A Workload Prediction Approach Using Models Stacking Based on Recurrent Neural Network and Autoencoder

Nguyen, Hoang Minh; Woo, Sungpil; Im, Jungho; Jun, Taejoon; Kim, Dae-Young

doi:10.1109/hpcc-smartcity-dss.2016.0133

Cited by 12 publications

(9 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Whereas, the later worked on VM workload prediction using N-LSTM or the novel LSTM and compared the results with other LSTM variants in terms of prediction accuracy but the training and testing time required was high in their proposed model. Nguyen et al [15] designed and implemented a new approach to predict workload by stacking Recurrent Neural Networks and Autoencoder on different datasets to compare prediction accuracy. Better prediction accuracy results may be possible if LSTM and autoencoder combination was used.…”

Section: A Machine Learning Methodsmentioning

confidence: 99%

See 1 more Smart Citation

Resource Utilization Prediction in Cloud Computing using Hybrid Model

Shivakumar¹,

Nagaraja²

2021

IJACSA

View full text Add to dashboard Cite

In cloud environment, maximum utilization of resource is possible with good resource management strategies. Workload prediction plays a vital role in estimating the actual resource required for successful execution of an application on cloud. Most of the existing works concentrated on predicting workloads which either showed clear seasonality/trend or for irregular workload patterns. This paper presents a new perspective in forecasting both seasonal and non-seasonal workloads. To accomplish this, a hybrid prediction model which is a combination of statistical and machine learning technique is proposed. Suppose the seasonality exists in the workload pattern, Seasonal Auto Regressive Integrated Moving Average (SARIMA) model is applied for prediction. For non-seasonal workloads Long Short-Term Memory networks (LSTM) or AutoRegressive Integrated Moving Average (ARIMA) model is used based on the results of normality test. This paper presents a prediction model which forecasts the actual resource required for diverse time intervals of daily, hourly and minutes utilization. The experimental results confirm that accuracy of the prediction of LSTM model outperformed ARIMA for irregular workload patterns. The SARIMA model accurately forecasts the resource usage for forthcoming days. This work actually helps the cloud service provider (CSP) to analyze the workload and predict accordingly to avoid over or under provisioning of the cloud resources.

show abstract

Section: A Machine Learning Methodsmentioning

confidence: 99%

“…Lesser MAPE value indicates better prediction accuracy in terms of percentage. It is defined as in (15).…”

Section: Evaluation Criteriamentioning

confidence: 99%

Resource Utilization Prediction in Cloud Computing using Hybrid Model

Shivakumar¹,

Nagaraja²

2021

IJACSA

View full text Add to dashboard Cite

show abstract

“…To predict workload time series in cloud, this author [96] uses another performed ensemble learning approaches called stacking combining first the unsupervised autoencoder used to learn a "representation" of the input, and a recurrent neural network (RNN) layer trained in order to provide final predictions. In order to avoid the exploding and vanishing gradient issues in learning dependencies, three models were selected as candidates for this prediction model.…”

Section: Ensemble Learning Methodsmentioning

confidence: 99%

A Survey on the Current Challenges of Energy-Efficient Cloud Resources Management

et al. 2020

View full text Add to dashboard Cite

Given the perpetual surging of cloud services' requests, energy consumption of cloud data centers with their related CO 2 emissions still represents major issues. Efficient use of cloud's resources becomes then the driven force ensuring both, the satisfaction of service-level agreements and the sobriety of cloud's energy consumption. This paper purports to survey for the first time, a comprehensive literature of some actual challenges facing various cloud's resources utilization scheduling approaches dealing with energy conservation. Indeed, cloud resources involve not only computing servers, but also a wide set of intra-and inter-cloud network's resources to be considered. These resources are mainly provisioned through two wellknown technologies: virtualization and/or containerization. As existing studies in the area of cloud resources provisioning are generally categorized into different groups, this research depicts a complete taxonomy of energy-efficient cloud resources scheduling. In this same perspective, we survey some recent efforts made in energy-efficient virtual and containerized resources, by emphasizing first the most used reactive then proactive techniques for managing the whole cloud resources energy efficiency scheduling.

show abstract

“…Reliable workload prediction of monitored devices becomes critical in order to proactively manage the capacity of connected infrastructure, mitigate cyber security risks and simply respond early to the anomalous behaviour of the monitored IT infrastructure [1]. Accurate forecasting of the future host workload plays also a central role for robust scheduling and resources management in data centers and cloud computing and among many expected benefits could lead to reduced operational cost, for example in a form of eliminated or cut idle time of the devices [2], [3], [4].…”

Section: Introductionmentioning

confidence: 99%

“…In [2], an adaptive model was developed for highly-variable workloads prediction by integrating a Top-Sparse Auto-Encoder (TSA) and Gated Recurrent Unit (GRU) blocks into RNN. In [3], workload sequences in Cloud and Grid systems were predicted by developing a model of stacking prediction algorithms using RNN and Autoencoder. An approach based on the Long Short-Term Memory (LSTM) encoder-decoder network with attention mechanism was proposed in [11].…”

Section: Introductionmentioning

confidence: 99%

Deep Bi-Directional LSTM Networks for Device Workload Forecasting

Ruta

Cen

Vu³

2020

Annals of Computer Science and Information Systems

View full text Add to dashboard Cite

Deep convolutional neural networks revolutionized the area of automated objects detection from images. Can the same be achieved in the domain of time series forecasting? Can one build a universal deep network that once trained on the past would be able to deliver accurate predictions reaching deep into the future for any even most diverse time series? This work is a first step in an attempt to address such a challenge in the context of a FEDCSIS'2020 Competition dedicated to network device workload prediction based on their historical time series data. We have developed and pre-trained a universal 3-layer bi-directional Long-Short-Term-Memory (LSTM) regression network that reported the most accurate hourly predictions of the weekly workload time series from the thousands of different network devices with diverse shape and seasonality profiles. We will also show how intuitive human-led post-processing of the raw LSTM predictions could easily destroy the generalization abilities of such prediction model.

show abstract

A Workload Prediction Approach Using Models Stacking Based on Recurrent Neural Network and Autoencoder

Cited by 12 publications

References 17 publications

Resource Utilization Prediction in Cloud Computing using Hybrid Model

Resource Utilization Prediction in Cloud Computing using Hybrid Model

A Survey on the Current Challenges of Energy-Efficient Cloud Resources Management

Deep Bi-Directional LSTM Networks for Device Workload Forecasting

Contact Info

Product

Resources

About