[ Alexis AUBRY / Mario LEZOCHE]
Enterprise Information Systems: a proposition for conceptualisation and interoperability evaluation
Alexis AUBRYMario LEZOCHE
2010, 10th November
[ Alexis AUBRY / Mario LEZOCHE]
Contents
• Background• Enterprise Systems Models Semantic
interoperability• Proposed methodology
• Fact Oriented Semantic conceptualisation• Semantics Structure enactment: the semantic
blocks
• Future works
2
[ Alexis AUBRY / Mario LEZOCHE]
Background
• Product-centric paradigm in the Enterprise Information Systems:
3
Informational flow
Physical flowProduct
Problem of synchronisation between the informational and the physical flows
Business
Manufacturing
ProductCRAN(Morel et al., 2003)
Fully integrated system (CIM)
The product is considered as an information system (Tursi, 2009) The informational and physical flows are synchronised via the system product
From one monolithic Information System… … to a set of heterogeneous and autonomous ISs aggregated into one SoIS (Auzelle, 2009)
System of information systems (SoIS)
CRM ERP
MESSCE
APS
CRM
ERP
APS
MESSCE
[ Alexis AUBRY / Mario LEZOCHE]
Background
4
• Two constats:1. Increasing complexity of the information flows
(quantity, diversity…)
2. Distribution of the information in the whole supply chain Enterprises use an increasing number of software applications These enterprise information systems have to interoperate… … in order to achieve the global performances for the full
manufacturing processes.
Product
CRM ERP
MESSCE
APS
Semantic interoperability
[ Alexis AUBRY / Mario LEZOCHE]
Semantic interoperability
5
• Definition (Whitman et al., 2006)Two Information Systems are interoperable if they are able:i) to share specified information and,ii) to operate using that information according to a shared
semanticsiii) in order to realise a specified mission in a given context.
Main issues:
1.Studying the semantic gap between different Information System models and concepts2.Highlighting the non-interoperation effects
Product
CRM ERP
MESSCE
APS
[ Alexis AUBRY / Mario LEZOCHE]
Proposed Methodology
CRM IS1
IS2IS
APS
Conceptualisation
Conceptual Model 1
PRODUCT+Weight-val[0..1]:int+Weight-un[0..1]:UNIT
IS2
IS1
Conceptual Model 2
PRODUCT
WEIGHT+val[1]:int+un[1]:UNIT
0..11
Non-interoperation
Problems: 1. The two models are not made by the same expert. The modelling experiences are not the same Several possible conceptual representations (UML, MERISE, ORM…)
2. Majority of conceptual models are built a posterioriImplementation-based functionalities and constraints can disturb the semantics of the ISs: hiding, overloading...
Fact-oriented modelling
6
[ Alexis AUBRY / Mario LEZOCHE]
IS2
Proposed Methodology
CRM IS1
IS2IS
APS
Conceptual Model 2
Conceptual Model 1
Conceptualisation
Product
Weight-un
Weight-val
0..1Has
1
0..1Has
1
Fact-oriented Conceptual
Model 1
PRODUCT+Weight-val[0..1]:int+Weight-un[0..1]:UNIT
IS1
PRODUCT
WEIGHT+val[1]:int+un[1]:UNIT
0..11
Fact-oriented Conceptual
Model 2
PRODUCT
WEIGHT
val
un
1Has 1 1 Has1
1 0..1
Fact oriented modelling
Non-interoperation
- Using Object-Role Modelling which is the most popular fact-oriented approach.
