Download presentation
Presentation is loading. Please wait.
2
Content Semantic Information Integration
crossvision Information Integrator Use Case Customer Information Experiences Waterfeld / / 2
3
Getting Information customer sales Support Management existing manual
needed Information in vielen verschiedenen Systemen Oft nur 1 existierende Innerhalb Firma noch irgendwie manuell Aber bei Kunden geht es gar nicht Legacy CRM ERP SFA logistics Waterfeld / / 3
4
Information heterogeneity
Heterogenous data sources Technical heterogeneity Different formats Incomplete data Redundant data Different access system Different user interfaces Different Semantic Formate fängt an bei Dezimalzahlen . Und , Datumsformate, Unvollständige Daten, für 1 Kunde Telefon, für anderen pro Informationsquelle noch mehr pro Informationsquelle: Redundante Daten: Änderung nur an 1 Stelle Verschiedene Bedeutung: sehr häufig, anspruchsvolles Problem Waterfeld / / 4
5
Semantic Information Integration
Current technologies EAI ETL (data warehouse) Information Integration Direct access Integrated information model Semantic Information Integration Ontologies as integrated information modell Ontologies Mapping Mapping Mapping Waterfeld / / 5
6
Layered Ontology Approach
views integration ontology manual mappings automatic mappings source ontologies sources <article> <articleid>a-5634</articleid> <category>printer</category> <name>hp81</name> <price currency=‘USD’>500</price> <producer>hp</producer> <resolution>1960 dpi</resolution> <type>laser</type> …. </article> Waterfeld / / 6
7
crossvision Information Integrator
Studio (Eclipse-based) Web Service Semantic Server CentraSite Repository Query Queries Ontologies Load Ontologies Mapping Metadata User Security WebDAV JAXR UDDI XML Tables WSDL ADA Waterfeld / / 7
8
Usage Information Integrator
Application Query Builder Deploy Web Service End Web Service Integrator Studio Integration Ontology Semantic Server Source Ontology Mapping to Source Ontologies Import Metadata Start Metadata Waterfeld / / 8
9
Use Case: Customer Information Gateway
Does a customer raising a support issue for a specific product have a valid maintenance contract for that product? What is a customer‘s experience with our products? Adabas Customer Information System SQL Support Information System Situation at beginning of project 2 systems consisting of databases and applications Sometimes user mainly working with one system, e.g. support engineer, is interested in questions that can only be answered by the other system. So, systems are somehow related to each other, somebody knows, how to combine data but this is not automatically available client contact information contract customer contact information support case Waterfeld / / 9
10
Customer Information Gateway – Source ontologies
Class „File151“ property „AA“ type integer property „AB“ type date property „AC“ type date property „AD“ type string … Class „File87“ Class „CUSTOMER“ property „cid“ type integer … Class „CASE“ property „desc“ type string property „cust “ type integer property „belongsTo“ type „CUSTOMER“ Customer Information System (Adabas) Support Information System (SQL) File 151 Fields „AA“, „AB“, „AC“, „AE“ File 87 Fields „AA“, „AB“, „AC“ DATA SOURCE IMPORT Source ontologies Contain rule with built-in predicate which implement access to external data sources Translation of foreign keys No such foreign keys with adabas Open issue: user access information Table „CUSTOMER“ Table „CASE“ Waterfeld / / 10
11
Automatic Source ontology mapping
FORALL X,Y c(“FILE151”, X):FILE151_CONTRACT[FILE151_AA->X; FILE151_AE->Y] <- accessToSource(connectionInfo, “FILE151”, “AA”, X, “AE”, FORALL X, Y, Z c(“FILE152”, X):FILE152_CLIENT [FILE152_AA->X; FILE152_AB->Y; FILE152_AC->Z] <- accessToSource(connectionInfo, “FILE152”, “AA”, X, “AB”, Y, “AC”, FORALL X, Y, Z c(“CUSTOMER”, X):CUSTOMER[id->X; name->Y; addr->Z] <- accessToSource(connectionInfo, “CUSTOMER”, “id”, X, “name”, Y, “addr”, Z) FORALL X,Y,Z c(“CASE”, X):CASE[caseId->X; status->Y; customerId -> Z] <- accessToSource(connectionInfo, ”CASE”, “caseId”, X, “status”, Y; “customerId”, Z) Waterfeld / / 11
12
Integration Ontology Class Customer Class Contract
property id property contractId property name property contractEndDate property address property contractor Class SupportRequest property id property status property issuedBy File 151 -> Contract AA -> contractId AB -> contractStartDate AC -> contractEndDate AD -> contractStatus File 87 -> Client … Class „File151“ Class „File 87“ Class „CUSTOMER“ Class „CASE“ Mapping source ontologies to business ontologie Using application/designer know-how Adabas SQL Waterfeld / / 12
13
Integration Ontology Mapping
Class and property mapping FORALL X, Y, Z X:Contract[contractId->Y; <- X:FILE151_CONTRACT[ FILE151_AA->Y; Object property mapping FORALL X, Y, Z1, Z2 X:SupportRequest[issuedBy->c(“Customer”, Z1, <- and c(“CUSTOMER”,Y)[name->Z1; Waterfeld / / 13
14
Customer Information Gateway - Mapping
Class Customer Class Contract property id property contractId property name property contractEndDate property address property contractor Class SupportRequest property id property status property issuedBy Object identity Data type transformations Encoding of values Generation of object references Generation of inverse references N:M relations Class „File151“ Class „File 87“ Class „CUSTOMER“ Class „CASE“ Mapping issues Surrogate values might be but often are no good identifiers (SSN, SAP numbers) Waterfeld / / 14
15
Mapping tool Waterfeld / / 15
16
Customer Information Gateway - Queries
For given SupportRequest search for the corresponding Contract Class Customer Class Contract property id property contractId property name property contractEndDate property address property contractor Class SupportRequest property id property status property issuedBy Use one language for modeling and querying Queries for data access (like database queries) Explanations (only) if it does not work. Waterfeld / / 16
17
Experiences: Problems
F-Logic functionality Basic datatypes (primitive XML Schema types) Mapping F-Logic to XML Processing of nested hierarchical lists in F-Logic order Null values User adoption No rule language standard F-Logic syntax too cryptic, not keyword-oriented Only possible via tools e.g. graphical rules editor No accepted ontology and rule model API Utilizing query capabilities of data sources Stratification needed due to recursive subclass definition Waterfeld / / 17
18
Benefits Query language as sublanguage of rule language
High expressiveness of rule language for ontology mapping Adding additional knowledge to integration ontology Ontologies as executable models for information integration Waterfeld / / 18
Similar presentations
© 2025 SlidePlayer.com. Inc.
All rights reserved.