Domain Oriented Meta-Modelling for change

Management of Information System

Julie Chapron

, Xavier Boucher

, Patrick Burlat

, Pierre Lebrun

Ecole Nationale Supérieure des Mines de St Etienne, Centre G2I

158, cours Fauriel, 42023 Saint-Étienne cedex 2, France

ST Microelectronics – Architecture and Strategy Team

Z.I. av Olivier Perroy, 13790 Rousset, France

Abstract : This paper presents the first results of a research-partnership

between Ecole Nationale Supérieure des Mines de St Etienne (G2I) and St-

Microelectronics (Rousset - France). We explain the results of a first stage of

research works which aim at developing a method for the management of the

evolution of information systems. Our current contribution consists in a

domain-oriented enterprise modelling approach. We justify the structure of the

meta-model proposed and we show its links with evolution management

1 Introduction

The research works presented below focus on the evolution management of the

Information System (IS) applied to a specific industrial field (the microelectronics

industry). This research is a partnership (including the PhD of J.Chapron) between the

Ecole des Mines de Saint Etienne (G2I department) and the Architecture and Strategy

Team of ST-Microelectronics on the site of Rousset (France). The final aim is to

define a methodology and tools for the management of the evolution of the

information system, with the following requirements :

• Transparency of the information system : the ability to provide a global,

coherent and useful model of the information system;

• Models of change process: Characterization of evolution situations, risk

and impact analysis, formalization of the change process;

• Decision aid tools to master reactivity facing situations of evolution, and

taking into account the specificities of the microelectronics field.

In the last 20 years, research on enterprise modelling has produced several

descriptive methods aiming at providing a methodological and operational support for

the design, the analysis and the reengineering of industrial firms : methodologies such

as PERA [10], CIMOSA [1], GRAI [6], are based on distinct meta-models

formalizing the information required for their specific enterprise models [7]. In order

Chapron J., Boucher X., Burlat P. and Lebrun P. (2004).

Domain Oriented Meta-Modelling for change Management of Information System.

In Proceedings of the 1st International Workshop on Computer Supported Activity Coordination, pages 261-266

DOI: 10.5220/0002684902610266

 SciTePress

to integrate these approaches using their specificities in the best possible way (for

instance a focus on life cycle for PERA or on decision system for GRAI), GERAM

[7] intends to provide a global integrative framework on enterprise modelling.

Research on GERAM is linked to normalization works which lead in Europe to

standards like (ISO 14 258, ENV 40 003). After different European normative

advances [3], UEML language [8] is the last normalization work intending to develop

a unified language to support exchanges between currently existing modelling tools.

In spite of all these results, lack in enterprise modelling is still to be underlined

concerning the management of the evolution of organisational structures and of their

information systems. To fulfil such deficiency, Information System Urbanism [5]

appears as an innovative industrial approach within the information system field. It

provides the necessary structuring and basic concepts to build a real method for

change management for information systems. Urbanism aims at transforming the

information system to enable it to anticipate and to adapt to different kinds of changes

within the firm (strategic, organizational, legal…). The approach requires an

urbanistic program on the information system, including (i) the targeted information

system, (ii) urbanistic rules insuring large possibilities of evolution for the new

system, and (iii) the roadmap to reach the target system. A global methodology is

already available, due to industrial experiences. However much scientific work

remains to be done, in order to clearly formalize the concepts and methods, before

being able to build up better scientific tools aiming at the management of information

systems evolution.

This paper focuses on a first contribution to integrate the urbanism concepts into

existing methodologies for enterprise modelling. In section 2, we develop a meta-

model used to map the information system, taking into account the requirements for

evolution management and linked with urbanism principles. The section 3 presents an

application of this meta-model to the microelectronics industry.

2 Meta-model Proposal

The meta-model we elaborated puts forth two major partitions of modelling concepts.

First we justify a partition between models of processes and the models of

organisational environments : in an evolution perspective these two domains don’t

change with similar constraints and characteristics (§2.1). Thus, we specify the

interface between processes and organisational environments, with generic concepts,

independent for the operational constraints of the firm (§2.2). Secondly we explain a

partition between an information-technology oriented modelling and a domain-

oriented modelling (§2.3). This partition is namely justified by urbanism principles.

2.1 The Need to Separate Processes and Organizational Environment

Referring to CIMOSA [1] or more recently GERAM [7], we propose a conceptual

partition between functional modelling (focused on processes models), and

organizational modelling. Figure 1 highlights partition by a “business processes”

modelling view and by an “organizational environment” modelling view. Each view

262

gathers various models, each of them gathering modelling objects. Here, we do not

put all the modelling objects into detail.

