Nowadays data-intensive software systems have to meet user expectations in ever-changing execution environments. The increasing space of possible context states and the limited capacity of mobile devices make no longer possible to incorporate all necessary software functionalities and data in the system. Instead, the system database has to be adapted to successive context changes, in order to include all the information required at each stage. This adaptation process may translate into frequent and costly reconfigurations, in turn affecting negatively system stability and performance. This paper presents an approach to context-dependent database reconfiguration that aims to improve system stability by anticipating future information needs. The latter are specified by means of an annotated probabilistic task model, where each state is associated with a database subset. Experiments suggest that this approach has a positive impact on the stability of the system, the gain depending on the degree of similarity of the successive tasks in terms of database usage.
|Title of host publication||2nd International Conference on Context-Aware Systems and Applications|
|Subtitle of host publication||2nd International Conference on Context-Aware Systems and Applications|
|Publication status||Published - 2013|
- self-adaptive data-intensive systems
- context-aware schema-driven database adaptations schema
- adaptation stability
Mori, M., Cleve, A., & Inverardi, P. (2013). A Stability-aware Approach to Continuous Self-adaptation of Data-intensive Systems. In 2nd International Conference on Context-Aware Systems and Applications: 2nd International Conference on Context-Aware Systems and Applications Springer Verlag.