Presentation is loading. Please wait.

Presentation is loading. Please wait.

301-Getting Started with BW

Similar presentations


Presentation on theme: "301-Getting Started with BW"— Presentation transcript:

1 301-Getting Started with BW
Andy Heldt – BW/BI&A committee member Kimberly-Clark Corporation 5/28/2003

2 Assumptions/Goals Assumptions Goals
You may not use BW but are aware of what BW is Goals Provide a broad overview of BW as a tool Provide some of the ‘essentials’ you need to know when getting started Provide tips and tricks from experiences

3 Kimberly-Clark SnapShot
Consumer products company Kleenex, Scott, Huggies, Pull-Ups, Kotex and Depend Products are manufactured in 43 countries and sold in more than 150. 5 BW landscapes + 3 APO Started in 1999 on BW 1.2B currently BW 3.1C Over 200 Production InfoCubes, even more ODS structures 3-4 Terabytes of data

4 Agenda BW Overview in 4 slides Fundamentals
Reporting tools (Front-End) Extraction, Transformation, & Loading Open Hub Services Other Things to Know Informational Resources

5 What is BW? BW is SAP’s tool to enable business analysis, analytical applications, data mining, etc. Integrated into the SAP environment R/3, CRM, APO, EBP, etc. MySAP HR, CRM, SCM, PLM, FIN, etc. Integrated at both the component level (CRM, APO, EBP, SEM) and the business solution (mySAP HR, CRM, SCM, etc.). If you are already a R/3 customer looking to grow your SAP investment BW is the logical choice. It is the foundation for many of the other business solutions from SAP.

6 BW Overview The toolset for querying or analyzing information in BW
The toolset for managing the BW infrastructure The toolset for taking information out of BW The toolset for bringing data into BW The reporting and analysis capability at the front end. BW comes with two interfaces, BEX Analyzer built off of MS Excel and web built off of the SAP provided web application designer (WAD). Both are effective and full functioned. There are a host of 3rd party providers available that will interface with BW to provide front-end capability. The extraction, transformation and loading at the back end. BW has numerous capabilities to take data from external sources and bring it into BW. BW provides multiple places where data transformations can occur. One thing to note, BW is currently a pull technology and does not support a data source handing off data to BW. Open Hub Services is new with BW 3.0 and provides the mechanism to take data out of BW. This is relatively new and supports BW being the hub of information to feed into external data sources. Currently, BW pushes the data out instead of the external system pulling the data. The administration workbench has a slue of tools to assist with developing and managing the environment. This includes the tools needed to build all of the objects and structures within BW, loading and monitoring the data in BW, performance tuning the warehouse, managing change control (not version control), scheduling activities, and data management like archiving. The middle is not discussed at this time. The middle section represents the various structures that host the data within BW.

7 BW positioning – Part of the Foundation
BW is not the solution, but one of the tools that the solution is built upon Key points in this slide. BW is one of the core underlying tools of the SAP environment. BW is not the solution, but one of the tools that the solution is built upon. BW is a technology for enabling business analysis, analytical applications, data mining, etc. This is a fundamental shift that has been showing up in the SAP information. It is important to understand this separation as it will help in defining how BW fits into the organization both as a technology and business solution. Separating these out also helps in understanding what the knowledge base for BW needs to be and the scope around where BW will be used. MySap BI is the SAP provided business solution, not the underlying tool

8 SAP Positioning - NetWeaver
One of the tools to support developing and deploying business solutions BW is not the solution but the tool. Re-enforcing the idea around BW being a part of the foundation.

9 Agenda BW Overview in 4 slides Fundamentals
Reporting tools (Front-End) Extraction, Transformation, & Loading Open Hub Services Other Things to Know Informational Resources

10 Fundamentals Terminology Naming standards Security Team structure Methodology System Maintenance Shared master data

11 Terminology - Data model
Fact table Dimensions Characteristics Key figures

12 Terminology InfoObjects – Base data element
Characteristics – Master data Navigational attributes – attributes of a characteristic that you can drill on Key Figures – Measures within fact table Dimensions – Logical grouping of characteristics InfoCubes / Cubes – dimensional model of data from which most analysis will occur InfoSource – Definition and business rules for the data flow into a BW object DataSource – Definition of the input record layout for a dataflow

13 Fundamentals – Naming Convention
Need to define a naming convention for all the BW objects. Security is object based and a naming convention is essential to properly securing the environment. Consider how to be consistent with other ABAP conventions that your company may have. A solid naming convention … supports securing the various BW objects and projects … supports managing your environment This is critical to your long term success!