Domain

Oriented

Modeling

IT Oriented

Modeling

is carried out by

use 4

is perform by

has the capacity

use 2

Is composed by 1

Is composed by 3

Is composed by 2

call 2

call 1

use 3

is responsible for 3

is implement with

has the capability

use 1

is based on 2

is based on 1

is based on 3

Material Capability Model

Business Rule Model

Material Ressource Model

Organisationnal Model

Hardware Model

Architecture Model

Role model

Capacity Model

Process Map

Operational Process Model

Informationnal Process Model

Fonctionnality Model

IT process model

Enterprise

Process View

Inferface View Organizationnal

Environment View

Fig. 1. Meta-model of our approach

The “business processes” view describes all the functional aspects linked with the

processes, using different granularity levels. The central element of the view is the

informational process used to represent management and control processes requiring

informational and decisional activities (partially automated). The “organizational

environment” view gathers the various models which represent part of the

environment of the process. There are bi-directional interactions between the

processes and their environment, the former including material resources, actors, IT

resources (hardware, software, data, documents, …).

This conceptual partition will later be used for evolution management : it is

important to clearly distinguish between changes in the processes themselves and

changes in their organizational environment. For instance, a change among the firm’s

employees, in the data base structures or even in the software can be introduced

without any transformation of the process (i.e. its structure and model) in itself. Thus,

this partition between processes and environment leads to a classification of evolution

situations, but also to a better impacts analysis.

2.2 Interface View

We have just explained that the two previous views are built of several models. These

models are dependent on each other, and one of the aims of the enterprise modelling

approach is to manage these dependency links. The third “interface” view aims at

263

formalizing the inter-dependencies between models of the first two views. These

inter-dependencies are formalized in a generic way, which means that they remain

valid for various versions of the organizational environment.

Thus, the « interface » view will specify the requirements of the business processes

as regards the organizational environment. For instance a “functionality model” will

specify the dependency link between an information process model and an IT

architecture model. The “functionality” model is a generic specification of IT tools

meeting the requirements of the informational process. It remains independent from

the existing IT architecture.

2.3 Partition between Domain-oriented Modelling and IT Oriented Modelling

Figure 1, shows a second conceptual partition between a “domain-oriented”

modelling and an “information-technology” oriented modelling. These two modelling

layers concern the three views presented above.

In order to manage the evolution of processes at ST Microelectronics, the partition

between these modelling layers will allow the management of a progressive

rationalization, aiming at transforming classical enterprise processes into computer

supported processes. Indeed, the Microelectronics field is a high-technological

environment with a high level of automation which never stops improving. This

induces a permanent optimization of the management and control tasks. In this

perspective, the evolution management aims at realizing a progressive encapsulation

of the enterprise processes, based on Information Technology. Thus, at the global

scale of the firm, the partition between domain-oriented modelling and IT oriented

modelling gives the possibility to measure IT deployment level among the processes

and to identify critical zones within the firm from that point of view.

Furthermore, for methodological requirements, it is necessary to figure out the

frontier between what we could call the “business” world and “information

technology” world [4], as well as their correspondence links. Indeed, several

methodologies aim at managing information system evolution using the principles of

urbanism applied to IT systems [5]. These methodologies insist on the necessity to

manage the consistence between these two “worlds” : the IT urbanization zones in the

firm must be formally related to organizational zones.

3 Meta-model Implementation on the Microelectronic Domain

This study is based on the industrial needs of ST Microelectronics. Indeed, the

microelectronics industry is characterized by fast and dramatic evolutions of products

and processes. In this section, we illustrate our approach with an existing operational

process named “control card management” from STM. The control cards are used to

test the electrical functionality of silicon wafers (on which the integrated circuits are

produced). These cards test the validity of the chip and represent an essential step in

the production process.

In fig. 2, we show a part of the application of the meta-model to the “control card

management” process. It was developed using Adonis, a tool developed by BOC

264

GmbH. Adonis offers the possibility of changing its meta-model. This process occurs

in a complex environment (simplified in fig. 2) and it is characterized by its

interactions with (i) other processes and sub-processes (these interactions are

modelled in the process map), (ii) actors and organizational structure (depicted in

fig.2 by the models “Organization” and “role”) and (iii) the information technology

architecture and infrastructure (shown in fig.2 by the models “Function” and “IT

system”).

Fig. 2. Example: part of the process “Control Card Management”

This process provides an example of a situation of evolution. Indeed, at the

information system level, a study of the process has shown the importance of

implementing a workflow system to support the process which was previously

manual. A change like the implementation of a workflow will have impact on other

areas such as the database by trying to access data not yet available and interact with

other existing application. By modelling the IT infrastructure and architecture we will

be able to manage the evolution of this system and control any eventual perturbations.

