Abstract

The case study provided in this work package will cover the area of development of large scale Network Management Systems. Comarch has already developed well established methodology of development of systems in the software product line using model driven methods. The intent of the case study will be to use these existing sources as a starting point and to adopt MOST technical advancement to increase productivity and software quality. MOST tools and approaches will be evaluated against scenarios representing real challenges of large scale software development.

Mission

Facing global competition, European software houses are striving to respond to customer's need for flexible, large scale, software systems that can be easily and quickly customized, configured and deployed. One of the main challenges during development of such systems is the volume and fragmentation of software-related artefacts and the level of necessary expertise to develop and maintain software products that increases with scale of the system. This problem is particularly challenging in development of software product lines where modularized software parts produced by remote developer teams are composed and customized to meet very specific customer requirements.

Hence, the objective of the Comarch case study is to apply and evaluate MOST technology in three areas:

1. Providing overview of software product line in context of model driven development to reduce the complexity of software development and leverage reuse.
2. Support for configuration of the software system by integrated configuration metadata management.
3. Provide guidance in the software engineering process using MOST technologies.

Use Scenarios

Considering the growing number of developers and products available in the suite, the need for effective communication among domain experts, developers and implementation teams grows. The promise to provide comprehensive overview of the product line using MOST technology has a great potential to improve collaboration of the stakeholders.

Equally important for Comarch is the MOST promise to increase the quality of produced software. Given that MOST tools will provide means to synchronize software artefacts, validate them and trace the impact of changes, as well as provide guidance through error prone activities, the benefits of adopting MOST technology would be better software quality.

Product Line Overview

The first area of the case study is to apply MOST technology to provide overview of software product lines in context of model driven development appliance will be maintaining and extending the core software models and systems by managing the different branches of MDA transformations in an integrated manner with information about the models. This description expressed in the ontology along with applied traceability techniques will provide overview of the whole software product line and simplify management dependencies between artefacts.

Equally important is to allow composing domain models from smaller parts and clearly separate management of core domain ontologies, ontologies formalizing specific network technologies and ontologies describing customer specific extensions. The reuse process will be improved and its complexity will be reduced by applying traceability techniques to control the impact of changes in a reused ontology fragment on the other parts of the system. Since the dependencies between parts of the product line as well as the domain model definition itself change in time, the use case will provide an integrated strategy of version controlling of both aspects that will allow simultaneous development of domain ontologies for multiple customers as well as development of core ontologies.

Configuration Metadata Management

MOST technology will be also applied to provide a strategy for integrated configuration metadata mana gement. Corresponding to the given physical network management system, the software provided by Comarch will be adapted to customers' needs using ontological reasoning about the configuration and will prevent illegitimate software configurations.

Process Guidance

MOST research outcomes will be also adopted to provide guidance in the software engineering process. This will involve formalization of the customized software process used by Comarch in the area of Network Management Systems in a form of ontology. This guidance ontology will be used together with the product line ontology to provide step by step guidance when developing systems tailored to customer needs. Though this development process may vary depending on the size and complexity of the target customer system, it will always contain common types of activities such as partial system reuse, development of core parts, or customizations via specific MDA transformations. The process guidance will cover these core areas of software development for network management systems, benefiting from inference capabilities, such as reasoning about indirect and implicit dependencies.

Thus the customized product structure can be visualized and software engineers can be guided in such activities as installing or upgrading particular parts of the customized product.

MOST technology will also enable re a soning about software libraries and technologies basing on the customized product structure. For instance, when reaching the stage in the software architecture where use of libraries has been specified, the process guidance will ask for checking conformance of these libraries with project requirements (e.g. regulations on encryption, such as length of key, differ for various project targets according to country-specific regulations). Thus, quality checks are integrated into the software development process in a seamless and non-intrusive manner leading to improved software quality.

Evaluation

MOST tools and approaches will be evaluated against series of scenarios representing real challenges of large scale software development. The test cases will be performed both with and without using the MOST tools in comparative conditions to provide credible measurements. The evaluation will have two phases. In the part of this task the demonstrator will be produced to provide WP1-WP4 with feedback on their research outcomes performance. The second phase of evaluation will result in producing final evaluation report.

Download this text in PDF