Transcript Slide 1

Evoluzione dei Sistemi Software e Reverse Engineering
Software Evolution and Reverse Engineering
Founded in 2004
Main aims:
Analyze and define approaches, methodologies, techniques, tools able to
support software evolution and maintenance
•
•
● scientists: 2 faculties, 3 PhD students
● external collaborators
•
Contact: Francesca Arcelli Fontana, [email protected]
•
Website: http: //essere.disco.unimib.it/
Software Evolution and Reverse Engineering
System evolution carries:
– Software Architecture violation
– Not up to date documentation
– .....
Expensive maintenance
Software Evolution and Reverse Engineering
Main research areas:
• Design Pattern Detection (DPD) for Reverse
Engineering
• Software Architecture Reconstruction (SAR)
– Program Comprehension
• Legacy System Modernization
– Migration of legacy systems to SOA
– Model-driven reverse engineering
– Function and Data Reengineering
Design Patterns Detection (DPD)
•
•
•
•
•
DPD for reverse engineering
DPD for forward engineering (Design Advisor)
Quality design assessment
Detection based on DP sub-components
Architectural Patterns detection
– Pattern detection for data reverse engineering
• Exploiting both static and dynamic analysis
Software Architecture Reconstrution (SAR)
• We have experimented several tools for SAR and
Program Comprehension on exixting systems of
different complexity
– Expecially for Java, C++ systems
– Cobol, RPG
• We are developing a tool to support Program
Comprehension providing:
– Views, diagrams on the system
– Metrics
– High level architectural abstractions
– Anti Patterns detection
Data Mining Techniques for Reverse Engineering
• We experimented data mining techniques for
DPD:
– Supervised Classification (SVM, Naive
Bayes,..)
– Unsupervised Classifcation(Clustering, Weka
SimpleKMeans)
• experimenting other tecniques for DPD and SAR
Systems Modernization
• How applications become legacy?
• Virtually every company has legacy applications of some form
• Often companies don’t have enough money for new projects
because existing applications cost too much to operate and maintain
• Only some legacy applications will be reusable: which is the right
level of granularity to support reuse?
Solutions??
SOA are really the future state target for IT?
Archtecture Driven Modernization – ADM (OMG standard)
Migration to SOA
• We are currently working on the analysis and
experimentation of tools and methodologies for
legacy systems migration towards Service Oriented
Architectures.
– Program Comprehension and SAR
– DP Detection
– Anti Pattern Detection (modularity, dependecy relations)
– We are experimenting the COMPASS (COde Migration
Planning and ASSesment) workbench framework of IBMHaifa to understand applications and assist in isolating the
business component in VB code towards SOA
DPD for Migration to SOA
• Exploiting the knowledge associated to some
Design Patterns to provide a kind of guidance in
designing SOA-based systems.
• Identifying the design patterns more relevant:
– in the design of a SOA based system;
– in the transformation process towards SOA when
identified in an existing system or in a system under
development, as they provide additional information
on the underlying architecture and behaviour.
– i.e. DP Mediator as a service or service composer,
DP Façade,..
Software modernization exploiting architectural models
Archtecture Driven Modernization – ADM (OMG standard-2004)
•
•
•
Capture architectural aspects of existing applications
Consolidate best practices leading to successful modernization
Modernization from an analysis and design based perspective to prevent
propagation of obsolete concepts and designs in modern language and
platforms
•
Aims to create interoperability standards for modernization tools
– KDM (Knowledge Discovery Metamodel)-2007: Defines an initial meta-model
that allows modernization tools to exchange application meta-data across
applications, languages, platforms and environments
– Modisco (MOdel DISCOvery): An Eclipse GMT (Generative Modeling
Technology) component for model-driven reverse engineering
Software modernization and Marple
Marple (Metrics and ARchitecture
Reconstruction PLugin for Eclipse)
• AST of the Java language
• Module that extracts Basic Elements and
Metrics
• Module allowing to visualize
Metrics and Design Patterns
• Module that detects Design Patterns using
Basic Elements contained in the model
12
• ASTs of the supported
languages
• AST to “MetaAST” transformation function
• “MetaAST” that allows the representation
of “all” object oriented languages
• Module allowing to analyze the “MetaAST” in
order to extract the information to be inserted
in the model (e.g. Metrics Collector: calculates
the metrics)
• Module allowing to visualize and
explore all the information within the
model
• Module allowing analyses on the model that
enrich it with the detected information (e.g.
Design Pattern Detection: detects Design
Patterns using basic elements within the
model)
13
• Allows to implement new
languages AST transformation
functions (e.g. Python)
• Allows to implement new
analyses on the “MetaAST” (e.g.
DataFlow Analysis)
• Allows to implement
new visualizations and
explorations of the
model
• Allows to implement new
analyses on the model (e.g.
correlation analysis)
14
D t
a a
S
o
f
t
15
w
a
r
e
Collaborations with other DISCo-Laboratories
• ITIS (Batini, Viscusi), Function and Data
Reverse Engineering
• MAD Data Mining for Reverse Engineering
• LTA, Dynamic Analysis for DPD and SAR
• SAL, Detection of architectural patterns,
detection of interfaces/services, software
evolution through architectural reflection
• Open-IT, Application of reverse engineering
techniques to close software, software protection
(decompiling techniques)
• TangoLab, stage on projects supporting
activities in Ciad.
Collaborations
Research Collaborations activated or to start:
Università di Salerno
Università del Molise
Università di Roma Tre
The Technion – Israel Institute of Technology, Unified
catalog of Micro architectures
University of Szeged, Ungheria,Benchmark for DPD
INRIA (ATLAS Group, Nantes), Model driven reverse
engineeering
Microfocus (APM, Metrics)
Replay (System Modernization)
IBM Haifa (Migration to SOA-Compass)