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.



Published on

Published in: Technology
  • Be the first to comment

  • Be the first to like this


  1. 1. Enterprise Data Governance Reference and Master Data Management Semantic Modeling By Pierre Bonnet - 2010, May Creative Commons To order the book:
  2. 2. What we deliver Open community working on MDM+BRMS and the IS Rating Tool Open Community working on Semantic Data modeling and EA applied to MDM ² Software Vendor in Model Driven MDM and Data Governance
  3. 3. MDM… Why?
  4. 4. Where are your ref/master data? A product (Database No. 1) is only created if the factory code (Database No. 2) which it is affiliated with is under the management of an organization (Database No. 3) that has an ongoing production agreement (Database No. 4) with the company or one of its partners (Database No. 5), that has been active for the past twelve months, at its disposal Database No.1 Database No.2 Database No.3 Database No.4 Database No.5 ? Product Factory Organization Agreement Partner
  5. 5. Solution: MDM and Data Governance approach Database No.1 Database No.2 Database No.3 Database No.4 Database No.5 MDM A product (Database No. 1) is only created if the factory code (Database No. 2) which it is affiliated with is under the management of an organization (Database No. 3) that has an ongoing production agreement (Database No. 4) with the company or one of its partners (Database No. 5), that has been active for the past twelve months, at its disposal Master Data Integration (EAI/ESB/ETL) Rich and shared master data model Product Factory Organization Agreement affiliated managed role 2 role 1 Data Steward Product Factory Organization Agreement Partner
  6. 6. Three axis for every data element
  7. 7. Data Governance: version management Version management: compare and merge your data versions
  8. 8. Data Governance: contexts management Contexts management: data authoring by uses contexts Permissions delegation Inheritance of values Corporate Partner 1 Partner 2 Threshold = [10, 100] Threshold = [20, 90] Threshold = inherited Agency Web Threshold = [25, 50] Partner Call-center Web Corporate Sales Channels = use context
  9. 9. Data Governance: other key features <ul><li>Ability to customize default UI </li></ul><ul><li>Permissions management </li></ul><ul><li>Data history management </li></ul><ul><li>Data hierarchy management </li></ul><ul><li>Data approval workflow </li></ul><ul><li>Data archiving </li></ul><ul><li>Data query and reporting </li></ul><ul><li>Data web services </li></ul><ul><li>../.. </li></ul>
  10. 10. Which data quality is the right one? Level 1 – BASIC DATA QUALITY <ul><li>Sticking cleansing tools on existing silos and databases </li></ul><ul><li>Lack of data governance </li></ul>Level 2 – EXTENDED DATA QUALITY MDM <ul><li>Data repository with limited data governance functions as it doesn’t rely on a semantic data model. Its data model has a similar structure to existing databases schemas </li></ul>Level 3 – TOTAL DATA QUALITY MDM <ul><li>Data repository based on a rich data model independent of existing databases schemas. It brings all data governance functions aligned with the semantic data modeling </li></ul>Data Steward Low Medium Optimized
  11. 11. Model-driven lives with OLTP systems Model-driven MDM Rich data model over OLTP RDBMS Business objects, lifecycles, validation rules, permissions Governance time CDI (OLTP system) Rigid data model in OLTP RDBMS Data Injection To data governance Run time Consolidation and data enrichment Transactional IT System RDBMS Tables RDBMS Tables Data Steward Usual users of the IT system All data governance functions are automatically driven by the rich data model
  12. 12. Towards what IS transformation?
