Using Neural Networks on Cloud Container’s Performance Comparison By R on Docker (ROCKER)

Recently, containerization could give much experience with just run command and the containers help make software deployment run faster and more efficient. Docker is a `container'. On this container could run within existing operating system and build an environment that could build and testing the code. This paper presents our experiment work, and the advantage using neural networks on the cloud container's based application, explores the features from it and to measure the performance from the host environment operating system and from the virtualization environment system. The comparison results from work metrics of two applications by R native environment and R on Docker (ROCKER) when processing the datasets, we find that computational time of neural networks from ROCKER when processing dataset with maximum time of 10 hidden neuron dataset is 29s.