Upload
dinhkhue
View
219
Download
4
Embed Size (px)
Citation preview
The Data Warehouse - Strategy
Service profile for
dispositive informationprocessing
Diplom-Informatiker Peter K. Albrecht
Head of the division
Systems-Consult
(Version: 01.05.1999)
DWH001.PPT; 7/13/99 14:09 ; Seite 2© Mummert + Partner Unternehmensberatung AG
Content
• Mummert + Partner Consultancy AG
• Why Data Warehouse Projects with Mummert + Partner
• The Mummert + Partner - Service profile for Data Warehouse Projects
DWH001.PPT; 7/13/99 14:09 ; Seite 3© Mummert + Partner Unternehmensberatung AG
1960 Dr. Olaf Mummert founded the Mummert consultancy in Karlsruhe on April 1st.
1996 Transformation of the consultancy Ltd. into a PLC.
We accompany our customers up to the target.
Entwicklung und Umsetzung von Branchenlösungen 42%Strategie-, Management und Organisationsberatung 39%IV-Beratung 15%Sonstiges 4%
Öffentlicher Sektor
Sonstige Industrie
42%
39%
15%
4%
0% 5% 10% 15% 20% 25% 30% 35% 40% 45%
Entwicklung und Umsetzung vonBranchenlösungen
Strategie-, Management undOrganisationsberatung
IV-Beratung
Sonstiges
revenuerevenue
0
50
100
150
200
250
300
88 89 90 91 92 93 94 95 96 97 98
employeesemployees
0
100
200
300
400
500
600
700
800
900
1000
88 89 90 91 92 93 94 95 96 97 98
DWH001.PPT; 7/13/99 14:09 ; Seite 4© Mummert + Partner Unternehmensberatung AG
The target leading combination of business- and specialistknow-how.
Banks - Investment Banking - Retail Banking - Asset Management- Banking Services
Insurances - Customer Relation - Market/Product/Sales
- Controlling - Insurance architecture
- Financial architecture - Asset Management - Business Solutions
Federal and stateauthorities
- Management and organizational consulting - Financial systems
- Facility- and Assetmanagement
Public sector - Financial systems - Logistic - Facility- and Assetmanagement
Energy - Customer management - Logistic
- Finanzsysteme - Facility- und Assetmanagement
Telecommunication - Customer Relation - Debitor management
Trade - Customer management – Logistic - Financial systems
Industry - Customer management - Construction (CAX-Technologien)
- Logistic - Financial systems
- Data Warehousing
- Document Management
- CRM-Technology
- Quality Management
- Software Factory
- Technology Management (Infrastructur)
- Human Ressources Management
DWH001.PPT; 7/13/99 14:09 ; Seite 5© Mummert + Partner Unternehmensberatung AG
We are where our customers need us.
BerlinFrankfurtHamburgCologneLeipzigLondonMilanMunichMünsterNew YorkViennaZurich
DWH001.PPT; 7/13/99 14:09 ; Seite 6© Mummert + Partner Unternehmensberatung AG
Content
• Mummert + Partner Consultancy AG
• Why Data Warehouse Projects with Mummert + Partner
• The Mummert + Partner - Service profile for Data Warehouse Projects
DWH001.PPT; 7/13/99 14:09 ; Seite 7© Mummert + Partner Unternehmensberatung AG
Extensive initial solutions ensure the project success
1. Specialized procedural model
2. Proven methodology
3. Certified product know how
4. Data Warehouse laboratory
5. Skilled and experienced employees
6. Successful customer relationship
On the basis of these 6 components, Mummert + Partner is the competent partner for Data Warehouse projects.
DWH001.PPT; 7/13/99 14:09 ; Seite 8© Mummert + Partner Unternehmensberatung AG
1. The M+P procedural model for Data Warehouse projects
Operation/ maintenance
Helpdesk AutomationVerificationScalability
Capacity Tuning BackupChange-Manage-
ment
Modeling DataWarehouse
DataflowData MartsMetadata Frontend Architecture
PrototypingAdvance study Targetdefinition
Is - analysis Know-howExternal
DataFocus
Economic viability
Project definitionROLAP
vs.MOLAP
3-TIER vs.2-TIER vs.
