Architectural styles for early goal-driven middleware platform selection [Elektronische Ressource] / von Simon Giesecke

carl_von_ossietzky_universitat_oldenburg

Le téléchargement nécessite un accès à la bibliothèque YouScribe
Tout savoir sur nos offres

279 pages

English

Le téléchargement nécessite un accès à la bibliothèque YouScribe
Tout savoir sur nos offres

A propos
Informations
Extrait

Description

Sujets

Fakultät II – Informatik, Wirtschafts- und Rechtswissenschaften
Department für Informatik
Architectural Styles
for Early Goal-driven
Middleware Platform Selection
Dissertation zur Erlangung des Grades eines
Doktors der Ingenieurwissenschaften
von
Dipl.-Inform. Simon Giesecke
Gutachter:
Prof. Dr. Wilhelm Hasselbring
Prof. Dr. Ralf H. Reussner (Univ. Karlsruhe)
Tag der Disputation: 2008-09-262Contents
Contents 3
1 Introduction 11
1.1 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
1.2 Solution Approach . . . . . . . . . . . . . . . . . . . . . . . . 12
1.3 Scientiﬁc Contribution . . . . . . . . . . . . . . . . . . . . . . 13
1.4 Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
I Foundations 17
2 Software Quality and its Evaluation 19
2.1 Introduction to Software Quality . . . . . . . . . . . . . . . . 19
2.2 Basic Terms . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
2.3 Realms of Software Quality . . . . . . . . . . . . . . . . . . . 21
2.4 Software Quality Models . . . . . . . . . . . . . . . . . . . . . 25
2.5 Software Architecture Evaluation: Goals and Methods . . . . 27
3 Software Design & Architecture 29
3.1 Basic Deﬁnitions . . . . . . . . . . . . . . . . . . . . . . . . . 29
3.2 General Concepts of Architectural Description . . . . . . . . 34
3.3 The Component-and-Connector Viewpoint . . . . . . . . . . . 41
3.4 Architectural Design Constraints . . . . . . . . . . . . . . . . 47
3.5 Arc Description Languages . . . . . . . . . . . . . . 57
II The MidArch Approach 67
4 Research Design and Methods 69
4.1 WP1: Analysing Architectural Styles and their Usage . . . . 69
4.2 WP2: Extending ISO Standard 42010 to Represent Architec-
tural Rationale . . . . . . . . . . . . . . . . . . . . . . . . . . 71
4.3 WP3: Modelling Middleware-oriented Architectural Styles . . 71
4.4 WP4: Developing a MidArch Style-Based Middleware Selec-
tion Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
4.5 WP5: Providing Tool Support for the MidArch Design Method 75
3CONTENTS
5 Basic Deﬁnitions and Assumptions 77
5.1 Software Architecture . . . . . . . . . . . . . . . . . . . . . . 77
5.2 Middleware . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
5.3 Architecture-level Design Exploration . . . . . . . . . . . . . 84
5.4 Formalisation . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
5.5 Design Knowledge . . . . . . . . . . . . . . . . . . . . . . . . 88
5.6 Software Development Methods and System Quality . . . . . 91
5.7 Architectural Styles and the ISO 42010 Reference Model . . . 96
6 Taxonomy of Architectural Style Usage 101
6.1 Types of Codiﬁed Architectural Design Knowledge . . . . . . 101
6.2 Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
6.3 Usage-based Taxonomy of Architectural Styles . . . . . . . . 103
7 Modelling Middleware-oriented Architectural Styles 117
7.1 Relationship of Middleware, MidArch Styles and Architectural
Descriptions . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118
7.2 Assumptions . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
7.3 Design Decisions in Middleware-oriented Architectural Style
Modelling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
7.4 Language Speciﬁcity . . . . . . . . . . . . . . . . . . . . . . . 125
7.5 MidArch Style Taxonomies . . . . . . . . . . . . . . . . . . . 130
7.6 General MidArch Style Modelling Method . . . . . . . . . . . 136
7.7 Style-enabled ADLs . . . . . . . . . . . . . . . . . . . . . . . 139
7.8 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150
8 The MidArch Design Method for Early Middleware Selection 153
8.1 Assumptions . . . . . . . . . . . . . . . . . . . . . . . . . . . 153
8.2 Activities of the MidArch Design Method . . . . . . . . . . . 155
8.3 MidArch Style Taxonomy and Repository Speciﬁcation . . . . 156
8.4 Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164
8.5 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170
III Evaluation 171
9 Overview 173
9.1 Rationale . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173
9.2 Case Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . 176
9.3 Experimental Validation . . . . . . . . . . . . . . . . . . . . . 176
4CONTENTS
10 Case Studies 179
10.1 RegIS Online Case Study . . . . . . . . . . . . . . . . . . . . 179