7
[ Alexis AUBRY / Mario LEZOCHE]
IS2
Product
Weight-un
Weight-val
0..1Has
1
0..1Has
1
Proposed Methodology
8
CRM IS1
IS2IS
APS
Conceptual Model 2
Conceptual Model 1
Conceptualisation Fact oriented modelling
Fact-oriented Conceptual
Model 1
Fact-oriented Conceptual
Model 2
PRODUCT+Weight-val[0..1]:int+Weight-un[0..1]:UNIT
IS1
PRODUCT
WEIGHT+val[1]:int+un[1]:UNIT
0..11
PRODUCT
WEIGHT
val
un
1Has 1 1 Has1
1 0..1
Semantics structure enactment
Non-interoperation
Semantic Blocks of Model 2
- Pointing the concepts that are mandatory to semantically define a given concept in each fact-oriented conceptual model
Semantic Blocks of Model 1
[ Alexis AUBRY / Mario LEZOCHE]
IS2
Product
Weight-un
Weight-val
0..1Has
1
0..1Has
1
Proposed Methodology
9
CRM IS1
IS2IS
APS
Conceptual Model 2
Conceptual Model 1
Conceptualisation Fact oriented modelling
Fact-oriented Conceptual
Model 1
Fact-oriented Conceptual
Model 2
PRODUCT+Weight-val[0..1]:int+Weight-un[0..1]:UNIT
IS1
PRODUCT
WEIGHT+val[1]:int+un[1]:UNIT
0..11
PRODUCT
WEIGHT
val
un
1Has 1 1 Has1
1 0..1
Semantics structure enactment
Non-interoperation
Semantic Blocks of Model 1
Semantic Blocks of Model 2
Semantic gap evaluation
Semantic gap¬ IS1 ⊓ IS2
Common semanticsIS1⊓ IS2
Semantic gapIS1 ⊓¬ IS2
Semantic gapIS1 ⊔ IS2
- Using the semantic blocks for identifying the semantic gaps between the two information systems- Highlighting the non-interoperation effects
[ Alexis AUBRY / Mario LEZOCHE]
IS2
Product
Weight-un
Weight-val
0..1Has
1
0..1Has
1
Proposed Methodology
10
CRM IS1
IS2IS
APS
Conceptual Model 2
Conceptual Model 1
Conceptualisation Fact oriented modelling
Fact-oriented Conceptual
Model 1
Fact-oriented Conceptual
Model 2
PRODUCT+Weight-val[0..1]:int+Weight-un[0..1]:UNIT
IS1
PRODUCT
WEIGHT+val[1]:int+un[1]:UNIT
0..11
PRODUCT
WEIGHT
val
un
1Has 1 1 Has1
1 0..1
Semantics structure enactment
Non-interoperation
Semantic Blocks of Model 1
Semantic Blocks of Model 2
Semantic gap evaluation
Semantic gap¬ IS1 ⊓ IS2
Common semanticsIS1⊓ IS2
Semantic gapIS1 ⊓¬ IS2
Semantic gapIS1 ⊔ IS2
[ Alexis AUBRY / Mario LEZOCHE]
Fact Oriented Semantic Conceptualisation(Lexical enactment and primary structuring)
For each Information System
1. Reverse engineering from the database structure to the first conceptual model (automatic)
2. Conceptual model examination by the domain expert according to the good practices in the enterprise and the data (manual)
3. Fact-oriented transformation (automatic)
11
[ Alexis AUBRY / Mario LEZOCHE]
Reverse Engineering :DB structure -> First Conceptual Model
12
The reverse engineering process will extract the database’s entities, attributes, relationships from a given database to a conceptual model (in our case described in UML format).
Reverse Engineering
1° Step 2° Step
Physical Relational Conceptual
DBMS ER UML
[ Alexis AUBRY / Mario LEZOCHE]
Reverse Engineering:two different use cases
13
Flexnet
Sage X3
Relational DiagramRelational Diagram Class DiagramClass Diagram
Class DiagramClass Diagram
[ Alexis AUBRY / Mario LEZOCHE]
Expertwork
- Modifying the multiplicities of the attributes- Adding explicit names to the concepts, the
attributes and the associations- Others operations to fit the conceptual model to
the “real” use of the Enterprise Information System.