  13. 13. IS transformation path: stage 1 R/M Functional silo R/M Functional silo R/M Functional silo MDM Ref/Master data Unified data model Step 1. MDM is synchronized with existing databases Data governance Referential/Master data R/M Transactional data
  14. 14. IS transformation path: stage 2 R/M Functional silo R/M Functional silo R/M Functional silo MDM Ref/Master data Unified data model BRMS Business Rules Step 2. BRMS is used for extracting business rules locked in silos Data and Rules governance Referential/Master data R/M Transactional data
  15. 15. IS transformation path: stage 3 New system Functional silo R/M Functional silo MDM BRMS BPM Business Rules Processes Master Data R/M IS governance (Data, Rules, Processes) Referential/Master data R/M Transactional data
  16. 16. Linking value: MDM+BRMS+BPM source: SMABTP insurance company
  17. 17. MDM… Which modeling procedures? Caveats: this presentation doesn’t tackle organization issues when deploying and managing a MDM. To get information about this governance topic, please read the book (chapter 7) and/or visit our website
  18. 18. Which MDM is the right one? Virtual MDM Static MDM DANGER Semantic MDM Richness and expressiveness Level of modeling Semantic data modeling Classic data modeling IT Business Ambiguous
  19. 19. Semantic MDM gains: example 1 MDM ../.. 1 1 Partner 5 1 Subsidiary 1 0 Head Office Cardinality (p) Cardinality (n) Use context Produit Usine * on use context (n,p ) Product Factory
  20. 20. Semantic MDM gains: example 2 Semantic modeling Author Book * * writes <ul><li>This (very simple) semantic data model must be translated into a logical data model runnable by the MDM software </li></ul><ul><ul><li>either via join table (static MDM) </li></ul></ul><ul><ul><li>or via an object-oriented approach relying on a MDA architecture </li></ul></ul>
  21. 21. Semantic MDM Static MDM => entails hard-coding to hide join tables => what you model is what you get Join Table (IT point of view) Logical data model with join table AuthorBook FKAuthor : real FKBook : real Author Book PKAuthor : real PKBook : real Logical data model without join table (semantic is maintained) Model-driven: the screen is automatically generated from the rich data model Author Book PKAuthor : real PKBook : real lstFKAuthor[*] : real
  22. 22. Method: mustn’t be dedicated to the MDM Regardless of any specific method, at the very least these key points must be enforced to succeed in MDM projects: <ul><li>Semantic modeling </li></ul><ul><li>Enterprise Data Architecture </li></ul><ul><li>Logical data modeling </li></ul><ul><li>… and that’s all </li></ul>Praxeme example – figure from Praxeme Institute (creative commons) see detailed information in the book
  23. 23. Enterprise Data Architecture Domain of business objects Domain of business objects Domain of administrative objects Domain of administrative objects Category Category Category Category Business Object Business Object Business Object Category Category Business Object Business Object Business Object Business Object Data form Foreign Key Data form Administrative Object Administrative Object Administrative Object Domain of business objects Domain of business objects Domain of business objects Domain of business objects
  24. 24. Data Category Data category Client Data category Order Tightly coupled Loosely coupled Classification PKClient : real FKOrder[*] : real Classification : Classification Preference[*] : Preference Preference Client Order <<comment>> Linked with nested data type <<comment>> Linked with a list of nested data types <<comment>> Linked with foreign keys (FK)
  25. 25. Enterprise Data Architecture
  26. 26. Let's take a quick example Business object: Address
  27. 27. This is not a semantic data model <ul><li>lack of referential integrity constraints </li></ul><ul><li>no business lifecycle, i.e. no state attribute whereas this is an important ref/master data attribute </li></ul><ul><li>very poor expressivity as this data model is a flat point of view (data flows) </li></ul>If this data model is used to manage Address within the MDM… this is the end of MDM Figure stemming from OASIS
  28. 28. Semantic modeling – Address stage 1
  29. 29. Semantic modeling – Address stage 2
  30. 30. Semantic modeling – Address stage 3 Business lifecycle (state machine)
  31. 31. Decision table applied to Address Carried out by the MDM automatically when checking data modifications of Address. It means less approach of hard-coding software and more auditability and traceability of integrity constraints applied to ref/master data MDM Stored into the MDM
  32. 32. Which project management? It depends on the maturity level of the MDM tool <ul><li>either a nightmare with significant risks of MDM collapse </li></ul><ul><li>or easier way relying on Model-driven approach </li></ul>
  33. 33. see the book
  34. 34. How to drive and measure the IS transformation based on MDM?
  35. 35. How to assess the IS Value? <ul><li>Use Value </li></ul><ul><ul><li>Deals with an assessment of company’s working procedures used by IS stakeholders to communicate, build, maintain and run their Information Systems </li></ul></ul><ul><li>Business Value </li></ul><ul><ul><li>Regardless of working procedures, it measures the ability to align the IS with business strategy and operational requirements. It also includes financial measurement </li></ul></ul><ul><li>Intrinsic Value </li></ul><ul><ul><li>It is the actual value of IS Assets regardless of its Use Value and its Business Value, based on ref/master data, business rules and processes </li></ul></ul>ISO900x, CobiT, CMMI, PRINCE2, ITIL,TOGAF , etc. CobiT, ITIL, IAS-IFRS, etc.
