A framework for dependable QoS adaptation in probabilistic environments

Distributed protocols executing in uncertain environments, like the Internet, had better adapt dynamically to environment changes in order to preserve QoS. In a previous work, it was shown that QoS adaptation should be dependable, if correctness of protocol properties is to be maintained. In this paper we provide concrete strategies and methodologies to improve the implementation of dependable QoS adaptation. During its lifetime, a system alternates periods where its temporal behavior is well characterized, with transition periods where a variation of the environment conditions occurs. Our method is based on the following: if the environment is generically characterized in analytical terms, and we can detect the alternation of these stable and transient phases, we can drastically improve the effectiveness of dependable QoS adaptation. To prove our point, we conduct an evaluation based on "synthetic" data flows generated from one or more probabilistic distributions, and we show that the proposed strategies can indeed be effective and still dependable in the considered cases.