1-TIERPlaning
Teamstaffing
Product choice
ServerFAT-ClientTHIN-Client
Intranet
Implementation Network DevelopmentSecurity /
UsergroupsTesting Education
PropagationCustomizing
Datasources
DWH001.PPT; 7/13/99 14:09 ; Seite 9© Mummert + Partner Unternehmensberatung AG
2.The proven methodology is the key to success.
Frontend
Data Marts
Data Warehouse
Data sourcesMeta-Data
Static Data-models and dynamic Data-flow models
DWH001.PPT; 7/13/99 14:09 ; Seite 10© Mummert + Partner Unternehmensberatung AG
Relational Multidimensional
OLAP-CLIENT
OLAP-SERVER
RDBMS
ETL
3. Extensive know-how for an objective product choice
DWH001.PPT; 7/13/99 14:09 ; Seite 11© Mummert + Partner Unternehmensberatung AG
4. The Data Warehouse Laboratory - for developing and prototyping
Backupserver für alleNT-Server und Clients
IBM RS6000 F401 GB RAM
60 GB40 x Windows NT WS
64 MB RAM8/4 GB
IBM SP2 SilverNode AIX
4 Prozessoren3 GB RAM360 GB
Fünf NT-ServerPentium II
512/256 MB RAM18/9 GB
DWH001.PPT; 7/13/99 14:09 ; Seite 12© Mummert + Partner Unternehmensberatung AG
5. Professional technical and methodical know -how.
Data Warehouse - Team
Data Warehouse - Experts
System analysts
Graduates
Technology management
Producer
24 MA
6 MA
Business consultants
6 MA
Composition of the team
Goal: 60 employees up to the middle of 1999
More than 140 personal years of
Data Warehouse experienceMummert + Partner
1 1 2 49 10
172121
25272931313336
05
10152025303540
Mar-98
Apr-98
May-98
Jun-98
Jul-98
Sep-98
Oct-98
Nov-98
Dec-98
Jan-99
Feb-99
Mar-99
Apr-99
May-99
Jun-99
Jul-99
DWH001.PPT; 7/13/99 14:09 ; Seite 13© Mummert + Partner Unternehmensberatung AG
6. Successful projecthandling ensures successful customer relationship
DWH001.PPT; 7/13/99 14:09 ; Seite 14© Mummert + Partner Unternehmensberatung AG
Content
• Mummert + Partner Consultancy AG
• Why Data Warehouse projects with Mummert + Partner
• The Mummert + Partner - Service profile for Data Warehouse projects
DWH001.PPT; 7/13/99 14:09 ; Seite 15© Mummert + Partner Unternehmensberatung AG
Project handling or alternativly consulting single services
Operation/ maintenance
Modeling
Advance studies
Project definition
Implementation
Helpdesk AutomationVerificationScalability
Capacity Tuning BackupChange-Manage-
ment
DataWarehouse
DataflowData MartsMetadata Frontend Architecture
PrototypingTarget
definitionIs - analysis Know-how
ExternalData
FocusEconomic
viability
ROLAPvs.
MOLAP
3-TIER vs.2-TIER vs.
1-TIERPlaning
Teamstaffing
Product choice
ServerFAT-ClientTHIN-Client
Intranet
Network DevelopmentSecurity /
UsergroupsTesting Education PropagationCustomizing
Datasources
DWH001.PPT; 7/13/99 14:09 ; Seite 16© Mummert + Partner Unternehmensberatung AG
Targetdefinition
• Activities• Evaluation of the strategic areas and their dimensions• Definition of the expected targets from the economic point of view• Definition of the necessary controlling instruments to reach the target• Holding „strategy-workshops“ with the concerning departments
• Results• Documentation of demand concerning the subject• Model of a satisfactory information demand (high level)• Prioritization of set targets• Prioritization of controlling instruments that have to be created• Hierarchical dimension of the business
Advance study:
DWH001.PPT; 7/13/99 14:09 ; Seite 17© Mummert + Partner Unternehmensberatung AG
IS- analysis
• Activities· Analyzing existing operational systems· Analyzing existing dispositive systems· Analyzing the current user architecture
• Results· Grade of covering the information demand by data of existing systems· Expanding the targets by the integration of existing dispositive systems and reporting· Consolidation of necessary controlling instruments to reach the target
• Productcriteria• Accessibility to: Flat Files, VSAM, IMS, DB2, Oracle, Informix, Sybase, Adabas, Ingres,
SAP-R/2, SAP-R/3, Tandem-NonStop-SQL, NCR Teradata, MS-SQL-Server, MS-Excel,MS-Access, OBDC
Covering - preceding Systems
Advance study:
DWH001.PPT; 7/13/99 14:09 ; Seite 18© Mummert + Partner Unternehmensberatung AG
External Data
Covering - prec.Systeme - external sources
Advance study:
DWH001.PPT; 7/13/99 14:09 ; Seite 19© Mummert + Partner Unternehmensberatung AG
Know-how
Covering: - prec. Systems - external Sources - Know-how (System-Input)
Advance study:
DWH001.PPT; 7/13/99 14:09 ; Seite 20© Mummert + Partner Unternehmensberatung AG
