Editorial: Experiences with Auto-adaptive and Reconfigurable Systems (System Level Adaptation)

For this special issue on Experiences with Auto-adaptive and Reconfigurable Systems, we received over 70 submissions. Due to the high volume of submissions, we decided to divide the theme into two sub-topics addressing system-level and application-level concerns respectively, and dedicate a separate part of the special issue to each topic. After a stringent selection process, six papers were selected per part. There is a growing demand for software systems to perform in dynamic environments, where both functional and non-functional requirements may change. Typically, real-time constraints should be fulfilled and the desired quality of service should be provided. Furthermore, software systems should be able to adapt both statically and dynamically to changing demands. A current emerging trend in software systems requires software entities to be able to reconfigure themselves seamlessly and without user intervention in reaction to changes in their environment, while continuing to provide a guaranteed service. However, the potential system configurations needed in a particular context can hardly be anticipated beforehand. Dynamic non-invasive approaches for providing adaptive system behaviour are therefore required. The development of these systems is a challenge from both the theoretical and practical perspective. The first part of this special issue on Experiences with Auto-adaptive and Reconfigurable Systems focuses on at system level adaptation, and is composed of the following papers. ‘Dynamic service adaptation’ by Robert Hirschfeld and Katsuya Kawamura is concerned with next generation mobile communication systems and their increasing complexities. The authors anticipate an emerging need for such systems to be open for third-party service providers. Since customers keep switching their service providers, a crucial characteristic of these systems is for their service portfolio to be adjustable to the needs and the preferences of their customers and service level agreements. The paper also suggests infrastructure support for propagating adaptation modules, to coordinate their activation and deactivation, to detect and resolve conflicts and to address safety and security concerns related to mobile code. ‘Adapting workload distribution on software DSM clusters’ by Yen-Tso Liu, Tyng-Yeu Liang and Ce-Kuen Shieh presents the development of a novel method for distributing program threads onto the individual computers of software distributed shared memory (SDSM) clusters. The proposed method has been implemented on a test bed and verified through application to a set of benchmark user programs. The experimental results have shown that the new approach, which takes into consideration both the memory availability of the processors and the CPU power when performing workload distribution, has a significant beneficial impact in terms of maximizing the performance of SDSM applications.