Automaticwork
Manual Expert work:Adonix Sage example
14
Sage X3Relations,
Primary keys, trigger, index…
Manual « cleaning » of this conceptual model by the domain expert according to the good practices in the enterprise and the data
Software interface
[ Alexis AUBRY / Mario LEZOCHE]
Fact Oriented Transformation
Fact oriented transformation (automatic)
1. Making out the attributes: building an association between the concepts and its attributes with respecting the multiplicities
2. Transforming the different types of association in simple association by respecting some rules and adding, eventually, some semantics annotation
15
C1
A1 [1]A2 [0..1]
C1
A1
A21
1
0..1
1
[ Alexis AUBRY / Mario LEZOCHE]
Fact Oriented Transformation:Associations reduction 1
16
C1
A1A2
C2
A3
C1
Aggregation
C21 *
C1 C21 *
{C1 aggregates: C2}
Generalization
C2
A3
1
1
C1
A2A1
11
11
1 *
{C1 generalizes: C2}
1
1
1
1
[ Alexis AUBRY / Mario LEZOCHE]
Fact Oriented Transformation:Associations reduction 2
17
C2C1
Composition
C21 *
C1 C21 *
{C1 isComposedOf: C2}
C1
C3
C2C1
C3
Association Class
1
1
1
*
*
*
{C3 isTypeOf: associationClass(C1,C2)}
[ Alexis AUBRY / Mario LEZOCHE]
Fact Oriented Transformation:Adonix Sage example
18
Facility
Product
1
*
Third Part Reference [1]Creation Date [1]Article Code [1]Quality Card [1]Replacement Article [1]UCEE-US Coefficient [0..1]Export Number [0..1]
RIB Number [1]Default Addres [1]Associated Site [0..1]
Product
Third PartReference
CreationDate
ArticleCode
QualityCard
ReplacementArticle
UCEE-USCoefficient
Export Number
Facility
RIBNumber
DefaultAddress
AssociatedSite
1
11
11 1
1
1
0..1
0..1
1 1 1 1
1
1
0..11
1
1
1
*
[ Alexis AUBRY / Mario LEZOCHE]
IS2
Product
Weight-un
Weight-val
0..1Has
1
0..1Has
1
Proposed Methodology
19
CRM IS1
IS2IS
APS
Conceptual Model 2
Conceptual Model 1
Conceptualisation Fact oriented modelling
Fact-oriented Conceptual
Model 1
Fact-oriented Conceptual
Model 2
PRODUCT+Weight-val[0..1]:int+Weight-un[0..1]:UNIT
IS1
PRODUCT
WEIGHT+val[1]:int+un[1]:UNIT
0..11
PRODUCT
WEIGHT
val
un
1Has 1 1 Has1
1 0..1
Semantics structure enactment
Non-interoperation
Semantic Blocks of Model 1
Semantic Blocks of Model 2
Semantic gap evaluation
Semantic gap¬ IS1 ⊓ IS2
Common semanticsIS1⊓ IS2
Semantic gapIS1 ⊓¬ IS2
Semantic gapIS1 ⊔ IS2
[ Alexis AUBRY / Mario LEZOCHE]
Semantics Structure enactment(Semantic blocks)
20
C1
C2
C3
C4
1..*
C1C2
*
1C1C3*
1..*
C3C40..1
C5
1..*C2C5_20..1
1 C2C5_1 *
C8
1
C5C8
*
C6
1..*
C3C61..*
C7
1C4C7_21
1..*
C6C7
*
1
C6C80..1
*
C5C6
0..1
*
C2C3
*
1
C4C7_11..*
0..1
C1C8
0..1
*
C8C8
1
A semantic block, denoted as B(c), and associated with a concept c, represents the set of concepts necessary to the minimal semantics definition of the concept c.
• Semantic block: definition
[ Alexis AUBRY / Mario LEZOCHE]
C1
C2
C3
C4
1..*
C1C2
*
1C1C3*
1..*
C3C40..1
C5
1..*C2C5_20..1
1 C2C5_1 *
C8
1
C5C8
*
C6
1..*
C3C61..*
C7
1C4C7_21
1..*
C6C7
*
1
C6C80..1
*
C5C6
0..1
*
C2C3
*
1
C4C7_11..*
0..1
C1C8
0..1
*
C8C8
1
21
B(C2)
A semantic block, denoted as B(c), and associated with a concept c, represents the set of concepts necessary to the minimal semantics definition of the concept c.
