW 7.0 (non-unicode) DB2 on 2-tier architecture (on one server all components) HANA on 3-tier rachitecture (database and application on different servers) 15 15 © 2014 IBM Corporation Huge savings through DB2 Technology Often DB2 BLU needs 70-95% less HW COMPARATIVA ENTRE IBM POWER + DB2 contra ORACLE EXADATA NECESIDAD INICIAL: Sistema SAP (180 sistemas, 48 entornos de producción) CONTINUIDAD DEL NEGOCIO: Contingencia en 2 centros separados NOTA: • El ejercicio de sizing se ha basado en la metodología SAP con entornos para Producción, Pre-Producción y Desarrollo/Q • El nivel de rendimiento SAPS ha sido el mismo en ambos casos • La infraestructura IBM es POWER8 + AIX + DB2 10.5 y la de Oracle es EXADATA (INTEL`+ Oracle Linux + Oracle DB) • La opción de IBM permite virtualización • El ejercicio es una estimación y está orientada a mostrar las diferencias de infraestructura entre ambas soluciones Customer runs DB2 on POWER - 180 systems, 48 production - 26 HA (LPM*) + 26 DR (PowerHA) - 2 x data centers Possible Exadata implementation ** - 180 systems, 48 production - 26 HA + 26 DR clusters - 2 x BIGGER or more data centers 6 full racks for production + HA 6 full racks for DR 6 full racks for test/QA 6 full racks for dev 6 full racks for the rest 36 systems 4 x POWER servers (160 cores) ~30 full racks (5760 cores) 16 16 © 2014 IBM Corporation * LPM - AIX live partition mobility ** No virtualization + limited number of databases per rack (e.g. 8 database servers per full rack, max 24 processor per database) COMPARATIVA ENTRE IBM POWER + DB2 contra SAP HANA NECESIDAD INICIAL: Sistema SAP (180 sistemas, 48 entornos de producción) CONTINUIDAD DEL NEGOCIO: Contingencia en 2 centros separados NOTA: • • • • • • El ejercicio de sizing se ha basado en la metodología SAP con entornos para Producción, Pre-Producción y Desarrollo/Q El nivel de rendimiento SAPS ha sido el mismo en ambos casos La infraestructura IBM es POWER8 + AIX + DB2 10.5 y la de SAP HANA (INTEL+ Linux + HANA) La opción de IBM permite virtualización El ejercicio es una estimación y está orientada a mostrar las diferencias de infraestructura entre ambas soluciones 1 HANA UNIT = 64 Gb RAM = 13 K€ Customer runs DB2 on POWER - Customer runs DB2 on POWER - 180 systems, 48 production - 26 HA (LPM*) + 26 DR (PowerHA) 2 x data centers Possible HANA implementation ** - 180 systems, 48 production - 26 HA + 26 DR clusters - 2 x BIGGER or more data centers 48 appliance servers for production 52 appliance servers for HA+DR clusters up to 48 appliance servers for test/QA up to 48 appliance servers for dev up to 36 appliance servers for rest 4 x POWER servers (160 cores) 101-232 x HANA servers (4040-9400 cores) 17 17 © 2014 IBM Corporation * LPM - AIX live partition mobility ** No virtualization + limited number of databases per rack (e.g. 8 database servers per full rack, max 24 processor per database) Balluff GmbH – Game-changing boost to information delivery with IBM DB2 enables rapid insight 98 percent faster access to complex reports 30 percent typical report speed increase 50 percent faster SAP ERP response times Solution Components •SAP® Business Warehouse, SAP ERP, SAP ERP HCM, SAP CRM, SAP NetWeaver® Enterprise Portal, SAP PI •IBM® AIX®, DB2® for Linux, UNIX and Windows with BLU Acceleration, PowerHA® SystemMirror®, PowerVM®, System Storage SAN Volume Controller, Tivoli® Storage FlashCopy® Manager, IBM® Power® 750, FlashSystem™ 840, XIV®, IBM System and Technology Group Lab Services, IBM Software Group Services Business challenge: Balluff knew that slow access to finance and business reports threatened productivity and potential growth. How could executives gain fast insight into critical data to make better business decisions? The solution: The company moved its SAP Business Warehouse to IBM® DB2® with