Advance study: Focussing
•
DWH001.PPT; 7/13/99 14:09 ; Seite 21© Mummert + Partner Unternehmensberatung AG
PrototypingAdvance study:
•
DWH001.PPT; 7/13/99 14:09 ; Seite 22© Mummert + Partner Unternehmensberatung AG
Economic viabilityAdvance study:
•
DWH001.PPT; 7/13/99 14:09 ; Seite 23© Mummert + Partner Unternehmensberatung AG
ROLAP vs. MOLAPProject definition:
• Activities• Working out the necessary dimensions and their hirarchical order (drill down, slicing and
dicing, aggregation)• Evaluating the expected data volumina by regarding the sparsity• Discussing the MOLAP – attempt (multidimensional OLAP)• Discussing the ROLAP – attempt (relational OLAP)• Discussing the HOLAP – attempt (mixed use)
• Results• Basic decision about the usuable data architecture for the data marts and for the data
warehouse which will be build up in later steps• Productcriteria
• Support of MOLAP, ROLAP and/ or HOLAP• Support of personal OLAP• Export possibility to MS-Office• Add-in functionality for MS-Office
MOLAP ROLAP HOLAP
Operative Data
Data Warehouse
Data Marts
Personal OLAP
DWH001.PPT; 7/13/99 14:09 ; Seite 24© Mummert + Partner Unternehmensberatung AG
3-TIER vs. 2-TIER vs. 1-TIER
DB-Server
OLAP-Server
Client
Project definition:
DWH001.PPT; 7/13/99 14:09 ; Seite 25© Mummert + Partner Unternehmensberatung AG
FAT-Client vs. THIN-Client / IntranetProject definition:
•
DWH001.PPT; 7/13/99 14:09 ; Seite 26© Mummert + Partner Unternehmensberatung AG
ServerProject definition:
DWH001.PPT; 7/13/99 14:09 ; Seite 27© Mummert + Partner Unternehmensberatung AG
Product choiceProject definition:
• Activities· Assessment of the existing technologies on the market regarding the necessary data-
and C/S-arcitecture to reach the target· Consideration of the company’s demands regarding the product choice· Holding user-workshops to find out about the acceptance regarding possible front-end
design• Results
· Confirmation of the data- and C/S-arcitecture in case it had to be revised· Making a principle decision about the products that will be used· Making a principle decision about the sort of user-surface
• Productcriteria· Possibilities of combining the products between each other· Integration of the products/modules over common metadata· Maximum amount of transformed dimensions and hierarchical levels· Reputation of the producer
DWH001.PPT; 7/13/99 14:09 ; Seite 28© Mummert + Partner Unternehmensberatung AG
Team staffingProject definition:
• Data supply
Front-endDesign
Technique
Team of experts
Sponsor
customer
Qualitysafeguarding
producer
Support
Projectmanagement
DWH001.PPT; 7/13/99 14:09 ; Seite 29© Mummert + Partner Unternehmensberatung AG
PlanningProject definition:
•
DWH001.PPT; 7/13/99 14:09 ; Seite 30© Mummert + Partner Unternehmensberatung AG
Metadata Modeling:
• Activities• Determination/ fixing the repository to deposit metadata over datasources, data
warehouse, data marts as well as the data flow between all levels (staging areas)• Determination of metadata with accessibility for the user
• Results• A central active repository for every level of the system
• Productcriteria• Interface for data modeling• Storing of the transformation- aggregation and extraction rules• Storing of all the necessary operations• Storing of information about the user and his authorization on the dimension
level• Reporting of all the metadata including cross-referential-lists• Accessibility on the metadata by the user
DWH001.PPT; 7/13/99 14:09 ; Seite 31© Mummert + Partner Unternehmensberatung AG
Data Marts Modeling:
• Activities• Modeling of the Data Mart that will be developed in consideration to the agreed data
arcitecture• Consideration of aggregation and denormalization• Agreeing precisely on the data model with the help of the specialists team by holding a
navigation workshop• Documentation of the data model by using metadata from the arranged repository• Working out the cardinality and rates of compressions of the dimensions
• Results• Data model of the data mart that will be created in a denormalized form
• Productcriteria• Modeling of data in the way of ER with support of the denormalization and hierarchical
dimensions (star scheme, snowflake)• Generating the physical data storage• Automatic partition and intensification
DWH001.PPT; 7/13/99 14:09 ; Seite 32© Mummert + Partner Unternehmensberatung AG
