Predict Resource Utilization Research Articles

CNN-LSTM is all you Need for Efficient Resource Allocation in Cloud Computing

Many organizations have embraced cloud computing in recent years to provide new services, easily expand their IT resources, and reduce the cost of their IT infrastructure. This has been made possible through the implementation of resource allocation strategies by cloud service providers. One of the major challenges during resource allocation is to minimize power consumption while ensuring the required Service Level Agreement (SLA). To solve this problem, a new approach to efficiently allocate resources in cloud computing while optimizing energy consumption and guaranteeing the required service level agreement has been proposed. The main idea of this proposal is to leverage the CNN-LSTM architecture to accurately predict resource utilization in order to make the appropriate resource allocation decision. The proposed solution was validated in two steps: step 1) a comprehensive set of statistical performance analysis and step 2) an intensive simulation of the solution for resource allocation using cloudSim Plus tool. The results of the experimentation demonstrated that the proposed solution can help cloud service providers achieve energy savings while guaranteeing the required SLA.

Towards Efficient Resource Utilization Forecasting in Acute Heat Events Using OMOP.

Predicting resource utilization can help to optimize the distribution of limited resources in the healthcare system. This requires different climatic and medical data from different sources, which can lead to problems with interoperability. In the paper we describe which data is needed for the prediction and how the data can be made interoperable using OMOP CDM.

Development and validation of a machine learning framework for improved resource allocation in the emergency department

ObjectiveThe Emergency Severity Index (ESI) is the most commonly used system in over 70% of all U.S. emergency departments (ED) that uses predicted resource utilization as a means to triage [1], Mistriage, which includes both undertriage and overtriage has been a persistent issue, affecting 32.2% of total ED visits [2]. Our goal is to develop a machine learning framework that predicts patients' resource needs, thereby improving resource allocation during triage. MethodsThis retrospective study analyzed ED visits from the Medical Information Mart for Intensive Care IV, dividing the data into training (80%) and testing (20%) cohorts. We utilized data available during triage, including patient vital signs, age, gender, mode of arrival, medication history, and chief complaint. Azure AutoML was used to create different machine learning models trained to predict the 144 target columns including laboratory panels and imaging modalities as well as medications required during patients' ED visits. The 144 models' performance was evaluated using the area under the receiver operating characteristic curve (AUROC), F1 score, accuracy, precision and recall. ResultsA total of 391,472 ED visits were analyzed. 144 Voting ensemble models were created for each target. All frameworks achieved on average an AUC score of 0.82 and accuracy of 0.76. We gathered the feature importance for each target and observed that ‘chief complaint’, among others, had a high aggregate feature importance across different targets. ConclusionThis study shows the high accuracy in predicting resource needs for patients in the ED using a machine learning model. This can greatly improve patient flow and resource allocation in already resource limited emergency departments.

An efficient dynamic migration and consolidation method of VMs based on improved K-nearest neighbor algorithm and attribute weighting

Virtual machine (VM) consolidation and migration that only consider current workload can result in excessive unnecessary migrations. To address this issue, a VM consolidation algorithm based on resource utilization prediction is proposed. An improved K-nearest neighbor (KNN) classification algorithm weighted by attribute inconsistency is proposed to predict the workload of both the host and the VMs. Firstly, two distributions are partitioned according to the neighboring relationship for comparing consistency. Then, an inconsistency evaluation function based on earth mover’s distance (EMD) is designed to measure the inconsistency between the neighboring sample set of each sample under each attribute and the equivalent partition refined by the decision attribute. Finally, the inconsistency level of the neighboring samples is transformed into the importance of the corresponding attribute to implement the attribute weighting KNN classifier. When selecting the source host and target host for VM migration, both current and predicted overloads are considered to avoid unnecessary VM migrations. Simulation tests were performed with random and realistic workloads, and the results show that the proposed method can reduce the overall energy consumption of the host, while also reducing service level agreement (SLA) violations and VM migration.