  36. 36. <ul><li>Each IS Asset is gauged through a set of questions grouped by three facets </li></ul><ul><ul><li>Knowledge, Governance Features, IT Approach </li></ul></ul>IS Rating Tool matrix Note: The governance features facet is focus on a list of functions required to well manage Data, Rules and Processes such as Version management, Authoring, Traceability, etc. This is not the quality of processes applied to conduct changes and relationships between Executives, Business Users and IT department.
  37. 37. <ul><li>Each cell of the IS Rating matrix is defined through a list of questions to measure the IS Assets </li></ul>IS Rating Tool matrix <ul><li>Five marks levels are available for each question </li></ul><ul><ul><li>Bad: 0% to 15% </li></ul></ul><ul><ul><li>Basic: 15% to 30% </li></ul></ul><ul><ul><li>Medium: 30% to 50% </li></ul></ul><ul><ul><li>Advanced: 50% to 80% </li></ul></ul><ul><ul><li>Optimized: 80% to 100% </li></ul></ul>
  38. 38. <ul><li>A consolidated rating is obtained in the form of three letters </li></ul><ul><ul><li>Each letter gives a consolidated Performance Level Percentage (PLP) regarding your Data, Rules and Processes respectively </li></ul></ul>IS Rating approach Letter #1 Letter #2 Letter #3
  39. 39. IS Rating example – 1/4 This company has a huge risk in Data management (C) and Rules management (C). Its ability to well manage its Processes (B) doesn’t prevent this company from facing a serious IS crisis because of a data and/or business rules loss of control. More precisely, according to its detailed IS rating outcomes, we have to study why its data knowledge is so poor (only 5% = D)? From this first rating, the company’s stakeholders should demand detailed information describing the IS plan to manage data and rules risks. Once this plan will be delivered, a further IS rating will be done to gauge the IS improvement. May be, this company could decide to raise its data consolidated mark to the &quot;B&quot; level within a year and then attain a &quot;BCB&quot; rating which would be reassuring for its stakeholders. Then, after a full year of work on data knowledge improvement, the company’s rating goes up to &quot;BCB” (see the next rating card).
  40. 40. IS Rating example – 2/4 On the basis of this new rating, the company has a huge interest in improving its business rules management to reach a ‘B’ level rather than a poor ‘C’. The key factor to succeed seems to deploy a stronger approach in rules governance, as shown in its detailed IS rating outcomes (21% = C). After delivering works on rules management, this company achieves this objective and attains a new rating &quot;BBB&quot; (see the next rating card).
  41. 41. IS Rating example – 3/4 &quot;BBB&quot; is an excellent rating, very reassuring for all IS stakeholders. By analyzing its detailed percentage levels per IS assets (see its detailed IS rating outcomes) additional improvements can be identified to target the &quot;AAA&quot; level. Don't forget, the rating can be established both at the scale of the whole IS but also within a smaller scope such as a subsystem, a software package, a subsidiary, a web application, etc.
