TY - JOUR
AU - Misra, Rajiv
AB - [Cloud computing leverages virtualization as the most popular technique to deploy enterprise applications on virtual machines. Since the cloud system dynamically adapts to workload changes depending on the time of the day. It is required to ensure elasticity as a robust technique to efficiently model the changing workload requirements. However, it is an extremely challenging task, as several users may enter and depart from the cloud system over time. Predicting the different resource usage metrics of dynamically arriving jobs can help the cloud service providers (CSPs) in better capacity planning to fulfill the service level agreements (SLAs). In this paper, we propose a k clustering-based stacked bidirectional LSTM (BiLSTM) deep learners to model the multi-variate resource usage predictions for highly varying cloud workloads. We evaluate the proposed model on the Google cluster trace and validate its performance with the current approaches.]
TI - Internet of Things and Connected Technologies: k Stacked Bidirectional LSTM for Resource Usage Prediction in Cloud Data Centers
DA - 2021-05-30
UR - https://www.deepdyve.com/lp/springer-journals/internet-of-things-and-connected-technologies-k-stacked-bidirectional-9GnTbw86SL
DP - DeepDyve
ER -