• Semantic block: definition
Semantics Structure enactment(Semantic blocks)
[ Alexis AUBRY / Mario LEZOCHE] 22
C1
C2
C3
C4
1..*
C1C2
*
1C1C3*
1..*
C3C40..1
C5
1..*C2C5_20..1
1 C2C5_1 *
C8
1
C5C8
*
C6
1..*
C3C61..*
C7
1C4C7_21
1..*
C6C7
*
1
C6C80..1
*
C5C6
0..1
*
C2C3
*
1
C4C7_11..*
0..1
C1C8
0..1
*
C8C8
1
𝐶1 𝐶2
𝐶3
𝐶5 𝐶8
𝐶4 𝐶7
𝐶6
Associated directed graph
• An associated semantic-relationships graph
Semantics Structure enactment(Semantic blocks)
[ Alexis AUBRY / Mario LEZOCHE] 23
C1
C2
C3
C4
1..*
C1C2
*
1C1C3*
1..*
C3C40..1
C5
1..*C2C5_20..1
1 C2C5_1 *
C8
1
C5C8
*
C6
1..*
C3C61..*
C7
1C4C7_21
1..*
C6C7
*
1
C6C80..1
*
C5C6
0..1
*
C2C3
*
1
C4C7_11..*
0..1
C1C8
0..1
*
C8C8
1
𝐶1 𝐶2
𝐶3
𝐶5 𝐶8
𝐶4 𝐶7
𝐶6
Associated directed graph
• An associated semantic-relationships graph
Semantics Structure enactment(Semantic blocks)
[ Alexis AUBRY / Mario LEZOCHE] 24
C1
C2
C3
C4
1..*
C1C2
*
1C1C3*
1..*
C3C40..1
C5
1..*C2C5_20..1
1 C2C5_1 *
C8
1
C5C8
*
C6
1..*
C3C61..*
C7
1C4C7_21
1..*
C6C7
*
1
C6C80..1
*
C5C6
0..1
*
C2C3
*
1
C4C7_11..*
0..1
C1C8
0..1
*
C8C8
1
𝐶1 𝐶2
𝐶3
𝐶5 𝐶8
𝐶4 𝐶7
𝐶6
Associated directed graph
• An associated semantic-relationships graph
B(C2)
Theorem 1A concept c’ is included in the semantic block of c if and only if there exists a directed path from c to c’ in the associated directed graph.
Semantics Structure enactment(Semantic blocks)
[ Alexis AUBRY / Mario LEZOCHE] 25
• A three-phases procedure to build the semantic blocksi. Building the associated
semantic-relationships graph,
ii. simplifying the graph using graph theory properties and using Kosaraju-Sharir’s algorithm,
iii. and building the semantic block associated with each aggregated node. Deducing the semantic block of each concept.
𝐶1 𝐶2
𝐶3
𝐶5 𝐶8
𝐶4 𝐶7
𝐶6
C1
C2
C3
C4
1..*
C1C2
*
1C1C3*
1..*
C3C40..1
C5
1..*C2C5_20..1
1 C2C5_1 *
C8
1
C5C8
*
C6
1..*C3C6
1..*
C7
1C4C7_21
1..*
C6C7
*
1
C6C80..1
*
C5C6
0..1
*
C2C3
*
1
C4C7_11..*
0..1
C1C8
0..1
*
C8C8
1
𝑆𝐶𝐶1 𝑆𝐶𝐶2
𝑆𝐶𝐶3 𝑆𝐶𝐶4
Semantics Structure enactment(Semantic blocks)
[ Alexis AUBRY / Mario LEZOCHE]
Future works
26
IS2
Product
Weight-un
Weight-val
0..1Has
1
0..1Has
1
CRM IS1
IS2IS
APS
Conceptual Model 2
Conceptual Model 1
Conceptualisation Fact oriented modelling
Fact-oriented Conceptual
Model 1
Fact-oriented Conceptual
Model 2
PRODUCT+Weight-val[0..1]:int+Weight-un[0..1]:UNIT
IS1
PRODUCT
WEIGHT+val[1]:int+un[1]:UNIT
0..11
PRODUCT
WEIGHT
val
un
1Has 1 1 Has1
1 0..1
Semantics structure enactment
Non-interoperation
Semantic Blocks of Model 1
Semantic Blocks of Model 2
Semantic gap evaluation
(Yahia, 2011)
Semantic gap¬ IS1 ⊓ IS2
Common semanticsIS1⊓ IS2
Semantic gapIS1 ⊓¬ IS2
Semantic gapIS1 ⊔ IS2
[ Alexis AUBRY / Mario LEZOCHE]
Enterprise Information Systems: a proposition for conceptualisation and interoperability evaluation
Alexis AUBRYMario LEZOCHE
Thank you for your attention !
2010, 10th November