Using Variability Modeling Principles to Capture Architectural Knowledge

1
Using Variability Modeling Principles to Capture
Architectural Knowledge

• Marco Sinnema (University of Groningen),
• Jan Salvador van der Ven (University of
  Groningen),
• Sybren Deelstra (University of Groningen)

2
Outline

• Software Product Families
• COVAMOF
• Architectural Knowledge
• Mapping to COVAMOF
• Example

3
Software Product Families
4
Software Product Families
notepad interface
mail manager
network interface
built-in editor
mail manager
network interface
file storage
database interface
Customer 1
Customer 2
word interface
mail manager
network interface
built-in editor
mail manager
network interface
database interface
file storage
Customer 3
Customer 4
5
Software Product Families
Derive Product
Derive Product
Customer 1
Customer 2
6
Product Derivation
7
COVAMOF Variability Modeling
Product Configuration
1

Decision


Variant
1
Variation Point
with rationale

Dependency

Dependency Interaction
8
Architectural Knowledge

• Supports the development of the architectural
  design
• During this development, issues arise
• These issues lead to architectural choices
  between two or more alternatives
• Based on rationale, engineers make a decision for
  an architectural choice
• Making these decisions often involves making
  trade-offs between quality attributes

9
Commonalities
Product Family Engineering
Architectural Knowledge
10
Mapping on COVAMOF
System Design
Product Configuration
1

Decision
Decision


Alternative
Variant
1
Architectural Choice
Variation Point
with rationale
with issue

Quality Attribute
Dependency

Trade-off
Dependency Interaction
11
Example
Performance
Trade-off
Processor Usage
Quality Attributes
Memory Consumption
Editor
Persistent Data
Architectural Choices
Built-in Editor
Database Interface
Alternatives
Word Interface
File Storage
Notepad Interface
System Design
Decisions
12
Conclusion

• We can use Variability Modeling concepts to
  capture Architectural Knowledge

13
The End

• Any Questions?

More Information
Shark 2006 paper
m.sinnema_at_rug.nl
http//www.msinnema.nl/covamof
14
Vision
Product Family Engineering
Variability
15
Vision
Variability
16
Vision
17
Vision Example 1
18
Vision Example 2
19
Conclusion

• We can use Variability Modeling concepts to
  capture Architectural Knowledge
• In the future we should strive for an integrated
  solution for managing architectural choices and
  variation points

20
The End

• Any Questions?

More Information
Shark 2006 paper
m.sinnema_at_rug.nl
http//www.msinnema.nl/covamof
21
Benefits of COVAMOF

• Incremental Externalization
• Document Expert Knowledge
• Incorporate Existing Documentation
• Collect Reference Data
• Formalize Documented Knowledge
• Reduce Derivation Cost
• Reduce Expert Involvement
• Provide Structured Documentation
• Configuration Guidance
• Automatic Inference
• Automatic Consistency Checking
• Automatic Quality Attribute Estimation
• Tools integrate in development environment
  (Visual Studio .NET)

22
Tool support for COVAMOF
23
Example
24
Example
Memory Usage
Variation points
25
COVAMOF Variability Model
Legend
Product class
Variation
Point
Variant
Feature Layer
Full
Custom
Basic
Realization

Dependency
Realizes
Editor Component

Persistency
Component
Architecture Layer
Memory Usage
Notepad
Built-in
Word
Files
Database