Why sap hana

13,463 views

Published on

Why do we need SAP HANA? What is SAP HANA. High Performace Analytical App

Published in: Technology
  • For SAP online training register at http://www.todaycourses.com
       Reply 
    Are you sure you want to  Yes  No
    Your message goes here

Why sap hana

  1. 1. WHY SAP HANA ?<br />Ugur CANDAN<br />August2011<br />
  2. 2. Struggling to Keep Up With Expectations?<br />
  3. 3. The Gap Between CPU and GPU<br />ref: Tesla GPU Computing Brochure<br />
  4. 4. The Gap Between CPU and GPU<br />
  5. 5. Evolution of Intel Pentium<br />Pentium I<br />Pentium II<br />Chip area<br />breakdown<br />Pentium III<br />Pentium IV<br />Q: What can you observe? Why?<br />ref: Zhenyu Ye / Bart Mesman / Henk Corporaal “GPU Architecture and Programing”<br />
  6. 6. Extrapolation of Single Core CPU<br />If we extrapolate the trend, in a few generations, Pentium will look like:<br />Of course, we know it did not happen. <br />Q: What happened instead? Why?<br />
  7. 7. Evolution of Multi-core CPUs<br />Penryn<br />Chip area<br />breakdown<br />Bloomfield<br />Gulftown<br />Beckton<br />Q: What can you observe? Why?<br />
  8. 8. The Brick Wall -- UC Berkeley's View<br />Power Wall: power expensive, transistors free<br />Memory Wall: Memory slow, multiplies fastILP Wall: diminishing returns on more ILP HW<br />Power Wall + Memory Wall + ILP Wall = Brick Wall<br />David Patterson, "Computer Architecture is Back - The Berkeley View of the Parallel Computing Research Landscape", Stanford EE Computer Systems Colloquium, Jan 2007, link<br />
  9. 9. MapReduce<br />Dynamo<br />Google File System<br />Big Table<br />
  10. 10. Xpress disk (xpd) 1 TB: 20 GB/s IOPS: 600,000 reads 500,000 writes<br />The NVIDIA Tesla c2070 memory 6GB interface 384bit 144 GB/s<br />Fusion ioDrive Octal Capacity 5.12TB 6GB/s Read 4GB/s Write 1,000,000 IOPS<br />8K Camera: 260 km by a fiber optic network 3 GB/s<br />a 20 minute broadcast would require roughly 4 TB of storage<br />
  11. 11. RDBMS in MPP Architecture<br />Useful work of a RDBMS<br />Each SMB 6.4 GT/s <br />
  12. 12. “Typical” Business Intelligence Today<br />Query<br />Query<br />Slow<br />Painful<br />Expensive<br />Slow<br />Painful<br />Expensive<br />ETL<br />ETL<br />Copy<br />Copy<br />Reporting<br />Data Warehouse<br />Business Applications<br />Calculation Engine<br />Business Intelligence<br />Query Results<br />Data Warehouse<br />Business Applications<br />Calculation Engine<br />Business Intelligence<br />Query Results<br />DataMarts<br />Data<br />Data<br />Aggregates<br />Aggregates<br />Indexes<br />Indexes<br />Operational Data Store<br />Operational Data Store<br />
  13. 13. In-Memory Computing<br />Disk is 1Mx slower than direct memory: like a chef doing his shopping on mars <br />
  14. 14. In-Memory Computing Costs have Plummeted<br />Cost of 1 Mb of memory in 2000: ≈$1<br />Bosphorus bridge:<br />105m / 344ft<br />
  15. 15. In-Memory Computing Costs have Plummeted<br />Cost of 1 Mb of memory today: <1 cent<br />And shrinking….<br />Dell: Quad 10 Core server 512GB Ram, Hosting Services<br />$5,200/month<br />Child:<br />1.5m<br />
  16. 16. A Shift of Frontiers in Computer ScienceFreely Adapted from Jim Gray, Turing Award Winner 1998<br /><ul><li>Tape is Dead
  17. 17. Disk is new Tape
  18. 18. Main Memory is new Disk
  19. 19. CPU Cache is new Main Memory</li></li></ul><li>SAP ve In-Memory<br />2010 SAP High-Performance Analytic Appliance<br />2006 SAP NetWeaver BW Accelerator (BWA)<br />2004 SAP NetWeaver Enterprise Search (TREX)<br />1999 SAP Advanced Planner and Optimizer<br />
  20. 20. Operations and Analytics Together<br />Add ACID-compliant, row-based, in-memory<br />Single source of data<br />Faster, better BI and actionable intelligence<br />Faster, better applications<br />New application opportunities<br />Copy<br />Business Applications<br />Business Intelligence<br />Data<br />Analytic Appliance<br />
  21. 21. SAP HANA<br />
  22. 22. In-Memory Computing – The Time is NOW<br />HW Technology Innovations<br />SAP SW Technology Innovations<br />Row and Column Store<br />Multi-Core Architecture (8 x 8core CPU per blade)<br />Massive parallel scaling with many blades<br />One blade ~$50.000 = 1 Enterprise Class Server<br />Compression<br />64bit address space – 2TB in current servers<br />100GB/s data throughput<br />Dramatic decline in price/performance<br />
  23. 23. Fast – Software Optimization for Memory<br />conceptual view<br />Conventional databases store records in rows<br />Storing data in columns enables faster in-memory processing of operations such as aggregates<br /><ul><li>Columnar layout supports sequential memory access
