Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.

1310 success stories_and_lessons_learned_implementing_sap_hana_solutions


Published on

hana successes

Published in: Education
  • Be the first to comment

1310 success stories_and_lessons_learned_implementing_sap_hana_solutions

  1. 1. September 10-13, 2012 Orlando, FloridaSuccess Stories and Lessons Learned Implementing SAP HANA SolutionsSession 1310
  2. 2. About Me Jonathan Haun Consulting Manager with Decision First Technologies. Over 12 years implementing BI solutions utilizing Crystal Reports and BusinessObjects Experience with multiple ETL tools and RDMS Certified in multiple versions of BusinesObjects Enterprise, BusinessObjects Data Services and BusinessObjects Reporting tools Certified SAP Trainer Over one year of experience implementing BOBJ and BW solutions on SAP HANA Manager of the “All Things BOBJ BI blog”. Twitter Feeds @jdh2n
  3. 3. Learning Points What is SAP HANA ? Solutions Currently Available based on SAP HANA Tips and Tricks implementing solutions on SAP HANA standalone Performance Improvements ROI for Solutions Based on SAP HANA3
  4. 4. What is SAP HANA ?
  5. 5. What is SAP HANA ? SAP HANA is a true next generation database. It is a fusion of both software and hardware that is fully engineered to provide deep and rich access to billions of rows of data. Data on SAP HANA is stored and accessed in RAM which eliminates the traditional limitations of spinning magnetic disks. Data is stored physically close to the CPU allowing faster and more direct access to the data. The HANA software is the first of it’s kind to be built from the ground up to support multi-core CPU’s and in-memory direct access to data. Memory First, Persistent Storage Second. Data can be stored using either Row or Column based storage providing flexibility in supporting both Front End Applications and Business Intelligence on a single platform. SAP HANA has built in Multi Dimensional Modeling Views that can mimic the capabilities of OLAP without the need to duplicate data into Cubes. This reduces the TCO of storing the data while providing the flexibility to perform both MDAS or Operational reporting.5
  6. 6. Solutions Currently Available based on SAP HANA6
  7. 7. Solutions Currently Available based on SAP HANA SAP HANA Standalone Leverage the power of SAP HANA with data from any source. Move data into SAP HANA in real-time or in batch using replication or Data Services 4.0 Highly customizable for any data analysis requirements Supports Analytics, MDAS, Forecasting and Operational reporting in a single solution Analyze billions of records at amazing speeds using SAP BusinessObjects 4.0 Similar to the Traditional EIM strategy of building Data Marts SAP Netweaver BW 7.3 powered by SAP HANA Run BW directly on SAP HANA by replacing your existing RDMS with SAP HANA Reduce the overall data footprint of BW with direct integration with SAP HANA Easy solution to adopt for existing SAP BW and SAP ECC customers Load 3rd Party Data in BW utilizing Data Services Analyze billions of records at amazing speeds using SAP BusinessObjects 4.07
  8. 8. Solutions Currently Available based on SAP HANA Multiple rapid solutions based on your Industry or Line of Business COPA, CRM, Finance, ERP, Cash Forecasting and many others. See more at Over 20 prebuilt solutions and growing Prebuilt solutions that are easy to deploy while reducing development costs Ability to manage billions of transactions with sub-second response times Analyze billions of records at amazing speeds using SAP BusinessObjects 4.08
  9. 9. Tips and Tricks for SAP HANA standalone Data Modeling vs Analytic Modeling Understand the difference before you begin developing solutions on SAP HANA standalone. Modeling in HANA centers around creating rich MDAS views, forecasting , and complex cross fact calculations. When loading data in batch, model your data into a traditional Star Schema or Fully De- normalized Table before loading SAP HANA. Joining hundreds of physical normalized tables will reduce the performance of queries significantly. Also Attribute Views that contain millions of records should be modeled into the Analytic Foundation tables to increase performance. Project Experience: 100% Analytic Modeling: Normalized Data loaded into HANA, containing multiple tables and joins and modeled completely in SAP HANA: 16 sec response times. 90% Data Services Data Modeling: With the same data, data modeled into a highly deformalized table, with Attribute Views containing less then 10,000 records and fewer then 10 total joins produced results in < 1 sec. Data Quality must be addressed via business processes and Data Services Merging Dimensional Data from multiple sources should be addressed using Data Services Reverse Pivots to de-normalize data should be addressed via Data Services Hierarchy Parsing should be addressed via Data Services.9
