Supporting Evolution in Model-based Product Line Engineering
Deepak Dhungana Thomas Neumayer Paul Gru¨nbacher Rick Rabiser
Christian Doppler Laboratory for Automated Software Engineering
Johannes Kepler University Linz, Austria
dhungana@ase.jku.at
Abstract
Software maintenance and evolution are among the
most challenging and cost-intensive activities in soft-
ware engineering. This is not different for software
product lines due to their complexity and long life-span.
New customer requirements, technology changes and
internal enhancements lead to the continuous evolution
of a product line’s reusable assets. Due to the size of
product lines, single stakeholders or teams can only
maintain a small part of a system which poses addi-
tional challenges for evolution. This paper presents an
approach supporting product line evolution by organiz-
ing variability models of large-scale product lines as a
set of interrelated model fragments defining the vari-
ability of particular parts of the system. The approach
allows semi-automatic merging of fragments into com-
plete variability models. We also provide tool support
to automatically detect changes that would make mod-
els and the architecture inconsistent. Furthermore, our
approach supports the co-evolution of variability mod-
els and their respective meta-models. We illustrate the
approach with examples from an ongoing industry col-
laboration.
1. Introduction
There are estimates that more than 60% of software
costs are devoted to evolution and maintenance [9].
Product line engineering (PLE) is fundamentally chal-
lenged by evolution because the world cannot simply
be stopped for modelling purposes. The system has to
be continuously changed to meet evolving needs. Prod-
uct line approaches need to consider maintenance and
evolution as critical strategic elements due to the size,
complexity, and longevity of systems. Evolving and
maintaining a software product line requires dealing
with different types of changes caused by changing cus-
tomer needs, evolving technology, or market develop-
ments. Despite its importance surprisingly few papers
are available on product line evolution (e.g., [1, 14, 24]).
Different activities in domain and application engineer-
ing are usually based on the assumption that a product
line is fairly stable. However, such stability cannot be
taken for granted. PLE should thus treat evolution as
the normal case and not as the exception [5].
Model-based product line approaches use models
to define the knowledge required for maintaining and
evolving a product line, i.e., core assets, their common-
alities and variability [6]. For instance, a reference ar-
chitecture model describes the variability of architec-
tural elements like components, connectors, interfaces,
or ports. More general purpose variability modeling
approaches [8, 21] allow defining the variability of ar-
bitrary domain-specific assets. Domain-specific meta-
models define the elements of variability models such as
types of reusable assets, attributes of assets, and depen-
dencies [4]. Evolution support becomes success-critical
in a model-based development approach to ensure con-
sistency after changes to meta-models, models, and ac-
tual development artifacts.
An analysis of development practices of our indus-
try partner Siemens VAI revealed three important issues
related to evolution in model-based PLE: (i) Multiple
teams are involved in creating and maintaining a prod-
uct line. The knowledge about the system’s variability
is typically distributed across multiple stakeholders and
teams [5]. (ii) Different parts of the system evolve at
different speeds further contributing to the complexity
of evolution [7]. (iii) The domain meta-models evolve
in parallel to the variability models, which makes it hard
to keep them consistent with each other.
We have presented some basic ideas of our ap-
proach in a short paper [5]. Here, we elaborate further
on the underlying issues and present our experiences
of applying the tool-supported approach in a real-word
product line of Siemens VAI, the world’s leading steel
12th International Software Product Line Conference
978-0-7695-3303-2/08 $25.00 © 2008 IEEE
DOI 10.1109/SPLC.2008.26
319