  42. 42. IS Rating example – 4/4 A progressive path to leverage your IS Intrinsic Value and drive your Enterprise Architecture
  43. 43. IS Rating Tool - Excerpts On the web site
  44. 44. Facets of the rating: Data, Rules, Processes. Here is the Data Assessment spreadsheet Questions to measure IS Assets How to answer Guidelines The marks from Bad to Optimized
  45. 46. IS Rating Dashboard
  46. 47. S-IT-A in the EA and ERP fields <ul><li>Lot of EA approaches but lack of a common foundation to transform IS and measure outcomes </li></ul><ul><ul><li>ACMS brings the foundation </li></ul></ul><ul><ul><li>IS Rating Tool measures outcomes instead of processes </li></ul></ul><ul><li>ERP benefit from the ACMS foundation </li></ul><ul><ul><li>Assess your ERP through the IS Rating Tool </li></ul></ul>
  47. 48. When using the IS Rating Tool? <ul><li>Audit tool used within IT departments to encourage innovation and to assess Enterprise Architectures </li></ul><ul><li>Audit tool to assess IS scopes such as departments, business units, systems, applications, etc. </li></ul><ul><li>Communication tool between IT department and all IS Stakeholders </li></ul><ul><li>Tool  for getting key indicators required to establish an annual IS Assessment published to company's Stakeholders </li></ul><ul><li>Tool used to encourage bench-learning between companies </li></ul><ul><li>Audit tool when selecting software packages such as ERP, CRM, Supply Chain, etc. </li></ul><ul><li>Audit tool to complement with financial analysis when studying and driving M&A (Mergers and Acquisitions) </li></ul><ul><li>Tool to support IS innovation based on the ACMS foundation, in other words business repositories relying on MDM, BRMS and BPM </li></ul><ul><li>Tool to drive third parties IS/IT contracts by including it within RFQ and RFP as a set of commitments </li></ul>
  48. 49. CobiT mapping with ACMS & IS Rating FOUNDATION MEASURE / CONDUCT CobiT - PO1 Define a Strategic IT Plan PO1.1 IT Value Management PO1.2 Business IT Alignment PO1.3 Assessment of Current Capability and Performance PO1.4 IT Strategic Plan CobiT - PO2 Define the Information Architecture PO2.1 Enterprise Information Architecture Model PO2.2 Enterprise Data Dictionary and Data Syntax Rules PO2.3 Data Classification Scheme
  49. 50. Excerpt: PO1.1 IT Value Management d
  50. 51. Excerpt: PO2.1 IT Enterprise Information
  51. 52. How to contribute to Sustainable IT Architecture
  52. 53. Year plan 2010 focus on IS Rating Tool <ul><li>End of June 2010 </li></ul><ul><ul><li>First official version of the IS Rating Tool </li></ul></ul><ul><li>September 2010 </li></ul><ul><ul><li>Beta one-day training IS Rating Tool </li></ul></ul><ul><li>October/November 2010 </li></ul><ul><ul><li>Annual conference S-IT-A dedicated to the IS Rating Tool </li></ul></ul><ul><li>End of 2010 </li></ul><ul><ul><li>Official one-day training and certification IS Rating Tool </li></ul></ul><ul><ul><li>First implementation within EA Tools </li></ul></ul><ul><ul><li>New book “IS Assets and EA” </li></ul></ul><ul><ul><li>New version with a derivation to business users </li></ul></ul>
  53. 54. Year plan 2010 – To contribute <ul><li>Working group IS Rating Tool </li></ul><ul><ul><li>To improve, to test, to provide feed-backs and bench-learning </li></ul></ul><ul><li>EA Tools </li></ul><ul><ul><li>We need to partner with software vendors in EA field </li></ul></ul><ul><li>Annual conference on October/November 2010 </li></ul><ul><ul><li>We are willing to welcome sponsors </li></ul></ul><ul><li>Lobbying and exchanges </li></ul><ul><ul><li>Linkedin Group (more than 900 members) </li></ul></ul><ul><ul><li>Added-values by System Integrators over the core IS Rating Tool </li></ul></ul>
  54. 55. How to participate in Sustainable IT? <ul><li>To accelerate the adoption of ACMS and IS Rating Tool we need your support </li></ul><ul><ul><li>Systems Integrators can benefit from the S-IT-A web platform to give information about their added-values based on ACMS and IS Rating Tool </li></ul></ul><ul><ul><li>User Companies can show their interests in ACMS and IS Rating Tool, publish and share lessons learnt and therefore contribute to the effort of the community </li></ul></ul>Individual contributors and testers of the IS Rating Tool are welcome (more than 100 members) To stay tuned everyone is welcome to our linkedin group (more than 900 members) complete list on our website complete list on our website
  55. 56.
  56. 57. Thank you To contact Pierre Bonnet, Founder of Sustainable IT Architecture Community [email_address] To order the book: Book by Pierre Bonnet.  ISBN: 9781848211827 Publication Date: May 2010 ISTE - Wiley Table of contents The MDM Approach 1. The Company and its Data. 2. Strategic Aspects. 3. Taking into account ERP. 4. Return on Investment. MDM for Business Users 5. MDM Maturity Model. 6. Data Governance Features. 7. Organizational Aspects. The MDM for IT Specialists 8. Key Principles of Semantic Modeling. 9. Semantic Modeling Procedures. 10. Logical Data Modeling Procedures. 11. Organization Modeling Procedures. 12. MDM-IT integration
  57. 58. You can use this document for your own needs and publications under the condition to quote the source “Sustainable IT Architecture” This document is protected by a Creative Commons (see opposite figure) You can reuse this material