  10. 10. Tips and Tricks for SAP HANA standalone Using More SAP HANA “Analytic Modeling”10
  11. 11. Tips and Tricks for SAP HANA standalone Using More Data Services “Data Modeling”11
  12. 12. Tips and Tricks for SAP HANA standalone Analytic Modeling Tips Derived Attributes allow you to make copies of an existing Attribute View. The copies can not be modified and will always represent the configuration of their master. They are great for Date and Time based Attributes that are needed to create joins between multiple date columns found in the Analytic Foundation. When you update the master Attribute all child attributes will inherit the master’s columns. Similar to creating an alias of a table. Joins between an Attribute View and Analytic Foundation can not be defined on Decimal(x,y) columns. Joins on varchar or text columns do not perform as well as joins between Integer columns. SAP HANA has a built in Date / Time Dimension (Attribute View). There is no need to use Data Services to create a calendar date dimension. Joining multiple tables in the Analytic Foundation will greatly reduce performance. Better to use Data Services to combine the tables or Calculation Views to merge two independent Analytic Views. Every Column in an Analytic View must have a unique name. It is best to implement a naming standard before you begin your end-to-end Analytic View design phase. Be cautious making changes to existing Analytic Views (Post Development). Upstream content will stop working until it is re-activated. In some cases, removing columns, will require re-work on up-stream content.12
  13. 13. Tips and Tricks for SAP HANA standalone Universe Design Tips Using Analytic Views as the source for your Universe is great when that Universe is used primarily for Dashboards, Analytics and other reporting needs that contain Group By and Aggregate functions. All queries executed against an Analytic View must contain a measure. When you use an Analytic View within a Universe very little Universe design is required as most of the modeling and metadata were already defined in SAP HANA. It is also a good idea to leverage Analytic Views as current versions of SAP BusinessObjects Explorer and future versions of Webi will bind directly to SAP HANA Analytic Views. Using SAP HANA Columnar Tables within a Universe (Traditional Universe Design) will provide more flexibility for a variety of reporting requirements. Best if used for Operational Reporting, Multi-Fact Data Synchronization and situations where measures are not required. Multi-Fact Data Synchronization can be pushed down to SAP HANA by setting the JOIN_BY_SQL parameter to true within the Universe. The SAP HANA ODBC / JDBC drivers are only supported in the Information Design Universe (IDT.UNX) ODBC is easier to setup and maintain on Windows based BOE environments JDBC is easier to setup and maintain on UNIX / LINUX environments.13
  14. 14. Observations BW powered by SAP HANA BW powered by HANA Frequent SAP HANA patching was required during the initial POC. However, the current patch level (Patch 33) has proven to be more stable and we expect fewer and less frequent patching as the product matures. SAP has been very diligent in resolving issues but I would recommend adding time to any project estimate to account for unforeseen issues. Hardware vendor selection is important. Watch for vendors recommending supplemental hardware components. (Network Switches Etc..) Make sure that all supplemental hardware is compliant with your current environment. Frequent SAP HANA database settings were changed during the POC to find the right mix to support the customers in-house developed BW processes. Loading and processing of complex BW flows may require increasing parallel treads and memory settings on SAP HANA. Some BW modeling optimizations were required to fully support BW on HANA. In most cases the modeling code was poorly developed (not following best practices) in other cases SAP BW and or HANA required patches. Note: These issues had little or no impact on the project.14
  15. 15. Performance Improvements Customer’s History & Performance Gains BW with Legacy DB2 Legacy DB-BWA With Legacy DB-BWA With BW HANA With BOBJ HANA Stand Alone database. No BOBJ BOBJ 3.1 SP2 BOBJ 3.1 SP4 4.0 SP4 With BOBJ 4.0 SP4 HANA Standalone Baseline Faster Report Speed Faster Report Speed BW Powered by HANA Database •BWA 2.5x faster •Reports are faster •Early version of BOBJ not •Reports are faster then •BOBJ 3.1 SP4 very Stable then BWA and BW•No BWA or BOBJ stable BWA •Reports are faster due to Powered by HANAReports •No Improvement on Data •Delta Data Loads are Query Striping •Data Loads are very Loads much faster (3.2x) fastReport Response Times: • Reports: • Reports: • Reports: • Reports: •Avg 90 sec •Avg 40 sec •Avg 32 sec •Avg 6.4 sec •Avg < 1 sec •CRM 400 sec •CRM 161 sec •CRM 23 sec •CRM 4.6 sec •CRM < 1 sec •Data Loads 15 hrs •Data Loads 15 hrs •Data Loads 15 hrs •Data Loads 4.6 hrs •Full Loads 3.2 hrs * Directly Observed during recent POC with customer15
  16. 16. Performance Improvements Query Response Times 400 350 300 250 AVG (Sec) 200 CRM (Sec) 150 100 50 0 Base Line BWA BOBJ 3.1 SP2 BWA BOBJ 3.1 SP4 BW HANA HANA Standalone * Directly Observed during recent POC with customer16
  17. 17. Performance Improvements Data Load Times 16 14 12 10 8 Load Times 6 (Hrs) 4 2 0 Base Line BWA BOBJ 3.1 SP2 BWA BOBJ 3.1 SP4 BW HANA HANA Standalone* Directly Observed during recent POC with customer17
  18. 18. ROI for Solutions Based on SAP HANA Reduce Expensive SAN storage by moving data into memory with compression. Better integration of BW with the SAP HANA database appliance means more operations are managed by the in-memory database without the need to duplicate data multiple times. Multi-Providers on DSOs, in many cases, will work as well (if not better) then cubes on BWA. Further reducing the data footprint. Near Line storage, directly integrated into BW 7.3, can be utilized for infrequently accessed data. Allows for BW to maintain multiple storage tiers at different costs levels. User will experience faster response times making them more willing to deep-dive into the data. This will keep them focused on the Analysis task while increasing their knowledge and productivity.18
  19. 19. ROI for Solutions Based on SAP HANA Reduced Data Footprint with BW on SAP HANA Object Type BW on DB2 Step 1. Reduction in Decommission Reduce Layers with Near line Cleaning up to System Tables Obselete Objects HANA move to HANA PSA 5,842 100 100 100 100 100 Legacy DB Overhead 6,141 - - - - - Change Log 4,174 - - - - - DSO 1,487 1,487 1,487 1,312 1,201 1,201 System Tables 1,167 1,167 300 300 300 300 Cube 708 708 708 648 591 141 Master Data 408 408 408 408 408 408 Temp 7 7 7 7 7 7 Total (GB) 19,934 3,877 3,010 2,775 2,607 2,157 6 x Comp 3322 646 502 462 434 359 * Directly Observed during recent POC with customer.Your results might very based on the type and makeup of the data.19
  20. 20. ROI for Solutions Based on SAP HANA Overall Cost Savings Estimates ITEM SAVINGS SAN (Fast Storage) Eliminated (Very Expensive) DB2 (Database) Eliminated Total # of Servers Reduced by 30 -50% SAP Licenses and Hardware Increased Hardware and Software Staff Reallocated (DB2 and Storage) BOBJ No Change Total Estimated Savings (5 yr) 30-40% reduction in TCO * Directly Observed during recent POC with customer20
  21. 21. ROI for Solutions Based on SAP HANA Happy and Productive Users User Acceptance 10000 9000 8000 7000 6000 5000 User Acceptance 4000 3000 2000 1000 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 2121
  22. 22. Key Learnings SAP HANA is a fusion of both software and hardware. It was built from the ground-up to support in-memory data storage. Legacy vendors are still trying to develop in-memory solutions based on legacy DB architecture. There are three main solution categories for SAP HANA. SAP HANA standalone, BW powered by HANA and a wide variety of Rapid software solutions. For most organizations, SAP HANA solutions will yield significantly faster query response times. There will be an opportunity for most organizations to reduce their TCO for BW when powered by SAP HANA. User Acceptance is hard to measure in dollars, but when users are able to use BI in seconds and not minutes, they will be more productive and willing to leverage these tools.22
  23. 23. Links and References Data Modeling vs Analytic Modeling in SAP HANA – “All Things BOBJ BI” Creating Analytic Models in SAP HANA Studio - Question and Answer Responses – “All Things BOBJ BI” SAP BusinessObjects Explorer 4.0, powered by SAP HANA – “All Things BOBJ BI” SAP HANA FAQ – “All Things BOBJ BI” – SAP Hosted Site with great information on SAP HANA. SAP HANA product page - platform/index.epx23
  24. 24. Thank you for participating.Please provide feedback on this session by completing a short survey via the event mobile application. SESSION CODE: 1310Learn more year-round at