Data sourcesModeling:
•
DWH001.PPT; 7/13/99 14:09 ; Seite 33© Mummert + Partner Unternehmensberatung AG
Data WarehouseModeling:
•
DWH001.PPT; 7/13/99 14:09 ; Seite 34© Mummert + Partner Unternehmensberatung AG
Data-flow Modeling:
• Activities• Modeling the dataflows all the way from the data-sources up into the data marts on field level• Documentation of aggregations, transformations, extractions and operations which are
necessary to fill the data marts via the data warehouse• Documentation of the conditions of integrity on field level• Joint and detailed agreement upon the data flow modeling with the specialists team• Differentiation of additive, semi-additive and non-additive business figures
• Results• Active repository to regulate the data supply
• Productcriteria• Use of the modern structured analysis or similar methods• Possibility to deposit conditions of integrity on field level
DWH001.PPT; 7/13/99 14:09 ; Seite 35© Mummert + Partner Unternehmensberatung AG
FrontendModeling:
• Activities• Modeling the frontend design with the necessary details• Agreeing upon the navigation within the system on the basis of the modeled data mart• Documentation of the data flow from the data mart to the Frontend• Agreeing upon the frontend design between the specialists team and chosen users• Considerating the possibilities of the chosen product• Considerating the possibilities of intelligent software-agents
• Results• Confirmed model of the dialog that will be developed and used
• Productcriteria• Possibilities of navigation (drill down, drill up, slicing and dicing)• Maximum amount of on- and off spreadsheet dimensions• Possibilities of exception-reporting (traffic lights function?, balanced scorecard, sw-agents)• Graphical opportunities (2D, 3D, geographic analysis)• Flexibility in the layout of reports (standard-reporting, free navigation, EIS-functionality)• Flexibility by analyzing (ranking, comparison of different periods, cumulation)• User friendliness (Power-user functionality)• Export- and impressor functionality
DWH001.PPT; 7/13/99 14:09 ; Seite 36© Mummert + Partner Unternehmensberatung AG
ArchitectureModeling:
•
DWH001.PPT; 7/13/99 14:09 ; Seite 37© Mummert + Partner Unternehmensberatung AG
NetworkImplementation:
DWH001.PPT; 7/13/99 14:09 ; Seite 38© Mummert + Partner Unternehmensberatung AG
CustomizingImplementation:
• Activities• Implementation of the chosen product in every level of the data warehouse which is supported by
the product• Using the specific procedures and technics which are given by the chosen product• Agreeing upon further creations of system components due to the development• Considerating the adopted standards which have been worked out by the internal and external
quality assurance• Considerating the referential integrity which is visible by modeling
• Results• Implementation of the chosen product• Identification of further components of the system that will have to be developed
• Productcriteria• effort for the development of surface and data supply• effort for the integration of changes• effort for refresh, append and update of the data deposit• seperating surface and data supply
DWH001.PPT; 7/13/99 14:09 ; Seite 39© Mummert + Partner Unternehmensberatung AG
DevelopmentImplementation:
• Activities• Creating the data supply routine by using the information deposed on the repository dynamically• Programming the frontend that has to be developed• Integration of data supply and surface by using standardized interfaces• Considerating the adopted standards which have been worked out by the internal and external
quality safeguarding• Considerating the referential integrity which is visible by modeling
• Results• Completed system allowing nonrestricted access to all data
• Productcriteria• Effort For the development of surface and data supply• Effort for the integration of changes• Effort for refresh, append and update of the data deposit• Seperating frontend and data supply• Interfaces to higher programming languages over class librarys• Availability of an integrated software development location
Surface-
Design
Data-
supply
Integration
DWH001.PPT; 7/13/99 14:09 ; Seite 40© Mummert + Partner Unternehmensberatung AG
Security/ UsergroupsImplementation:
