1 / 41

Master Thesis Presentation Florian Niedermann, 07. February 2008

Development of a method for the integrated handling of process variables and operational data sources. Master Thesis Presentation Florian Niedermann, 07. February 2008. Agenda. Introduction and Motivation The Linking Method Editor Functionality Evaluation Review & Outlook. Motivation.

rian
Download Presentation

Master Thesis Presentation Florian Niedermann, 07. February 2008

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Development of a methodfor the integrated handling of process variables and operational data sources Master Thesis Presentation Florian Niedermann, 07. February 2008

  2. Agenda • Introduction and Motivation • The Linking Method • Editor Functionality • Evaluation • Review & Outlook Florian Niedermann

  3. Motivation PROBLEM • Separate storage of audit trail data and other data sources • Certain questions can‘t be answered • Some examples: • Factors influencing process durations • Reasons for cancelled processes Florian Niedermann

  4. Process Designer Operational Data Source Designer or Business Warehouse Designer Business Analyst Manual Match Description Direct BA Manual Match Description Direct PD Manual Match Description in Workflow Extension DB Manual Match Description in Workflow Extension Link Semantic Annotation of Workflow Data Semi-automatic Semantic Annotation of Operat. Data Semi-automatic Manual Match Description Direct PD Semantic Annotation of Workflow Data Semi-automatic Recognizer Automatic Ontology Semantic DDL & DML Parser Extension DB Workflow Parser Extension Link Reasoner Semantic dBDD The BIA Framework Florian Niedermann

  5. Agenda • Introduction and Motivation • The Linking Method • Editor Functionality • Evaluation • Review & Outlook Florian Niedermann

  6. Possible “annotatable” elements • interesting elements for analysis: • Process • automated selection (based on Semantic description) • limited use  Not pursued • Operation • querying audit trails for executed operations (based on Semantic description) • no help with linking  Not pursued • Variables • Link process data to operational data sources  must be included (see next slide) Florian Niedermann

  7. Annotation granularity of variables • Three alternatives exist: • Top-level; Only the variable • Message parts • Whole XSD data structure • (1) and (2) too simplistic • (3) is used, even though it‘s more difficult to implement Florian Niedermann

  8. OWL-S and WSMO • OWL-S and WSMO are similar in several respects: • tightly linked to a specific ontology language (OWL and WSML respectively) • own conceptual representation of Web services • Grounding: Match between conceptual representation and Web service description (e.g. WSDL file) • Powerful but complex frameworks for the semantic description of Web services Florian Niedermann

  9. SAWSDL • No own conceptual representation • Extension of WSDL by adding additional attributes (see next slides) • Usage together with arbitrary ontology languages • Easy to use • Quite good extension(just add new attributes) • useable for WSMO and OWL-S grounding An easy-to-use, flexible approach to the annotation of Web services and XML schema definitions SAWSDL is used as the base for the linking method Florian Niedermann

  10. Semantic annotation with SAWSDL • modelReference • References ontology concepts • schemaLiftingMapping • Usually references a file • Transforms XML elements into ontology instances • schemaLoweringMapping • Usually references a file • Transforms ontology instances into XML elements Florian Niedermann

  11. Extensions to the SAWSDL standard • Semantic Link • modelType Denotes ontology language (WMSO or OWL) • Direct Data Link • source Directly references a data source (e.g. a table or a column) • sourceType Indicates type of the data source (so far: relTable or relColumn) Florian Niedermann

  12. Linking method example Florian Niedermann

  13. Agenda • Introduction and Motivation • The Linking Method • Editor Functionality • Evaluation • Review & Outlook Florian Niedermann

  14. Florian Niedermann

  15. 1. Step: Set up connection to audit database, BPEL engines, ontologies, processes etc. Florian Niedermann

  16. 2. Step: Select the variable in the artefact explorer • Displays BPEL processes and XML documents • Expanding a process shows its variables Florian Niedermann

  17. Step 3: Select the node to be annotated in the annotation explorer • tree-style visualization of variables and documents • XML documents: visualizing the JDOM model • Variables: more difficult (see next slide) • annotation is restricted to certain nodes (per default) Florian Niedermann

  18. Step 3: Technical background • Data structure of BPEL variables corresponds to either • XSD type • XSD element • WSDL message • Reference is made by namespace and name • To resolve references, the editor internally keeps hash mapsstoring all types, elements and messages by their namespace Florian Niedermann

