
Migration Modeling and Learning Algorithms for Containers in Fog Computing
Abstract
Migration Modeling and Learning Algorithms for Containers in Fog Computing Such a mobility support mechanism can be critical, as in the industrial internet where humans, products, and devices are mobile.To fill in such gaps, we propose in this paper novel container migration algorithms and architecture to support mobility tasks with different application requirements.
System Configuration
Platform : cloud computing
Conclusion
In this paper, as a large-scale MDP problem, we have modeled the container migration problem of mobile application tasks in FC. We first define the system model whose cost function consists of delay, power consumption, and cost of migration. Then we proposed the deep container migration algorithms based on Q-learning. In Q-network update, we optimize domaction selection in exploration and DNN training strategy to achieve rapid decision-making.