VERA: Verifiable Aspect Models for Middleware Product Lines

Principal Investigator: Awais Rashid
Other Project Personnel: Gordon Blair (Co-investigator), Robert France (Visiting Fellow), Nelly Bencomo
Project Description
Related Publications
Grant Ref: EPSRC EP/E005276/1
Start Date: August 2006
End Date: January 2007

Project Description:

The overall aim of the VERA project is to develop a modelling framework for verifiable composition of aspect models pertaining to middleware families.
Middleware is emerging as an increasingly important technology in the construction of (often complex) distributed applications and services. The role of middleware is to mask out problems of heterogeneity and distribution and to provide a more abstract and transparent programming model to application developers. However, in recent years it has become much more difficult to provide comprehensive middleware platforms largely due to the significant rise in levels of heterogeneity. As well as traditional aspects of heterogeneity (language, platform, etc.), middleware providers must also deal with increasing heterogeneity of application domains (pervasive, multimedia, mobile, Grid, etc.) as well as similar rises in heterogeneity in the deployment environment (wireless, wired, ad-hoc, etc.). Furthermore, there is a need to support a wider variety of non-functional requirements in areas such as security and dependability, with more specialised solutions emerging, for example, in areas such as pervasive computing. It is now clear that middleware providers are struggling with such extreme heterogeneity and rather provide platforms that are overly bloated and complex to deal with the variety of situations where they might be used. It is equally clear that this is unsatisfactory and that a major paradigm shift is required to cater for such changing operational conditions.

The middleware community is reacting to these challenges and there is now strong interest in reflective middleware as a technique to support more configurability, and indeed reconfigurability. In this project, we focus exclusively on the problems of configurable middleware and, in particular, how middleware families can be generated using the inherent support provided by reflection. However, there is an element of complexity in dealing with the flexibility that reflection offers. It is therefore imperative that in parallel with developments in reflective middleware, we also study tools and techniques that support the automatic or semi-automatic generation of middleware families from higher level specifications. In this project, we study the marriage of three areas of technology in meeting this challenge:

1. Reflective middleware as discussed above;
2. Model driven engineering (MDE) to support the automatic construction of complex software from models (in our case middleware);
3. Aspect-oriented software development (AOSD) to provide a natural separation of concerns both in terms of the models and the resultant platform.

Publications:

• S. A. Lundesgaard, A. Solberg, J. Oldevik, R. France, J. O. Aagedal, and F. (2007) Construction and Execution of Adaptable Applications using an Aspect-Oriented and Model Driven Appro. 7th IFIP International Conference on Distributed Applications and Interoperable Systems (DAIS 2007).
• L. Yu, R. France, I. Ray, and K. Lano (2007) A light-weight static approach to analyzing UML behavioral properties. International Conference on the Engineering of Complex Computing Systems (ICECCS 2007), Auckland, New Zealand.
• R. France and B. Rumpe (2007) Model-driven Development of Complex Software: A Research Roadmap. International Conference on Software Engineering (ICSE 2007): Future of Software Engineering. Editor(s): L. Briand and A. Wolf, IEEE-CS Press.
• N. Bencomo, G. Blair, C. Flores-Cortes, and R. France (2007) A Model-based Approach to Managing Runtime Variability in Complex Distributed Systems. submitted to MoDELS'07, Nashville, USA.