Predicting Resource Utilization Trends with Southern California Petabyte Scale Cache

Large community of high-energy physicists share their data all around world making it necessary to ship a large number of files over wide- area networks. Regional disk caches such as the Southern California Petabyte Scale Cache have been deployed to reduce the data access latency. We observe that about 94% of the requested data volume were served from this cache, without remote transfers, between Sep. 2022 and July 2023. In this paper, we show the predictability of the resource utilization by exploring the trends of recent cache usage. The time series based prediction is made with a machine learning approach and the prediction errors are small relative to the variation in the input data. This work would help understanding the characteristics of the resource utilization and plan for additional deployments of caches in the future.

Triage Accuracy in Pediatrics Using the Emergency Severity Index

IntroductionAlthough the Emergency Severity Index is the most widely used tool in the United States to prioritize care for patients who seek emergency care, including children, there are significant deficiencies in the tool’s performance. Inaccurate triage has been associated with delayed treatment, unnecessary diagnostic testing, and bias in clinical care. We evaluated the accuracy of the Emergency Severity Index to stratify patient priority based on predicted resource utilization in pediatric emergency department patients and identified covariates influencing performance. MethodsThis cross-sectional, retrospective study used a data platform that links clinical and research data sets from a single freestanding pediatric hospital in the United States. Chi-square analysis was used to describes rates of over- and undertriage. Mixed effects ordinal logistic regression identified associations between Emergency Severity Index categories assigned at triage and key emergency department resources using discrete data elements and natural language processing of text notes. ResultsWe analyzed 304,422 emergency department visits by 153,984 unique individuals in the final analysis; 80% of visits were triaged as lower acuity Emergency Severity Index levels 3 to 5, with the most common level being Emergency Severity Index 4 (43%). Emergency department visits scored Emergency Severity Index levels 3 and 4 were triaged accurately 46% and 38%, respectively. We noted racial differences in overall triage accuracy. DiscussionAlthough the plurality of patients was scored as Emergency Severity Index 4, 50% were mistriaged, and there were disparities based on race indicating Emergency Severity Index mistriages pediatric patients. Further study is needed to elucidate the application of the Emergency Severity Indices in pediatrics using a multicenter emergency department population with diverse clinical and demographic characteristics.

MSHGN: Multi-scenario adaptive hierarchical spatial graph convolution network for GPU utilization prediction in heterogeneous GPU clusters

Accurately predicting GPU utilization is crucial for effectively managing heterogeneous GPU clusters, yet existing prediction methods are tailored to homogeneous clusters or ignore the unique characteristics of heterogeneous ones. To address this problem, we propose the Multi-Scenarios Adaptive Hierarchical Spatial Graph Convolution Network (MSHGN) model. This model leverages the hierarchical relationships among users, tasks, and machines to construct multiple scenarios' undirected graphs and uses Graph Convolution Neural (GCN) to capture the spatial dependency relationships between the computational resources consumed during execution. The MSHGN model accurately predicts the future GPU utilization rates of each machine in a heterogeneous GPU cluster, achieving an RMSE of only 0.0101, an MAE of 0.0072, and an MAPE of 0.0641. We evaluate the MSHGN model on a real-world Alibaba dataset collected from a heterogeneous GPU cluster and find that it achieves superior accuracy and robustness in predicting resource utilization, outperforming other baseline models.

Multi-resource predictive workload consolidation approach in virtualized environments

The revolution of virtualization technologies and Cloud computing solutions has emphasized the need for energy-efficient and Service level agreement (SLA)-aware resource management techniques in cloud data centers. Workload consolidation in Infrastructure-as-a-Service (IaaS) providers allows for efficient utilization of hardware resources and reduced energy consumption by consolidating workloads onto fewer physical servers. To ensure successful workload consolidation, it is crucial for IaaS providers to carefully estimate the host state and identify overloaded and underloaded hosts, thereby avoiding overly aggressive consolidation. Existing proposals determine the host state depending on its current resource utilization or a single anticipated resource utilization value, and often consider only a single resource type of the host, such as CPU. These limitations may lead to unreliable host state estimations, resulting in excessive and needless service migrations between physical machines (PMs). This, in turn, can lead to extra delays in service execution, degraded performance, increased power consumption, and SLA violations. To address these challenges, we propose a workload consolidation approach that leverages a multi-resource and multi-step resource utilization prediction model. Based on this model, our overload and underload decision-making algorithms consider the forecasted future trend (sequence of future value) of each host resource's utilization, including CPU, memory, and bandwidth. Through extensive experimentations conducted with two real-world datasets, we demonstrate that our approach can significantly reduce power consumption, SLA violation rate, and the number of migrations compared to existing benchmarks.