14 Fundamentals – Naming convention
Section Example Max Digits First Qualifier Next Qualifier(s) Standard Prefix SDUS 4 1-2: Application prefix 3-4: Country/Region InfoArea SDUS_SALES 30 1-4: Standard Prefix 5-30: Unique Identifier InfoCube SDUSCSLS 9 * 5: Sensitivity Indicator. C - Confidential P - Public 6-9: Unique Identifier InfoObejct Catalog InfoObject SDUSKF001 9 5-9: Unique Identifier Application Component InfoSource (Master Data or Text Data) 1-9: Previously defined InfoObject (Transactional Data) SDUSC_SALES_ITEM 6-30: Unique Identifier InfoCubes should only be 8 characters maximum to support export datasources. If the cube name is more than 8 there is a weird message that a person gets when trying to use an export datasource. Other basic techniques that can be considered include following SAP coding conventions and replacing the ‘0’ with ‘Z’ Infocubes, InfoSets, ODS Objects, Multi-providers all share the same namespace. May want to consider something in the naming convention to declare the type of object.

15 Fundamentals – Naming Convention
Section Example Max Digits First Qualifier Next Qualifier(s) Standard Prefix SDUS 4 1-2: Application Prefix 3-4: Country/Region Hierarchy SDUSC_CUSTOMER_HIER 30 1-4: Standard Prefix 5: Sensitivity Indicator. C - Confidential P - Public 6-30: Unique Identifier Source System (R/3, BW, APO) QC1300 10 1-3: SID of source system 4-6: Client of source system (PC File) PCFILE PCFILE - Constant Queries SDUSC_CASES_SHIPPED_BY_CUSTOMER Development Class / Package ZSDUS 1: Constant "Z" 2-5: Standard Prefix

16 Fundamentals - Security
How are you going to do security? Follow the SAP delivered roles Custom develop your own Be careful Allowing the same objects to be secured through multiple roles can create problems with controlling access. Data level security can be dangerous if is not implemented correctly Can not support giving a person different views into the same data. (regional vs. detail) Same rule may or may not apply across multiple cubes

17 Fundamentals - Security
Queries S_RS_COMP – Access to run queries, reports, build variables. S_RS_COMP1 –Enhancement to above to control by Owner S_RS_ICUBE – InfoCube access and maintenance S_RS_ISET – InfoSet access and maintenance S_RS_ODSO – ODS Object access and maintenance S_RS_HIER – Admin Workbench – Hierarchy S_RS_IOBJ – Admin Workbench – InfoObject Maintenance S_RS_IOMAD – Admin Workbench - Maintain Master Data S_RS_ISOUR – Admin Workbench - InfoSource - Transaction Data S_RS_ISRCM – Admin Workbench – InfoSource - Master Data S_USER_AGR – Authorizations: Role check S_USER_TCD – Authorizations: Transactions in roles Data & Structure ETL The following list contains the essential authorizations needed for building and executing queries, saving to favorites, and publishing queries to the web reporting interface. S_BDS_DS - BC-SRV-KPR-BDS: Authorizations for Document Set S_BDS_DS allows the maintaining of workbooks and web reporting queries to favorites and roles. The class BW_CATALOG is the favorite's capability. SAP keeps track of personal favorites and ensures that one person can not get at another person's favorites. The class BW_META_OBJECTS is set up to support putting workbooks and web reporting queries into roles to be used by multiple users. This authorization is needed by the publisher in order to publish BW workbooks and web queries to roles for general use by those authorized to the roles. Parameters: Activity:01, 02, 03, 06, 30 Class Name:BW_CATALOG, BW_META_OBJECTS Class Type:OT S_RS_COMP - Business Explorer - Components S_RS_COMP is the authorization that allows the execution of BW queries and workbook. Without this authorization a client can not run the query interface. The authorization can be as generic as a given InfoArea or as specific as a given query. A query is associated to an InfoCube and an InfoCube is associated with an InfoArea. This authorization does not allow access to data in an infocube or allow access to a role that contains a given workbook or web report. Activity:03, 16 InfoArea:0SDInfoCube:0SD_C01, 0SD_C02 Reporting component name:0SD_C01*, 0SD_C02*, SDC01_*, SDC02_* Reporting component type:* S_RS_HIER - Administrator Workbench - Hierarchy S_RS_HIER is the authorization that allows access to BW hierarchies. If a BW hierarchy is used within a query this authorization is required to allow the client to use the hierarchy. Without this authorization the hierarchy is not available to the client even though the client has access to the query. Activity:03, 71 Hierarchy name:* InfoObject:* Hierarchy version:* S_RS_ICUBE - Administrator Workbench - InfoCube S_RS_ICUBE is the authorization that allows access to the data within an InfoCube. This authorization is required to be able to present data from an InfoCube. The authorization can be as generic as a given InfoArea or as specific as a given InfoCube. Activity:03 InfoCube subobject:DATA InfoArea:0SD InfoCube :0SD_C01, 0SD_C02 S_USER_AGR - Authorizations: Role check S_USER_AGR is the authorization that allows access to add, change and delete workbooks and web queries from a role. This authorization is required to be able to publish workbooks queries to the various roles for general access. This authorization does not give access to a role. It only gives the ability to put workbooks and queries into roles. Activity:01, 02, 06 Role Name:0SD* S_USER_TCD - Authorizations: Transactions in roles S_USER_TCD is the authorization that allows the transaction RRMX to be put into various roles. Each role that has workbooks and queries in it needs to have access to the transaction RRMX, BEX Browser. Transaction code:RRMX