The complete meta-model is currently being implemented by ST Microelectronics

in a pilot project. Of course it will with other types of changes: at a basic level, an

evolution may occur in the architecture of the database or in the structure of the data.

RoleOrganisation Process

M. Smith M. BlancM. DurantM. MartinM. Dupont

Device Engineer

Production controlSPCProcess controlProcess Engineer

request for change

of control card

Realisation

Technical

validation

Modification

study

Process Validation

IT systemFonction

Data production

treatment

DB updateWorkflowWorkflowWorkflow

CAM system Control Car serverECNS SPC RAD ECNS SPC RAD ECNS SPC RAD

Send the

New card

Operation success

diffusion

WS test

update

Request

reception

CC BDCC server I/O Manager CC Manager

Ask card

name

VFEI driver

Eq Manager

Check the

version

RoleOrganisation Process

M. Smith M. BlancM. DurantM. MartinM. Dupont

Device Engineer

Production controlSPCProcess controlProcess Engineer

request for change

of control card

Realisation

Technical

validation

Modification

study

Process Validation

IT systemFonction

Data production

treatment

DB updateWorkflowWorkflowWorkflow

CAM system Control Car serverECNS SPC RAD ECNS SPC RAD ECNS SPC RAD

RoleOrganisation Process

M. Smith M. BlancM. DurantM. MartinM. Dupont

Device Engineer

Production controlSPCProcess controlProcess Engineer

request for change

of control card

Realisation

Technical

validation

Modification

study

Process Validation

IT systemFonction

Data production

treatment

DB updateWorkflowWorkflowWorkflow

CAM system Control Car serverECNS SPC RAD ECNS SPC RAD ECNS SPC RAD

Send the

New card

Operation success

diffusion

WS test

update

Request

reception

CC BDCC server I/O Manager CC Manager

Ask card

name

VFEI driver

Eq Manager

Check the

version

265

The model allows us to analyze the quantitative and qualitative impacts of such a

change on all the processes. At the organizational level, internal restructuring are

quite frequent. The organizational models linked with the process models allows a

better management of the correspondence between the actors and the process.

4 Conclusion And Perspective

This paper focuses on how we can take into account the evolution perspective within

enterprise modelling. The first results discussed above, show a meta-model intended

to fulfill our requirements. The structure of this meta-model constitutes a first step

towards evolution management. However, as far as enterprise modelling is concerned,

additional modelling concepts are still to be formalized in the future, like evolution

events, process transformation, evolution scenario, or evolution processes…. Such

concepts will certainly be necessary to fulfill our goals [2].

Furthermore, such a meta-model provides only the first step towards evolution

management. The structure of the meta-model and the associated methodology are the

starting point that are used in order to develop further a decision aid approach for IS

evolution management. In this perspective, the meta model is currently used on a pilot

project in order to generate a chart of the enterprise (with all the necessary models)

focused on a limited pool of the firm’s processes. This data will be used to develop

two types of evolution management tools : (i) tools to define consistent processes

pools to be jointly managed within change scenarii ; (ii) tools to diagnose the impacts

and risks related to a selective change situation.

References

1. AMICE Consortium: Open System Architecture for CIM. Rapport de recherche du projet

ESPRIT 688. Springer-Verlagn (1989)

2. Barrios, J., Rolland, C., Nurcam, S.: Une méthode la définition de l’impact organisationel

du changement. Ingénierie des Systèmes d'Information vol. 7 n°4. Hermès. France (2002)

3. Chen, D., Vernadat, F.: Enterprise Inter- and Intraorganizational Integration, Building

International Consensus. Proceedings of the ICEIMT'02, K. Kosanke et.al (2002)

4. Chelli, H.: Urbaniser l'entreprise et son système d'Information. Guide des entreprises agiles.

VUIBERT. Paris France (2003)

5. Club Urba-SI: Pratique de l’urbanisme des systèmes d’information en entreprise.

Publibook. France. (2003)

6. Doumeight, G.: Méthode GRAI: méthode de conception des systèmes en Productique.

Thèse d’Etat, Automatique. Université Bordeaux I, France (1984)

7. IFAC-IFIP: Generalized Enterprise Reference Architecture and Methodology, Version

1.6.3. IFAC-IFIP Task Force on Architecture and Methodology, ISO IS 15704. (1999)

8. Themantic Network Project: Requirements analysis: initial core constructs and architecture.

European project IST–2001–34229. Deliverable 3.1, (May 2003)

9. Verdanat, F.: Techniques de Modélisation en Entreprise : applications aux processus

opérationnels. Ed. Economica. Paris France (1999)

10. William,T.J.: The Perdue Enterprise Reference Architecture. Computer in Industry,

Volume 24, Elsevier. (1994)

266