Self-Adaptive Middleware for Model-Based Network Adaptations
Martin Pfannemüller · 2018
Adding self-adaptive capabilities to systems is one way to reduce administrative effort and to cope with changing contexts at runtime. Trends in pervasive computing such as IoT increase the number of devices and the amount of context information which strengthens the need for self-adaptive software. Existing frameworks for building self-adaptive software can help to reduce development effort. However, modeling the reconfiguration behavior of a system including its planned execution context is still a challenging task. It should easily be possible for developers without knowledge about building adaptation logics to specify adaptation possibilities. Further, it would be beneficial for these interconnected devices to support network adaptations as concepts such as SDN can help to change networks at runtime. The goal of our research is to enhance distributed and pervasive systems by developing a middleware for defining network adaptations using a model-based approach. Utilizing such a middleware could help to always achieve the best network performance in distributed settings at runtime.