18 Fundamentals – Security Example
S_RS_COMP is the authorization that allows the execution of BW queries and workbook. This authorization does not allow access to data in an infocube or allow access to a role that contains a given workbook or web report. Parameters: Activity: 03, 16 InfoArea: 0SD InfoCube: 0SD_C01, SDUSCSLS Reporting component name: 0SD_C01*, SDUSC* Reporting component type: *

19 Fundamentals – Security Example
S_RS_ICUBE is the authorization that allows access to the data within an InfoCube. This authorization is required to be able to present data from an InfoCube. Parameters: Activity: 03 InfoCube sub-object: DATA InfoArea: 0SD InfoCube: 0SD_C01, SDUSCSLS

20 Security – One option… 4 essential profiles by BW application/solution
Power User – true ad-hoc capability for developing queries Display User – only has the ability to run something already built (use free characteristics to allow them to drill) Publisher – control point for proofing reports prior placing them into display user roles. Administrator – access to maintain & load BW Additional roles to hold the reports. These roles are more function based or solution based. For us it provided a cookie cutter approach for security whenever a new application or solution was being deployed.

21 Fundamentals – Team Structure
Consider how to break up responsibilities Data architect (design and structure build) Data steward (day-to-day loading/monitoring) Front-end (build/manage queries & roles) Others… Security (corporate computer security) Basis (BW is not R/3) Business involvement (required… business processes affect what goes into the warehouse) Success will be judged by your ability to meet business needs… make sure you are organized to have success! You need to think about how your company is going to organize to support the BW initiatives. The initial deployment is often a single centrally located team. That may not be what your organization need to look like after two, three, or four efforts. Some aspects are more suited to being centralized like basis, security, and data architect. Other aspects are more suited to sit with the business areas like data steward, front-end, business involvement. Bottom line: BW will ultimately be judged by the successfulness of business usage. Many sure your BW organization is aligned and in support of the business needs.

22 Fundamentals – Methodology
Use BW ASAP (Accelerated SAP) as the starting point. Work through it and add in steps as your company needs Make sure you check out the BW Best Practices

23 Fundamentals – System Maintenance
What is your system maintenance strategy? Remember BW is integrated with… R/3 (Plug-In), APO, CRM, EBP, SEM Are you a bleeding edge company? Plug-in – The interface between R/3 and BW Understand the constraints… based on R/3 extractors being used R/3 down/idle time may be required An overall strategy for the maintenance of your system is needed. Make sure you consider the other components of your SAP environment. BW is a critical part underneath many parts of SAP including R/3 (plug-in), CRM, APO, SEM, EBP. The plug-in interface has a number of significant constraints depending on what extractors your are using. The constraints range from requirements of the R/3 system to be idle while data is initialized for BW to risks associated with the application of support packages that can cause data lose on the BW interface.

