A Dependable Service Path Searching Method in Distributed Virtualized Environment Using Adaptive Bonus-Penalty Micro-Canonical Annealing

In Distributed Virtualized Environment, service components on a dependable service path will be selected to implement service composition. Searching for the optimal dependable service path is the key to implement dependability assurance, which is a Multi-Constrained Optimal Path problem. However, the existing algorithms have disadvantages of high complexity and low performance, and lacking the consideration of trust relationships and evidence spread among service components during service construction and composition. We proposed the concept of QoD, the Quality of Dependability, introducing some attributes(e.g. component intimacy) to describe and restrict the dependable service path searching in distributed virtualized environment. We also applied Adaptive Bonus-Penalty Micro-canonical Annealing(ABP-MA) to dependable service path searching, and chose service components on the optimal dependable service path to satisfy users' demands for service dependability. The experimental results showed that ABP-MA has the advantages of fast convergence and high search success rate.