  24. 24. A simple aggregate can be processed in one linear scan </li></ul>mapping to memory<br />organize by row<br />A<br />10<br />€<br />B<br />35<br />$<br />C<br />2<br />€<br />D<br />40<br />€<br />E<br />12<br />$<br />organize by column<br />A<br />B<br />C<br />D<br />E<br />10<br />35<br />2<br />40<br />12<br />€<br />$<br />€<br />€<br />$<br />memory address<br />
  25. 25. In-Memory Computing – The Time is NOW<br />HW Technology Innovations<br />SAP SW Technology Innovations<br />Row and Column Store<br />Multi-Core Architecture (8 x 8core CPU per blade)<br />Massive parallel scaling with many blades<br />One blade ~$50.000 = 1 Enterprise Class Server<br />Compression<br />Partitioning<br />64bit address space – 2TB in current servers<br />100GB/s data throughput<br />Dramatic decline in price/performance<br />No Aggregate Tables<br />
  26. 26. SAP HANA™<br /><ul><li>In-Memory software + hardware(HP, IBM, Fujitsu, Cisco, Dell)
  27. 27. Data Modeling and Data Management
  28. 28. Real-time Data Replication
  29. 29. SAP BusinessObjects Data Services for ETL capabilities from SAP Business Suite, SAP NetWeaver Business Warehouse (SAP NetWeaver BW), and 3rd Party Systems</li></ul>Capabilities Enabled<br /><ul><li>Analyze information in real-time at unprecedented speeds on large volumes of non-aggregated data
  30. 30. Create flexible analytic models based on real-time and historic business data
  31. 31. Foundation for new category of applications (e.g., planning, simulation) to significantly outperform current applications in category
  32. 32. Minimize data duplication</li></ul>SAP In-Memory Appliance (SAP HANA™)<br />SAP BusinessObjects tools<br />Other query tools<br />SQL<br />MDX<br />BICS<br />SQL<br />SAP HANA <br />SAP In-Memory Computing Studio<br />SAP In-Memory Database<br />Row & Column Storage<br />Calculation and Planning Engine<br />Real-Time Data Replication<br />SAP Business Objects Data Services<br />SAP Business<br />Suite<br />Other data sources<br />SAP NetWeaver Business Warehouse<br />
  33. 33. HANA Combines Software and Hardware<br />In-Memory Computing Engine (Software)<br />+<br />Pre-Installed Systems (Hardware)<br />
  34. 34. S+<br />S<br />XS<br />HANA Appliance “T-shirt” sizes Specifications & Approximate Data Volumes<br />Starts at S and scales up to M<br />
  35. 35. L<br />M<br />M+<br />HANA Appliance “T-shirt” sizes Specifications & Approximate Data Volumes<br />Starts at M and scales up to L<br />
  36. 36. Speed-up Existing Work<br />
  37. 37. Discover New Options<br />
  38. 38. Customer Testimonials<br />Consumer and Health Products<br />Manufacturing<br />Energy<br />Communications and Media<br />Services<br />Technology<br />
  39. 39. The First 35 Years: Innovated with ERP & LOB Apps<br />Three Years Ago: Innovated with Analytics<br />Last Year: Innovated with Mobility<br />This Year: Innovating the Database<br />HANA Accelerates Data, Applications, Analytics<br />Long Term: HANA Is the Database<br />Mobility<br />Accessible Systems<br />ERP + LOB<br />Business Analytics<br />Data “In”<br />BICS<br />Info “Out”<br />Systems of Record<br />Systems of Engagement<br />Business Applications Performance Bound by Data<br />Oracle<br />DB2<br />SQL<br />Other<br />HANA<br />In Memory Database<br />HANA<br />In Memory Database<br />Oracle<br />SQL<br />DB2, etc.<br />ELT or ETL<br />ELT or ETL<br />
  40. 40. The Truth About Enterprise Data Landscapes<br />The Value of HANA<br />The Future of “The Stack” – HANA Nirvana<br />?<br />Queries<br />Ad-Hoc<br />Reports<br />ETL<br />DATA QUALITY<br />Dashboard<br />Information<br />Semantic<br />Universe<br />OLAP<br />OLAP<br />OLAP<br />OLAP<br />Dimensional<br />HANA<br />HANA<br />Mart<br />Mart<br />Mart<br />Mart<br />Mart<br />Mart<br />Mart<br />Marts<br />BW<br />TeraData<br />Netezza<br />IQ<br />BW/Netezza/Teradata/IQ<br />Warehouses<br />Oracle<br />Oracle/DB2/SQL/Other<br />DB2<br />Other<br />SQL<br />Operational<br />Applications<br />
  41. 41. SAP BusnessObjects Event Insight <br />Event InsightNode<br />Event Insight<br />Node<br />Event InsightNode<br />WWW<br />Supplier<br />Mfg Plant<br />Apps<br />Apps<br />Event InsightNode<br />Event Insight<br />Apps<br />Custom DB<br />Event InsightNode<br />DSD DistributionCenter<br />Retailer A<br />Distribution Center<br />Distribution Center<br />Event InsightNode<br />Event InsightNode<br />Apps<br />BW<br />Retail Store A<br />Retail Store B<br />
  42. 42. SAP BusnessObjects Event Insight <br />Event InsightNode<br />Event InsightNode<br />Event InsightNode<br />WWW<br />Event Pattern<br />Supplier<br />Event Pattern Identified<br />Define Event Pattern<br />Mfg Plant<br />Apps<br />Apps<br />Event InsightNode<br />Event InsightNode<br />Apps<br />Custom DB<br />Event InsightNode<br />DSD DistributionCenter<br />Retailer A<br />Distribution Center<br />Distribution Center<br />Event Insight<br />Node<br />Event Insight Node<br />Apps<br />BW<br />Retail Store A<br />Retail Store B<br />

×