Attention-Enhanced Bidirectional Deep Learning for Short-Term CPU Usage Forecasting in Cloud Resource Allocation

Enukapally, Hemanth Kumar (2025) Attention-Enhanced Bidirectional Deep Learning for Short-Term CPU Usage Forecasting in Cloud Resource Allocation. Masters thesis, Dublin, National College of Ireland.

Preview	PDF (Master of Science) Download (1MB) | Preview
Preview	PDF (Configuration Manual) Download (585kB) | Preview

Abstract

Cloud computing environments have experienced highly dynamic and unpredictable workloads by making powerful and good CPU resource allocation as a challenge for maintaining performance, scalability, and cost efficiency. There are some traditional rule-based and reactive autoscaling mechanisms who fail to anticipate short-term workload fluctuations which results in resource underutilization or service degradation. This study solves the problem of short-term CPU usage forecasting for intelligent cloud resource allocation using deep learning–based predictive models. The study has been proposed an attention-enhanced Bidirectional Long Short-Term Memory model with Cross-Attention Tensor Fusion (BiLSTM-CATF) to improve forecasting accuracy by selectively focusing on the most relevant temporal dependencies in cloud workload time-series data. This study have implemented and deployed on an AWS cloud environment using EC2 for computation, Cloud9 for development and experimentation and SNS for performance notification and monitoring. This study shows that BiLSTM-CATF achieved the highest prediction error (RMSE = 0.0223, R² = 0.9709) which outperforms baseline models while capturing complex temporal patterns. This study advances the state of the art by combining cross-attention into bidirectional temporal modeling for cloud workloads.

Item Type:	Thesis (Masters)
Supervisors:	Name Email Gupta, Shaguna UNSPECIFIED
Uncontrolled Keywords:	Cloud Computing; CPU Usage Forecasting; Bidirectional LSTM; Attention Mechanism; Resource Allocation
Subjects:	Q Science > QA Mathematics > Electronic computers. Computer science T Technology > T Technology (General) > Information Technology > Electronic computers. Computer science T Technology > T Technology (General) > Information Technology > Cloud computing
Divisions:	School of Computing > Master of Science in Cloud Computing
Depositing User:	Ciara O'Brien
Date Deposited:	31 Aug 2026 12:55
Last Modified:	31 Aug 2026 12:55
URI:	https://norma.ncirl.ie/id/eprint/9698

Actions (login required)

View Item