• Activities• Defining all the dimensions which are relevant for security• Defining the usergroups in collaboration with the specialists team• Implementation of security informaion in the repository• Expanding the routines of data access by the consideration of security information in the repository
• Results• Security administration system• Completed system by considerating the aspects of security• Concept for different usergroups
• Productcriteria• Considerating power-user as well as end-user• Variable classification of users to different user groups• Taking care of security information by the concerning department
DWH001.PPT; 7/13/99 14:09 ; Seite 41© Mummert + Partner Unternehmensberatung AG
TestingImplementation:
• Activities• Definition of error categories (e.g. data supply, surface, technic)• Definition of error classes /grades (e.g. 1= serious error...4= nice to have)• Defining equivalence classes to reach a high functional covering by the test• Carrying out the test to eliminate errors steadily• Carrying out stress- and masstests above the equivalence classes
• Results• Approving the system
• Productcriteria• Debugging- functionality• Automatical generation of test data
DWH001.PPT; 7/13/99 14:09 ; Seite 42© Mummert + Partner Unternehmensberatung AG
EducationImplementation:
•
DWH001.PPT; 7/13/99 14:09 ; Seite 43© Mummert + Partner Unternehmensberatung AG
PropagationImplementation:
•
DWH001.PPT; 7/13/99 14:09 ; Seite 44© Mummert + Partner Unternehmensberatung AG
HelpdeskOperation/ maintenance:
DWH001.PPT; 7/13/99 14:09 ; Seite 45© Mummert + Partner Unternehmensberatung AG
AutomationOperation/ maintenance:
• Activities• Creating automatical job-processes for the data supply of the data marts• Creating automatical job-processes for the supply of data of the client-system (pull-principle)• Considerating available time frames• Handing over the data supply routines to the productive system of the customer
• Results• Productive system
• Productcriteria• Possibility of automation of the data supply- and data spreading routines• Considerating recommencement in case of an interruption of the data supply• Query automation of integrity conditions as well as free eligible filters• Exception-handling by proactive presentation of critical raw data• Independent data supply of the clients• Possibilities of partial loading of the system in case one source is not available
•
DWH001.PPT; 7/13/99 14:09 ; Seite 46© Mummert + Partner Unternehmensberatung AG
Change-ManagementOperation/ maintenance:
•
DWH001.PPT; 7/13/99 14:09 ; Seite 47© Mummert + Partner Unternehmensberatung AG
CapacityOperation/maintenance:
• Activities• Finding out about the expected need of store capacity• Considerating given key figures due to the modeling like
-amount of dimensions as well as their cardinality-amount of hierarchies per dimension as well as their rates of compression-amount of different features per hierarchy-in case of MOLAP solutions, the amount of non usable spaces (sparsity)-in case of ROLAP solutions, length of a line of the central entity-type
• Consideration of the additional requirement due to the indication• Consideration of the additional need due to aggregation
• Results• Qualified projection of the required capacity of the system
• Productcriteria• Algorithm for the determination of capacity requirements
DWH001.PPT; 7/13/99 14:09 ; Seite 48© Mummert + Partner Unternehmensberatung AG
TuningOperation/maintenance:
• Activities• Recommendation of aggregations for standard analysis• Recommendation of indices for frequent used analysis• In case of higher data volumina (>100 GB) implementation of parallel architectures• Moving data from the data mart to the client ( subsetting, personal OLAP)• Dismanteling and parallelising the raw data supply• Partitioning of the data deposit
• Results• Improvement of the running system
• Productcriteria• Automated logging on statement level• Automated recommendation of aggregations and index• Log-management-system to evaluate the use of the system• Possibility of using bitmap-index• Effort for placing an index
DWH001.PPT; 7/13/99 14:09 ; Seite 49© Mummert + Partner Unternehmensberatung AG
BackupOperation/maintenance:
•
DWH001.PPT; 7/13/99 14:09 ; Seite 50© Mummert + Partner Unternehmensberatung AG
Verification/Scalability
PredictionDeviationAnalysisClassification
SegmentationAssociation
•
Operation/maintenance: