How to Deploy Enterprise Analytics Applications With SAP BW and SAP HANA Peter Huegel SAP Solutions Specialist
Agenda MicroStrategy and SAP Drilldown MicroStrategy and SAP BW Drilldown MicroStrategy and SAP HANA Ways to leverage HANA - Modeled Schema and Data Import Performance Recommendations Questions and Answers 2
MicroStrategy with (NextGen) SAP BW Queries BW4/HANA SAP HANA S/4HANA (OLTP) SAP R/3 - ECC SAP HANA
MicroStrategy with NextGen SAP Web services OLAP BAPI MDX ODBC/SQL ODBC/SQL Attribute Views* Analytic Views* Calculation Views* plus Database Views (*information model views) views BW Queries BW4/HANA SAP HANA SAP R/3 - ECC views S/4HANA (OLTP) SAP HANA Operational reporting directly against new generation SAP ERP through SQL Information model views plus CDS Views, database views
MicroStrategy with the new SAP in the big Picture Web services OLAP BAPI MDX ODBC/SQL ODBC/SQL Big Data views BW Queries SAP BW SAP HANA views S/4HANA EDW SAP R/3 - ECC SAP HANA
Agenda MicroStrategy and SAP Drilldown MicroStrategy and SAP BW Drilldown MicroStrategy and SAP HANA Ways to leverage HANA - Modeled Schema and Data Import Performance Recommendations Questions and Answers 6
Integration with SAP BW Protocols Data Retrieval We connect to SAP BW through RFC/ODBO MicroStrategy ODBC (SQL) ODBO (MDX) (Web Servces) BW Queries SAP BW 7
Working with SAP BW What can you expect to get? Key figures (standard, calculated, restricted) Characteristics Navigational attributes Filter BW variables (Selection, Authentication) Structures Cell references/calculations SAP exits (in key figures and variables) Customer exits (in key figures and variables) Hierarchies
Aspect Performance What can you expect to get? Standard Setup MicroStrategy with SAP BW 1. Good performance for small and medium data sets from BW (filter) 2. Less good performance for big data sets from BW Larger data sets are possible > 1m cell yet, performance definitely not google-like Best results: use MDX interface for what it is designed for (1) SAP brought SAP HANA to market for a reason Handling of large data sets Get more control over performance Get some agility comes to mind
Agenda MicroStrategy and SAP Drilldown MicroStrategy and SAP BW Drilldown MicroStrategy and SAP HANA Ways to leverage HANA - Modeled Schema and Data Import Performance Recommendations Questions and Answers 10
Integration with SAP HANA Protocols Data Retrieval We connect to HANA through ODBC/SQL MicroStrategy ODBC (SQL) SAP HANA Cloud Connector flavors of HANA On Premise Cloud Stand alone BW S/4 11
Integration with SAP HANA Protocols - Authentication MicroStrategy takes advantage of HANAenforced per-user security (Analytical Privileges) MicroStrategy ODBC (SQL) SAP HANA 12
Integration with SAP HANA Content Tables Table Row-based Table Columnar Information Views Attribute Views Analytical Views Shown as Calculation Views Tables other Views CDS Views Standard Views 13
Integration with SAP HANA Table Components Information Views Attributes Measures Calculated Columns Hierarchies Variables Input Parameters read as read as Columns Table-level prompts 14
Integration with SAP HANA Table Components Information Views Attributes Measures Calculated Columns Hierarchies Variables Input Parameters unused By now. See workshop Hierarchy Reporting Covered by Filter/Prompts > WHERE CLAUSE 15
Agenda MicroStrategy and SAP Drilldown MicroStrategy and SAP BW Drilldown MicroStrategy and SAP HANA Ways to leverage HANA - Modeled Schema and Data Import Performance Recommendations Questions and Answers 17
Governed data access and Self Service data access How relates Model Schema to Data Import? The best from both worlds with capability to combine it 18
Modeled Schema vs. Data Import Information Views in Data Import Feature Modeling Data Import Supports Information Views and/or Tables Both (Information) Views Only Attribute form identification based on column name Manual Automatic Identification/creation of metrics and default aggregation (sum, min, max, count) Manual Automatic Import of Input Parameters Automatic Automatic Input Parameter prompt remapping Yes No 19
OLAP Reporting with Data Import or Modeled Approach? When to use which approach? Use Data Import workflows preferably: End user driven data exploration End User can create data sources on their own End user experience is better New features will appear in Data Import first Use modeled way if Data Governance is key Administrators control access to data sources there are needs for advanced features (pass-through metrics, source order, etc.) there is a need to load data to Intelligent Cubes Eventually both this can be a process Data Import to explore and find data sets worthwhile being used Modelled approach to share with community at wider scale 21
Demo It all comes together HANA BW non SAP data 24
Agenda MicroStrategy and SAP Drilldown MicroStrategy and SAP BW Drilldown MicroStrategy and SAP HANA Ways to leverage HANA - Modeled Schema and Data Import Performance Recommendations Questions and Answers 25
Performance Recommendations 1 General guidelines Measure HANA MicroStrategy Can use Avoid transferring large data sets between HANA and MicroStrategy Filter as early as possible Aggregate data (by group by) Set Analytic Privileges Define Input Parameters Use Analytic Views (at lowest view level) Yes Yes Yes do calculation after aggregation (when business case allows) Modelling View Yes Model joins in HANA One view (output) providing needed information Yes 26
Performance Recommendations 2 Load Balancing across Nodes in HANA Make use of it Partitioning data in HANA F F F F F F MicroStrategy 10 HF1 onwards: Enable Prepared Statements against HANA [HKEY_LOCAL_MACHINE\SOFTWARE\Wow6432Node\MicroStrategy\DSS Suite\ConnectionLimit] "UsePrepareStatement"=dword:00000000 27
Performance Recommendations 3 MicroStrategy VLDB settings and modelling options for HANA Adjust the SQL to the situation Adjust situation Often use derived table instead of temporary tables 28
Is there an Intelligent Cube Story for SAP HANA? It should not but there is the situation SAP HANA can provide answers in sub seconds, seconds or a couple of seconds depending on what is happening in the view along the data model. So, it is possible that HANA provides a result set fast enough to match the attention span of mobile user used to Google performance An intelligent cube can guarantee an instantaneous response time as he has only to that: response to data request, no processing of business rules. 29
Thank You Contact: phuegel@microstrategy.com 30