Edge assisted, forecast integrated ensemble learning based service management scheme for delay minimization in smart cities applications

As the Internet of Things (IoT) is maturing as a technology, innovative and cross-domain IoT applications have seen smart cities being conceived and designed across the globe, though with data and resource management challenges, Quality of Service (QoS) fulfilment challenges among others. These could also be addressed by means of context-aware fog computing at the edge of the network and also by incorporating intelligence at the network edge. Since workload at fog nodes can anytime see sudden changes in demand, hence load migration among fog nodes becomes viable. However, improper migration can lead to further migrations, eventually decreasing performance. In this paper, we present a multi-channel queuing model based smart distributed service management approach and an intelligent resource-aware forecasting technique to predict the required context and resource management. The scheme accomplishes live migration by a resource-aware ensemble forecast method that used current and predicted resource utilization and their context availabilities to address the delay requirement for cross-domain IoT applications. The proposed management algorithms are simulated using CloudSim simulator and the efficacy of the obtained results confirm the superiority of the proposed methods.

A Frailty Index to Predict Mortality, Resource Utilization and Costs in Patients Undergoing Coronary Artery Bypass Graft Surgery in Ontario

BackgroundPeople living with frailty are vulnerable to poor outcomes and incur higher health care costs after coronary artery bypass graft (CABG) surgery. Frailty-defining instruments for population-level research in the CABG setting have not been established. The objectives of the study were to develop a preoperative frailty index for CABG (pFI-C) surgery using Ontario administrative data; assess pFI-C suitability in predicting clinical and economic outcomes; and compare pFI-C predictive capabilities with other indices. MethodsA retrospective cohort study was conducted using health administrative data of 50,682 CABG patients. The pFI-C comprised 27 frailty-related health deficits. Associations between index scores and mortality, resource use and health care costs (2022 Canadian dollars [CAD]) were assessed using multivariable regression models. Capabilities of the pFI-C in predicting mortality were evaluated using concordance statistics; goodness of fit of the models was assessed using Akakie Information Criterion. ResultsAs assessed by the pFI-C, 22% of the cohort lived with frailty. The pFI-C score was strongly associated with mortality per 10% increase (odds ratio [OR], 3.04; 95% confidence interval [CI], [2.83,3.27]), and was significantly associated with resource utilization and costs. The predictive performances of the pFI-C, Charlson, and Elixhauser indices and Johns Hopkins Aggregated Diagnostic Groups were similar, and mortality models containing the pFI-C had a concordance (C)-statistic of 0.784. Cost models containing the pFI-C showed the best fit. ConclusionsThe pFI-C is predictive of mortality and associated with resource utilization and costs during the year following CABG. This index could aid in identifying a subgroup of high-risk CABG patients who could benefit from targeted perioperative health care interventions.

Advanced data analytics modeling for evidence-based data center energy management

Over the past few decades, the demand for Data Center (DC) services has significantly increased due to the world’s growing need for internet access, social networking, and data storage. Data Centers are among the most energy-intensive businesses, so optimizing IT operations in DC requires energy-efficient techniques. This paper presents AI based modeling strategies for effective energy management with a particular emphasis on DC’s two most energy intensive systems (i.e., cooling and IT systems). This study addresses the issues of IT equipment performance degradation, inappropriate IT room thermal conditions, inefficient workload placement, and excessive energy waste. This research entails the application of machine learning for DC thermal classification, and deployment of deep learning models to predict resource utilization and energy consumption in DC. Furthermore, a comparative analysis is performed with existing relevant methods to demonstrate the effectiveness and accuracy of the proposed AI techniques. The findings of this study also provide evidence-based recommendations for DC efficient energy management.

