Professional Documents
Culture Documents
Agenda
1. 2. 3. 4.
Overview why in-memory? SAP BW powered by HANA and its benefits Modeling opportunities Project approach
5.
1. 2. 3. 4. 5.
Overview why in-memory? SAP BW powered by HANA and its benefits Modeling opportunities Project approach Key points to take home
Option HANA
SAP ERP
RDBMS
SAP ERP
HANA 1.0 RDBMS
SAP NW BW
RDBMS
BWA
SAP NW BW
HANA 1.0
SAP ERP
RDBMS
requests
sends report
Business User
IT
IT departments require
Up to date accurate information Get insight from the large volumes of information Be able to navigate and explore information to get what I need (self service) Simulate scenarios and discover trends (ad-hoc analysis) React faster to events impacting my operations
Providing a landscape for real time reporting An architecture that scales to my business needs Access to data in a more detailed level Bring in a technology to fulfill new business requirements Reduce cost and time to implement new requirements
Hardware-Innovations
Multi-Core Architecture
(8 x 10 core CPU per Server)
Software-Innovations
many nodes
Compression
in current servers
No aggregates neccessary
Prices for RAM are
decreasing
Shift of Paradigm
Impact for applications
Classical Approach
Calculation
Application Layer
Current applications perform many dataintense operations in the application layer High performance applications delegate data-intense operations to the in-memory platform
Database Layer
Calculation
1. 2. 3. 4. 5.
Overview why in-memory? SAP BW powered by HANA and its benefits Modeling opportunities Project approach Key points to take home
Enterprise Data Warehouse and Data Mart Modeling with SAP NetWeaver BW
Planning Engine
Analytic Engine
Data Manager
Filter + aggregation
InfoCubes
DataStore Objects
EDW Processes
10
Scale
Flexible
11
12
Current Situation
SAP BW DataStoreObject (DSO)
DataStore Objects are fundamental building blocks for a Data Warehouse architecture They are used to create consistent delta information from various sources Reporting can be done on a detailed level In today's RDBMS-based implementation, the activation and querying operations are extremely performance-critical
Delta upload
Query
Activation
Parallel Upload
SAP World Tour 2012
13
traditional
In-Memory
14
4500
Using in-memory computing technology one of the most time consuming staging operations the request activation was speed up tremendously by factor 5 - 10 ... storage of redundant data was prevented
300 20 3 Delta: 0.1 M, Active: 1 M 41 Delta: 1 M, Active: 10 M BW 7.30 - RDMBS based In-Memory optimized
473
15
Facts
D MD
Conversion/New
In-Memory MD F Facts
MD
MD
16
In-Memory Planning
Simple Disaggregation Example
Traditional Approach
1. Determine the delta +50 2. Disaggregate (in application server)
per week (52) per branch (500) 26000 combinations / values
HANA-Based Approach
1. Determine the delta +50 2. Send 1 value to HANA database + instruction to disaggregate and how 3. Disaggregate (in HANA engine)
per week (52) per branch (500)
17
1. 2. 3. 4. 5.
Overview why in-memory? SAP BW powered by HANA and its benefits Modeling opportunities Project approach Key points to take home
18
18
SAP NetWeaver BW
Data Mart 1 Data Mart 2 Data Mart 3
Catalog of key figures and characteristics Data load und process handling Business analytics and planning Authorizations Change- und transport management Dataintegration SAP- and non-SAPdata
SAP NetWeaver BW
Metadata Management Architected Data Marts Reporting / Planning DataWarehouse Layer Acquisition Layer
2
SAP HANA
SAP HANA
MD
Structured data
DB
External data Unstructured data
19
BW
Query
CompositeProvider
Query
Link Models
Transient Provider
InfoCube
HANA
1
AnalyticView
20
US AP EU 1:1
US APJ EU
US APJ EU
US APJ EU
AMS APJ EU
US APJ EU
AMS APJ EU
1. 2. 3. 4. 5.
Overview why in-memory? SAP BW powered by HANA and its benefits Modeling opportunities Project approach Key points to take home
22
22
DB migration
Optimization
(LSA, opt. BW Objects)
Transportation
(BW Objects, partly)
Optimization
(e.g. LSA)
Preparation
Hist. DataLoad
Rebuild
(BW Objects, partly)
Optimization
(e.g. LSA, during rebuild of objects)
23
NetWeaver BW 7.30 SP6 (minimum SP5); HANA 1.0 SP3 Separation of the Java Stack prior to the Upgrade of the ABAP Backend, if not already done The new BI Authorization concept is now mandatory. Despite the fact that you either create a 7.0x System beforehand to migrate the old BW Authorization, a potential Upgrade from BW 3.5 to 7.30 would be possible. The following source databases are supported for database migration:
Oracle 11.2 IBM DB2 LUW 9.7 MaxDB 7.8 MS SQL Server 2008 DB2 for i61, 7.1 DB2 for z/OS V9, V19
24
1. 2. 3. 4. 5.
Overview why in-memory? SAP BW powered by HANA and its benefits Modeling opportunities Project approach Key points to take home
25
25
Summary
Overview Why in-Memory? Large data volume Real-time data Simulation scenarios Conclusion: faster reaction time and possibility for new business requirements SAP BW powered by HANA and its benefits In-Memory optimized objects for much better performance (activation/loading and reporting) Simplified InfoCube Calculations for planning on the in-memory plattform Conclusion: reduction timeframe of data loading and much better reporting/planning performance Modeling opportunities Combination of agile and architected data marts Simplified layered architecture (LSA) Conclusion: more flexibility and less data redundancies Project approach Full migration of existing SAP BW (DB migration) New installation and partial transport of existing data models Conclusion: Different project approaches depending on customers requirements
26
26
Thank You
Thank you
Contact information:
Thomas Kng Senior Solution Consultant HANA Center of Excellence (CoE) Switzerland thomas.kueng@sap.com
Backup
How does BW-HANA differ from a standard BW 7.3? (1) DB + BWA In-Memory DB (2) InfoCube (3) DSO (4) Planning In-Memory Optimized InfoCube In-Memory Optimized DSO In-Memory Based Planning Engine
30
How does BW-HANA differ from a standard BW 7.3? (1) DB + BWA In-Memory DB
One license, one instance No BWA required, no data replication Query performance at least as good as with BWA Full BWA feature support Pure DB Migration, same look&feel. Investments preserved
31
How does BW-HANA differ from a standard BW 7.3? (1) DB + BWA In-Memory DB (2) InfoCube
Simplified flat structure (only one fact and no dimension tables) Easier modeling, faster re-modeling Faster data loads (no DIMDIs) Simple conversion to new type
32
How does BW-HANA differ from a standard BW 7.3? (1) DB + BWA In-Memory DB (2) InfoCube (3) DSO In-Memory Optimized InfoCube In-Memory Optimized DSO
Activation step in HANA engine Shortened load window Fast reporting on DSO
33
How does BW-HANA differ from a standard BW 7.3? (1) DB + BWA In-Memory DB (2) InfoCube (3) DSO (4) Planning In-Memory Optimized InfoCube In-Memory Optimized DSO In-Memory Based Planning Engine
34