24 Fundamentals – Shared Master Data
Characteristics are shared across Infocubes and ODS structures. Timing of master data loads versus transaction data loads Set of values in different cubes can be different Take your time, there are good and bad choices How to integrate non-SAP data What are the options when compounding Managing data conflict between source systems When the reporting solution needs to be available to the business, the supporting master data needs to be available as well. If there are multiple set of the source data being loaded through out a 24 hour period, when should master data be loaded to ensure the full set of values is available. A particular master data element will most likely have more values than values known to a particular cube. For example, an open orders cube versus an invoiced orders cube. Both would share the same master data for customer but there will be customers that have invoiced orders but no open orders and vice-versa. It may appear that there is only a small set of master data when in reality there is a large set. Integrating non-SAP data into the overall strategy presents decisions. What is the potential data conflict between values and how can those conflicts be mitigated. They are only needs for the migration, how do you plan to retire that data. The values are being migrated from the value in the legacy system to the new system but we need reporting by both the new and old values for an interim period. Consider putting the old value as an attribute. Compounding is one way to reduce the possibility of data conflict but can introduce constraints into the overall strategy. For example, compounding region to customer to uniquely tell the difference between two customer from different regions with the same customer number may be necessary but it also requires someone running a query to know and indicate what region they are talking about in order to get the correct result. Consider mapping to a new set of values or keeping one field and concatenating an additional identifier on to the value to make it unique. When managing data conflict consider the above discussion and what requirements are there for drilling back to the source. Some of the options are more restrictive in the ability for a user to take a BW report and follow it back into the source system where the data originated.

25 Agenda BW Overview in 4 slides Fundamentals
Reporting tools (Front-End) Extraction, Transformation, & Loading Open Hub Services Other Things to Know Informational Resources

26 Reporting Tools (Front-end)
BEX Analyzer MS Excel based. Part of SAPGUI as an add-on Web interface Web Application Designer (WAD) main tool for developing web reports. Similar to MS FrontPage or other web development tools but integrated into BW Support download to MS Excel capability Query Designer – required to write queries regardless of web or BEX Analyzer presentation Other 3rd party tools supported Use the ODBO interface 3rd party tools that use the ODBO interface can at best be as fast as BW. 3rd party tools should be used to augment your environment or because you already have them and want to minimize change to the user interface. Be careful, not all 3rd party tools have the same interface when working with BW as they do with other databases.

27 Reporting Tools – Tips and Tricks
Sapbexc.xla – checks if BEX Analyzer is installed correctly. This Excel sheet is installed as part of SAPGUI. Web interface will be faster than BEX in most instances. 4-20 time faster… FrontEnd Performance article on SAP service marketplace  BW Performance Web interface has more capabilities for tuning (caching) commonly used reports. 3rd party tools live off the ODBO interface Requires a query to exist (still need to use query designer to build the query) Consider developing a standard template to be used for your of the reporting within the corporation. At a minimum, standardize the placement of objects on a the report like, variable prompts, free characteristics, support information, etc.

28 Reporting Tools – Tips and Tricks
Should queries/workbooks/web reports be developed directly in production? There isn’t a right answer Pros Immediate capability to run new reports/perform analysis Developing reports/analysis against real data Cons Direct maintenance capability in a production environment Risk in transports overlaying queries, workbooks, etc. Risk in impacting a production report through direct changes Consider your reasons for developing directly in production.

29 Agenda BW Overview in 4 slides Fundamentals
Reporting tools (Front-End) Extraction, Transformation, & Loading Open Hub Services Other Things to Know Informational Resources

30 Extraction, Transformation & Loading
The process of getting the data from the source to the target. 3 main steps or components Extraction – Selecting the data from the source system Transformation – Applying rules and checks to the data Realigning values Value consistency and reasonableness checks Referential integrity checks Load – moving the data into the BW components (PSA, ODS, InfoCubes & master data)

31 BW Data Flow - Terminology
DataSource – record layout of the data feeding into BW Allows BW to BW, APO, and R/3 to stay in sync Communication Structure – interim record layout between the business rules being applied to all the data and the business rules being applied just to the target. Transfer rules – the business rules being applied to all the data flowing into BW through a specific datasource. Update rules – the business rules being applied to the data flowing to a specific target Transfer routine – the business rules being applied to the InfoObject (or element). These are always applied.

