Short-Term Load Forecasting of Virtual Machines Based on Improved Neural Network

Wei Guo,Wei Ge,Xudong Lu,Hui Li

doi:10.1109/access.2019.2936875

Abstract

Cloud computing allows applications to share resources on physical machines in an independent manner so as to effectively increase the utilization of physical machines. To enable cloud service users to obtain high-quality services, it is significant for cloud service providers to reduce the time required to allocate and deploy virtual machines, as well as to provide time windows for the deployment of physical servers. To ensure demand is met, one effective implementation method is to predict the workload of virtual machines in the future, as this contributes to the efficient deployment of a data center's requests from virtual machines to dynamically changing physical servers. However, due to the complexity of virtual machine requests, virtual machine workload prediction remains a significant challenge at present; on the one hand, it is difficult for standard recurrent neural networks to capture long-term dependencies because of the disappearance of gradients, while on the other hand, the long short-term memory method cannot handle irregular intervals. Accordingly, a new virtual machine workload prediction method that is capable of managing irregular time intervals is proposed in this paper. The proposed method combines the amount and time intervals of historical virtual machine work requests in order to accurately predict the virtual machine's future workload using a fixed length of time as a unit. Experiments show that the proposed model can generate more accurate prediction results than the long short-term memory method.

Highlights

There is a conspicuous imbalance between the resource demands of a single application at different times and the resource demand between different applications
To address the aforementioned challenges in short-term load forecasting of virtual machines, we propose an integrated approach to forecast workload using a novel Long Short-Term Memory (LSTM), referred to as N-LSTM, which is a modified LSTM architecture that takes the elapsed time into consideration between the consecutive elements of a sequence to adjust the memory content of the unit
According to LSTM training method, the training process is as follows: 1) COMPOSITE LAYER The processed request quantity X and time interval ∆t serve as the input of the combination layer, while the merged information S is obtained after the merge operation is complete

Summary

INTRODUCTION

There is a conspicuous imbalance between the resource demands of a single application at different times and the resource demand between different applications. This virtualization technology effectively increases the resource utilization rate of physical machines. By benefiting from the convenience of virtual machines, a data center can provide flexible resource allocation to meet the applications’ different performance requirements For this to be executed successfully, the workload of the virtual machines first needs to be well predicted. Guo et al.: Short-Term Load Forecasting of Virtual Machines Based on Improved Neural Network deployment time of virtual machines As it takes some time for cloud service providers to deploy physical servers, it is of great significance to predict the number of physical servers that will need to be added when the utilization rate of these servers reaches a certain threshold. The final section summarizes this paper and provides some suggestions for future research

RELATED WORK

EXPERIMENTS

SUMMARY AND RESEARCH PROSPECTS