10.2 Thales Case Study . . . . . . . . . . . . . . . . . . . . . . . . 190
10.3 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . 195
11 Tool Support 197
11.1 Tool Support for the MidArch Design Method: MidArch
Repository Tool . . . . . . . . . . . . . . . . . . . . . . . . . . 197
11.2 Tool Support for Implementation Mapping . . . . . . . . . . . 199
12 Related Work 209
12.1 Style and Style-based Architecture Modelling . . . . . . . . . 209
12.2 Architecture-Level Evaluation and Selection of Middleware . 212
IV Conclusions and Future Work 217
13 219
13.1 WP1: Analysing Architectural Styles and their Usage . . . . 219
13.2 WP2: Extending ISO Standard 42010 to Represent Architec-
tural Rationale . . . . . . . . . . . . . . . . . . . . . . . . . . 222
13.3 WP3: Modelling Middleware-oriented Architectural Styles . . 222
13.4 WP4: Developing a MidArch Style-Based Middleware Selec-
tion Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . 224
13.5 WP5: Providing Tool Support for the MidArch Design Method225
14 Future Work 227
14.1 WP1: Analysing Architectural Styles and their Usage . . . . 227
14.2 WP2: Extending ISO Standard 42010 to Represent Architec-
tural Rationale . . . . . . . . . . . . . . . . . . . . . . . . . . 227
14.3 WP3: Modelling Middleware-oriented Architectural Styles . . 228
14.4 WP4: Developing a MidArch Style-Based Middleware Selec-
tion Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233
A Deﬁnition of the MidArch Design Method 237
A.1 Basic Deﬁnitions . . . . . . . . . . . . . . . . . . . . . . . . . 237
A.2 Roles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 237
A.3 Tasks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 240
List of Acronyms 249
List of Tables 251
List of Figures 253
5CONTENTS
Bibliography 255
6Summary
Business information systems are middleware-intensive applications, i.e. their
structure and behaviour is signiﬁcantly inﬂuenced by the chosen middleware
platform. The selection of a suitable middleware platform is a task in
the development process that is critical in fulﬁlling the system’s quality
requirements, e.g. availability or time eﬃciency. A systematic selection
should be done as early as possible within a development project, i.e. on the
architectural level. Up to now no adequate method for selecting a middleware
platform in a concrete project context exists, thus often ad-hoc decisions are
made.
As a solution, the MidArch approach was developed, which is presented
here. The main constituent is the MidArch Design Method, which uses
a repository of MidArch Styles and a MidArch Style Taxonomy for sup-
porting the systematic selection of a middleware platform and to guide the
speciﬁcation of software architectures.
A middleware-oriented architectural style (MidArch Style) is a model of
aare platform, which captures the structural constrains that are
imposed upon the application architecture by the underlying platform. From
another point of view, a MidArch Style characterises the family of software
architectures that can be implemented on a given middleware platform.
Within the MidArch Design Method, a MidArch Style Taxonomy is used as
the basis of a stepwise selection process of candidate MidArch Styles, which
match the given quality requirements. Consequently, software architectures
are speciﬁed that conform to these styles, and the architectures are evaluated
against the quality requirements. The evaluation results are lifted from
the level of individual architectures to the level of the MidArch Styles by
comparison. This information is stored in the repository as annotations to
the respectives MidArch Styles. In subsequent applications of the MidArch
Design Method, these are used to reﬁne the selection process.
For evaluation, two case studies using industrial systems are presented.
Furthermore, the thesis contains contributions in the areas of software archi-
tecture foundations and the deﬁnition of architectural styles and patterns.
7Summary
Zusammenfassung
Betriebliche Informationssysteme sind Middleware-intensive Anwendungen,
d.h. Softwaresysteme, deren Struktur und Verhalten durch die Middleware
wesentlich beeinﬂusst werden. Bei der Entwicklung ist die Auswahl einer
Middleware-Plattform eine kritische Aufgabe in Hinblick auf die Erfüllung
der gestellten Qualitätsanforderungen, die beispielsweise Verfügbarkeit oder
Zeiteﬃzienz betreﬀen. Eine fundierte Auswahl sollte möglichst früh in einem
Entwicklungsprojekt erfolgen, d.h. auf der Ebene der Softwarearchitektur.
Bisher existieren jedoch wenige Richtlinien oder systematische Vorgehens-
modelle für die Auswahl einer Middleware-Plattform in einem konkreten
Projektkontext, so dass oftmals Ad-hoc-Entscheidungen getroﬀen werd