32 BW Data Flow 1 2 R/3 BW 3 Source System InfoCube DataSource DataSource
Communication Structure PSA Update Rules Transfer Rules Transfer Routine Update Rules InfoCube Extractor Program(s) DataSource DataSource Comm. Structure ODS PSA Update Rules Transfer Rules Transfer Routine 1 2 Update Rules 3 ODS Extractor Program(s) Update Rules Here is the basic data flow from an R/3 system through BW. There are three places where transformation rules can be applied. During transfer rules as the data is moving out of a source. At this level the data is processed on a field by field basis and one record at a time. All fields in the input record are available but only the one being processed can be manipulated. You need to be careful because the fields get processed from top to bottom based on how they are listed. This can impact the results when referencing other fields than the one currently being processed. Note: if the source is feeding multiple targets the transfer rules are applied to all targets. During update rules as the data is moving into the target. At this level the data is processed on a field by field basis and one record at a time but there is something called a start routine that can optimize processing. All fields are available and can be manipulated by any routine on a field. Again, diligence is required to understand the impact of multiple changes and the order that they get applied. Note: update rules are only applied on the data going to the specific target identified. Transfer routines are specified on the InfoObject but are applied during the data load process. They apply to all instances where that InfoObject is being loaded. There is no control over when to apply and when not to apply. Transfer routines can be expensive since they are executed every time that InfoObject is loaded in master data, text, hierarchies, and transaction data. They should be the last choice for transformation requirements. Note: Transfer routines are applied immediately when the data comes into BW. That is prior to the data being stored in PSA or being used in transfer rules. DataSource DataSource Extractor Program(s) Comm. Structure InfoCube PSA Transfer Rules Update Rules Transfer Routine R/3 Tables

33 ETL – Tips and Tricks Transformations = increased load times
Perform the transformations as close to the source when ever possible. Outside of BW if possible. Manage the number of transformations needed against the expected throughput of load activities Consider using Start Routines within update rules instead of specific update rules. Allows the routine to be applied against the entire data packet instead of record by record. Consider segmenting the data and performing parallel loads to increase throughput. Build this into the overall design.

34 ETL – Tips and Tricks Seriously consider using the new method for UOM conversions instead of using transfer rules and/or update rules to pre-calculate and store the values. On SAP Marketplace  BW Services & Implementations  How to guides  How to Report Data in Alternate Units Our experience… what impacts throughput… # of transformations Width of record (# of infoobjects) Volume (# of records) Aggregates Master data exist? Each transformation represents some ABAP code that needs to run. Depending on how it is coded, this can mean lots of additional processing over the course of the load. Remember, BW does not use DB utilities for loading data. Instead there is an application program running that is performing inserts, updates and deletes. This is a significant difference than many other DSS tools which lay on top of whatever infrastructure or DB that holds the data. During the running of an InfoPackage you can control what to do if master data does not exist. The options are to not load the data or to generate a SID for the master data entry and continue. When BW has to generate a SID entry the load process switches to a different set of logic from that point forward. This does significantly slow down the load process. Normal loading BW is performing reads to determine SID ids but once there is a non-existent SID BW switches to performing inserts as the first check from then on. If the insert happens early in the load then all data loading after that point goes through the slower process.

35 Loading Data Text data – generates SID entries
Attribute data – loads attributes of characteristics Hierarchy/Attribute Change run Transaction data loads If transaction data requires attributes of characteristics for transformations, must run in order. Otherwise, can run 1, 4, 2, 3 Transaction data loads will not be ‘ready for reporting’ if any previous loads were not successful.