BLU Acceleration, running on IBM Power Systems™ with IBM AIX® and IBM PowerHA®. “IBM DB2 with BLU Acceleration is the ideal solution for us because we can gain new insights into business data more rapidly. Deploying IBM DB2 with BLU Acceleration was a low-risk project; implementation was quick and easy without affecting availability.” —Bernhard Herzog, Team Manager Information Technology SAP, Balluff Deep Blue IBM Solution SAP Stack 18 © 2014 IBM Corporation IBM SAP Alliance © 2014 IBM Corporation Big Data & Analytics Big Data & Analytics The Business Value, Why Speed is Money Business Analytics Accelerator Increase productivity & drive business growth Adding Value, not complexity Dynamic Query Dynamic Cubes Infrastructure That Matters 82X más rápido C1 C2 C3 C4 C5 C6 C7 C8 DB2 with BLU vs. Competitor Row Store Database on Ivy Bridge (x86)1 19 © 2014 IBM Corporation Compatible Query 1) Based on IBM internal tests as of April 7, 2014 comparing IBM DB2 with BLU Acceleration on Power with a comparably tuned competitor row store database server on x86 executing a materially identical 2.6TB BI workload in a controlled laboratory environment. Test measured 60 concurrent user report throughput executing identical Cognos report workloads. Competitor configuration: HP DL380p, 24 cores, 256GB RAM, Competitor row-store database, SuSE Linux 11SP3 (Database) and HP DL380p, 16 cores, 384GB RAM, Cognos 10.2.1.1, SuSE Linux 11SP3 (Cognos). IBM configuration: IBM S824, 24 cores, 256GB RAM, DB2 10.5, AIX 7.1 TL2 (Database) and IBM S824, 16 of 20 cores activated, 384GB RAM, Cognos 10.2.1.1, SuSE Linux 11SP3 (Cognos). Results may not be typical and will vary based on actual workload, configuration, applications, queries and other variables in a production environment. 82x calculation based on geometric mean calculation giving equal weighting to the report per hour (RPH) improvements in the three categories of simple, intermediate, and complex reports. GEOMEAN(RPH_simple,RPH_intermediate,RPH_complex) = GEOMEAN(18.85,40.07,747.63)=82.66 POWER + BLU Acceleration pero y …. ¿Cómo avanzamos? Elegir una opción de las dos configuraciones Small Configuration Large Configuration Existing Data Warehouse Existing Data Warehouse • • Supports up to 5 TB uncompressed active data (1 TB compressed data) Supports up to 10 TB uncompressed active data (3 TB compressed data) Hardware Configuration Hardware Configuration • Power S814: 8 cores, 3.72 GHz • Power S824: 24 cores, 3.52 GHz • Memory: 256 GB DRAM • Memory: 1 TB DRAM • Storage: • Storage: • • 146 GB (RAID 1) (OS/PGM VG) 2.4 TB HDD (RAID 5) (Data VG) • AIX Standard Edition • PowerVM Enterprise Edition Software Configuration • DB2 Advanced Workgroup Edition • 2.4 TB HDD (RAID 5) (Data VG) • 1.55 TB SSD (DB2 overflow cache) • 146 GB (RAID 1) (OS/PGM VG) • AIX Standard Edition • PowerVM Enterprise Edition Software Configuration • DB2 Advanced Workgroup Edition PRECIO MUY INTERESANTE ** PREGUNTE A SU VENDEDOR DE IBM ** 21 © 2014 IBM Corporation Configuración Try&Buy, incluye HW & SW New BLU POC Program Stack: Power 8, DB2 10.5 with BLU Acceleration and Cognos BI 10.2 with Dynamic Cubes Target Use Cases: “Business Intelligence Workload Accelerator” Provides acceleration for BI workloads in competitive or back-level environments Can work in any environment Commodity hardware accounts Microstrategy, Business Objects SQLServer, Oracle Is an accelerator to your existing environment Configurations 22 5 pre-defined Power 8, DB2 BLU and Cognos BI configurations to support various workload sizes © 2014 IBM Corporation Announcement Highlights – Announced: June, 2014 – Leverages established Power POC processes – Define the free trial period – Try with your own data Owners for Additional Information – lking@ca.ibm.com – kschlamb@ca.ibm.com GRACIAS