  19. Step 4.1: Enter the node‘s annotation with the attribute table • Shows “raw” annotation attributes • Manipulation of their values possible  easier to use graphical editors (see next slides) Florian Niedermann

  20. Step 4.2: Provide the semantic link with the ontology explorer • Tree-style visualization of ontologies • WSMO (WSML) ontologies: WSMO4J • OWL ontologies: OWL API of the Jena framework • “Set Model Reference” sets the values of the “modelReference” and “modelType” attribute Florian Niedermann

  21. Step 4.3: Provide the direct data link with the data connection explorer • Tree-style visualization of a database • Realization by a special query (see next slide) • “Set Data Source” sets the values of the “source” and “sourceType” attributes Florian Niedermann

  22. Step 4.3: Technical background Sample Oracle data dictionary query: SELECT t.TABLESPACE_NAME AS TABLESPACE_NAME, t.TABLE_NAME AS TABLE_NAME, c.COLUMN_NAME AS COLUMN_NAME, c.DATA_TYPE AS COLUMN_TYPE, c.DATA_LENGTH AS COLUMN_LENGTH FROM USER_TABLES t, USER_TAB_COLUMNS c WHERE t.TABLE_NAME = c.TABLE_NAME ORDER BY TABLESPACE_NAME, TABLE_NAME, COLUMN_NAME Florian Niedermann

  23. Saves all open tabs Saves the currently open tab Step 5: Save the annotated variable to the audit database Florian Niedermann

  24. Step 5: Technical Background Florian Niedermann

  25. Editor architecture Florian Niedermann

  26. Agenda • Introduction and Motivation • The Linking Method • Editor Functionality • Evaluation • Review & Outlook Florian Niedermann

  27. Evaluation background • The evaluation is qualitative • No quantitatively sound evaluation; requires empirical studies • The evaluation is largely based on an own developed case study for a car rental company Florian Niedermann

  28. Editor Evaluation Completely covers the linking method Easy to use Supports both OWL and WSML Can work with nearly arbitrary relational databases Performance issues when loading large ontologies Direct data link geared towards processing in Java + + + + – – Florian Niedermann

  29. Direct link evaluation Simple matches (especially ID attributes) Onlyannotation of process variables necessary Process designer needs to know the operational data sources + – + Florian Niedermann

  30. Semantic link evaluation Complex matches Matches with a priori unknown data sources + Requires reasoner Requires ontology Bothoperational data sources and variables need to be annotated – – + – Florian Niedermann

  31. A possible problem with both approaches... • Linking attributes used for joins between process data and operational data sources • Key assumption: One same real-world entity will have the same attribute values in both process data and operational data sources. • Possible scenario: Employee “ABC” has employee ID • 12345 in operational data sources • X-ABC-129 in process data Florian Niedermann

  32. Agenda • Introduction and Motivation • The Linking Method • Editor Functionality • Evaluation • Review & Outlook Florian Niedermann

  33. Overview of work done in the thesis • BPEL engine market study • Linking method • Annotation editor • Editor architecture • Java implementation • Manuals • Car Rental GmbH case study • Business background • Datawarehouse (populated with generated data) and BPEL processes Florian Niedermann

  34. Next Steps • Semantic annotation of operational data sources / DWH( Studienarbeit) • BIA ontology ( Diplomarbeit) • Development or selection of a reasoner • Consider standardization of audit data • Think of how to use the annotation data on an instance level Florian Niedermann

  35. Backup Florian Niedermann

  36. The Toolbar allows for quick access to most Functions (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) • Starts a wizard to add a new BPEL process to the workspace • Edits the currently selected process or document • Opens an XML document and adds it to the workspace • Removes the currently selected process / document • Saves the current tab open in the annotation explorer • Saves all currently opened tabs • Reloads the currently selected document / variable • Starts a wizard used to create, edit and delete data connections (see next slides) • Starts a wizard used to create, edit and delete ontologies • Starts a wizard used to create, edit and delete BPEL engines • Brings up the editor settings • Shows the “About” dialog Florian Niedermann

  37. Data Connection Wizard (1/5) Florian Niedermann

  38. Data Connection Wizard (2/5) Florian Niedermann

  39. Data Connection Wizard (3/5) Florian Niedermann

  40. Data Connection Wizard (4/5) Florian Niedermann

  41. Data Connection Wizard (5/5) Florian Niedermann

More Related