研究実績の概要 |
Surrounding the distributed virtual network function allocation in edge networks, I published three journal paper and seven conference paper in the past one year. I introduced three models for scheduled failures of physical machines in three different application environments with designing heuristic algorithms for solving the problems. Compared with non-model cases, the three models improve the continue servable time of services. For unpredictable failures, I introduced a model for keeping fault-tolerance of the network function allocations with a deep reinforcement learning approach for solving the model. In term of the implementations of the theoretical researches, I realized the allocations of real network functions in Kubernetes, which is an open-source and widely used system for managing the containerized applications, through controllers and scheduler in Kubernetes.
|