Electricity Consumption Forecasting System Using Deep Learning

Energy modeling in Smart Buildings (SB) and planning and operating the generation of power based on the information extracted are key components of the Smart Grid’s (SG’s) energy management system. In the world, buildings use a significant amount of energy that contributes to energy efficiency programs. Additionally, excessive utilization power generation appliance also including air conditioners and heater, breathing, and climate control (HVAC) units, improper microclimate control, and inappropriate start-up and ordering of power equipment waste a lot of heat pumps. The utility can mitigate energy generation costs when it anticipates electrical loads and schedules generation resources in accordance with the demand. To estimate electricity usage at varying tiers of utility grid systems, a range of techniques have already been used. The goal of this study will be to create a hybrid deep learning model can predict resource utilization in infrastructures. Model building and data cleaning are the two stages of the proposed framework. Data cleaning involves pre-processing techniques and adding additional lag values to raw data. This hybrid deep learning (DL) approach is made up of a series of completely connected layered and linear Long Short-Term Memory (LSTM) sections layered over bi-directional Long Short-Term Memory (LSTM) components and is based also on collected information. You could incorporate the dependence structure of electricity usage on regressors and increase computation efficiency, training time, as well as computational complexity by using the results obtained.

GMM-LSTM: a component driven resource utilization prediction model leveraging LSTM and gaussian mixture model

GMM-LSTM: a component driven resource utilization prediction model leveraging LSTM and gaussian mixture model

Utilization prediction-based VM consolidation approach

Reducing energy consumption and optimizing resource usage in large cloud data centers is still an essential target for the current researchers and cloud providers. The state-of-the-art highlights the effectiveness of VM consolidation and live migrations in achieving reasonable solutions. However, most proposals consider only the real-time workload variations to decide whether a host is overloaded or underloaded, or to trigger migration actions. Such approaches may apply frequent and needless VM migrations leading to energy waste, performance degradation, and service-level agreement (SLA) violations. In this paper, we propose a consolidation approach based on the resource utilization prediction to determine the overloaded and underloaded hosts. The prediction method combines a Kalman filter and support vector regression (SVR) to forecast the host's future CPU utilization. Simulations are conducted on Cloudsim using real PlanetLab workloads to verify the performance of our proposal against existing benchmark algorithms. Experimental results demonstrate that our consolidation technique significantly reduces the SLA violation rate, number of VM migrations, and energy consumed in the datacenter.

Allocation of Resources for Cloud Survivability in Smart Manufacturing

With the development of virtualization technology, cloud computing has emerged as a powerful and flexible platform for various services such as online trading. However, there are concerns about the survivability of cloud services in smart manufacturing. Most existing solutions provide a standby Virtual Machine (VM) for each running VM. However, this often leads to huge resource waste because VMs do not always run at full capacity. To reduce resource waste, we propose a smart survivability framework to efficiently allocate resources to standby VMs. Our framework contains two novel aspects: (1) a prediction mechanism to predict the resource utilization of each VM in order to reduce the number of standby VMs; and (2) a nested virtualization technology to refine the granularity of standby VMs. We will use an open-source cloud simulation platform named cloudsim, with real-world data, to verify the feasibility of the proposed framework and evaluate its performance. The proposed Smart Survivable Usable Virtual Machine (SSUVM) will predict resource utilization of VMs on Rack1 periodically. When errors happen in VMs, the framework will allocate standby resources according to the predicted result. The SSUVM will receive the latest running status of the failed VM and its mirror image to recover the VM's work.

Assessing Frailty Risk in Clinical Practice

Assessing Frailty Risk in Clinical Practice

The Age-Adjusted Modified Frailty Index: An Improved Risk Stratification Tool for Patients Undergoing Primary Total Hip Arthroplasty.