36 Agenda BW Overview in 4 slides Fundamentals
Reporting tools (Front-End) Extraction, Transformation, & Loading Open Hub Services Other Things to Know Informational Resources

37 Open Hub Services Used to get the data out of BW
Basic object is called an InfoSpoke Data Sources: InfoCube, ODS Object, and Master Data (Attributes and Text) Full and Delta Processing (Delta for InfoCubes and ODS Objects) Distribution targets: File (.csv format), BW DB Table Aggregation, Selection, Transformation Scheduling Direct Scheduling Process Chain Integration Extracting data from SAP BW by any means requires additional licensing! Contact your SAP representative.

38 Open Hub Service Open Hub Service BW Server Relational Table External
DataMart . . . InfoCubes Open Hub Service Flat File Controlled distribution of consistent data Target: file or DB table Central monitoring Select filter criteria and columns Scheduling Full or delta mode Master Data ODS Objects

39 Agenda BW Overview in 4 slides Fundamentals
Reporting tools (Front-End) Extraction, Transformation, & Loading Open Hub Services Other Things to Know Informational Resources

40 Where Do You Start… Infoobjects – Characteristics and key figures
Must define attribute objects first. InfoCubes/ODS structures InfoSources including transfer rules Update rules InfoPackages Queries & reports (* could be done after 2)

41 Use Business Content or Not?
Define: The SAP provided extractors, dataSources, InfoSources, Infoobjects, ODS, Infocubes, business rules, queries, etc. Very high return in using the SAP provided Extractors, Infoobjects and InfoSources. Let your gap analysis show how effective business content Infocubes and ODS will be. Use business content unless it is more cost effective to not. Myth: If business content is used my cubes, etc will change when changes to business content come out through SP or releases. ‘D’ = SAP Delivered ‘A’ = active BW objects ‘M’ = object currently being changed but not activated

42 Design Considerations - Dimensions
Dimensions are logical grouping of characteristics. A Dimension row is created for each unique set of characteristics within the dimension that are loaded into the fact table. Avoid combining common used query filters like status or type with things like document number Careful with a date dimension (365 * 365 * 365) like ordered date, requested date & delivery date Large dimensions (document number) should be turned into line item dimensions

43 Design Considerations - ODS
Make sure you understand how ODS key field works What characteristics identify this unique transaction, not what are all the characteristics of this transaction Ex. Document number, line item number NOT, document number, line item number, product, date, etc. The other characteristics are just data fields. ODS Implementation Whitepaper Service Marketplace  BW  Documentation  Enhanced  2.0B Whitepaper BW Operational Data Store

44 Design Consideration - Aggregates
BW 3.0 has removed most constraints related to aggregate and hierarchy/attribute change run contention Data loads (transaction and master data) can cause aggregate rebuilds to occur. Aggregate rebuilds are expensive What is your SPRO  Maintain Aggregate Realignment Run setting. SAP recommends 20%. General performance practice… Get BW Statistics turned on Monthly, use the BW tools to advise on aggregates Create specific ones based on critical queries as needed Justify aggregate use because they cost resources and time during load processes and they consume disk space

45 Managing Expectations
Manage response time expectations This is OLAP not OLTP Many queries will process thousands of rows of data Web is generally faster than BEX (4-20 times) Records / Cells 1,000 10,000 100,000 500,000 1 million 100 Fastest 5,000 20,000 Slowest

46 Managing Expectations
BW implementation timelines vary widely. Biggest factors.. Using Business Content out-of-the-box Simultaneous implementation with other SAP solutions, R/3, APO, CRM, etc. Experience of BW staff Availability of business process resources

47 Agenda BW Overview in 4 slides Fundamentals
Reporting tools (Front-End) Extraction, Transformation, & Loading Open Hub Services Other Things to Know Informational Resources

48 OSS ID required

49 www.service.sap.com/bw Services & Implementations
How to guides – wonderful guides on how to do many different things. They are also great for getting ideas. Know-How-Network SAP teleconference presentation held every other week Topics vary, see web site for schedule and call-in info Documentation Good supplemental information. Often more timely or directly related to SP enhancements/capabilities Lots of others… release info, SP info, performance

50 Other good sites: www.searchsap.techtarget.com www.techrepublic.com
Discussion forms  BW – BI&A

51 Thank you for attending! aheldt@kcc.com
Please remember to complete and return your evaluation form following this session. Session Code: 301


Download ppt "301-Getting Started with BW"

Similar presentations


Ads by Google