Frailty and increasing age are well-established risk factors in patients undergoing total hip arthroplasty (THA). However, these variables have only been considered independently. This study assesses the interplay between age and frailty and introduces a novel age-adjusted modified frailty index (aamFI) for more refined risk stratification of THA patients. The American College of Surgeons National Surgical Quality Improvement Program database was queried from 2015 to 2019 for patients undergoing primary THA. First, outcomes were compared between chronologically younger and older frail patients. Then, to establish the aamFI, one additional point was added to the previously described mFI-5 for patients aged ≥73 years (the 75th percentile for age in our study population). The association of aamFI with postoperative complications and resource utilization was then analyzed categorically. A total of 165,957 THA patients were evaluated. Older frail patients had a higher incidence of complications than younger frail patients. Regression analysis demonstrated a strong association between aamFI and complications. For instance, an aamFI of ≥3 (compared to aamFI of 0) was associated with an increased odds of mortality (OR: 22.01, 95% confidence interval [CI] 11.62-41.68), any complication (OR: 3.50, 95% CI 3.23-3.80), deep vein thrombosis (OR: 2.85, 95% CI 2.03-4.01), and nonhome discharge (OR 9.61, 95% CI 9.04-10.21; all P < .001). Chronologically, older patients are impacted more by frailty than younger patients. The aamFI accounts for this and outperforms the mFI-5 in prediction of postoperative complications and resource utilization in patients undergoing primary THA.

Parsimonious machine learning models to predict resource use in cardiac surgery across a statewide collaborative

ObjectiveWe sought to several develop parsimonious machine learning models to predict resource utilization and clinical outcomes following cardiac operations using only preoperative factors. MethodsAll patients undergoing coronary artery bypass grafting and/or valve operations were identified in the 2015-2021 University of California Cardiac Surgery Consortium repository. The primary end point of the study was length of stay (LOS). Secondary endpoints included 30-day mortality, acute kidney injury, reoperation, postoperative blood transfusion and duration of intensive care unit admission (ICU LOS). Linear regression, gradient boosted machines, random forest, extreme gradient boosting predictive models were developed. The coefficient of determination and area under the receiver operating characteristic (AUC) were used to compare models. Important predictors of increased resource use were identified using SHapley summary plots. ResultsCompared with all other modeling strategies, gradient boosted machines demonstrated the greatest performance in the prediction of LOS (coefficient of determination, 0.42), ICU LOS (coefficient of determination, 0.23) and 30-day mortality (AUC, 0.69). Advancing age, reduced hematocrit, and multiple-valve procedures were associated with increased LOS and ICU LOS. Furthermore, the gradient boosted machine model best predicted acute kidney injury (AUC, 0.76), whereas random forest exhibited greatest discrimination in the prediction of postoperative transfusion (AUC, 0.73). We observed no difference in performance between modeling strategies for reoperation (AUC, 0.80). ConclusionsOur findings affirm the utility of machine learning in the estimation of resource use and clinical outcomes following cardiac operations. We identified several risk factors associated with increased resource use, which may be used to guide case scheduling in times of limited hospital capacity.

A Resource Utilization Prediction Model for Cloud Data Centers Using Evolutionary Algorithms and Machine Learning Techniques

Cloud computing has revolutionized the modes of computing. With huge success and diverse benefits, the paradigm faces several challenges as well. Power consumption, dynamic resource scaling, and over- and under-provisioning issues are challenges for the cloud computing paradigm. The research has been carried out in cloud computing for resource utilization prediction to overcome over- and under-provisioning issues. Over-provisioning of resources consumes more energy and leads to high costs. However, under-provisioning induces Service Level Agreement (SLA) violation and Quality of Service (QoS) degradation. Most of the existing mechanisms focus on single resource utilization prediction, such as memory, CPU, storage, network, or servers allocated to cloud applications but overlook the correlation among resources. This research focuses on multi-resource utilization prediction using Functional Link Neural Network (FLNN) with hybrid Genetic Algorithm (GA) and Particle Swarm Optimization (PSO). The proposed technique is evaluated on Google cluster traces data. Experimental results show that the proposed model yields better accuracy as compared to traditional techniques.

Deep learning-based multivariate resource utilization prediction for hotspots and coldspots mitigation in green cloud data centers

Deep learning-based multivariate resource utilization prediction for hotspots and coldspots mitigation in green cloud data centers