Optimize Resource Usage Research Articles

Neural network inspired efficient scalable task scheduling for cloud infrastructure

The rapid development of Cloud Computing in the 21st Century is landmark occasion, not only in the field of technology, but also in the field of engineering and services. The development in cloud architecture and services has enabled fast and easy transfer of data from one unit of a network to other. Cloud services support the latest transport services like smart cars, smart aviation services and many others. In the current trend, smart transport services depend on the performance of cloud Infrastructure and its services. Smart cloud services derive real time computing and allows it to make smart decision. For further improvement in cloud services, cloud resource optimization is a vital cog that defines the performance of cloud. Cloud services have certainly aimed to make the optimum use of all available resources to the become as cost efficient and time efficient as possible. One of the issues that still occur in multiple Cloud Environments is a failure in task execution. While there exist multiple methods to tackle this problem in task scheduling, in the recent times, the use of smart scheduling techniques has come to prominence. In this work, we aim to use the Harmony Search Algorithm and neural networks to create a fault aware system for optimal usage of cloud resources. Cloud environments are in general expected to be free of any errors or faults but with time and experience, we know that no system can be faultless. With our approach, we are looking to create the best possible time-efficient system for faulty environments, Where the result shows that the proposed harmony search-inspired ANN model provides least execution time, number of task failures, power consumption and high resource utilization as compared to recent Red fox and Crow search inspired models.

Optimal Usage of Resources through Quality Aware Scheduling in Containers based Cloud Computing Environment

For cloud computing, the Quality Aware Scheduling of Containers (QASC) model has been proposed for delay-sensitive tasks. Plan your cloud tasks. Typically, time constraints are met while resources are used effectively. It's a really difficult undertaking. In order to distribute containers more effectively, QASC takes into account a number of performance factors. Containers and their make-span logs, as well as input quality metrics like I/O-intensive workload, startup time, hot standby failure rate, and inter-container dependencies, are collected by the QASC model. A metric coefficient that indicates each container's overall rating is calculated by normalizing and averaging these values to determine it. In order to determine how well scheduling performed, the model also includes a quality coefficient that calculates this metric-coefficient threshold. It's also critical for QASC to be able to determine the remaining energy in each container, which represents its request capacity. In order to optimize cloud resources, energy use is also taken into account by the model. From the cloud-sim simulation, an experimental dataset including 50 containers and 1,200 internet protocol-capable users was employed. For the make-span ratio, round-trip time, and energy consumption analysis, this produced 20,000 data points. The RLSched, DSTS, and ADATSA models were contrasted with the QASC model. The outcomes showed that QASC performed better than these models in a number of crucial areas. Tasks may be managed better with the higher average make-span ratio and lower volatility. Its superior job scheduling and resource use were further demonstrated by its shorter round-trip durations and lower energy usage across loads. The QASC model is an extremely complex scheduling method for container-based systems and a significant advancement in cloud computing research. Its approaches and methods enable for more intelligent energy use as well as high-quality services while also improving system performance, particularly for tasks that are delay-sensitive.

Empowering Agricultural Ecosystems: Leveraging 5G IoT for Enhanced Product Integrity and Sustainable Ecological Environments

The study explores the change-making potential of 5G IoT technology in improving the security of agro-products and contributing to conservation of environments in the ecological field. Through the process of a series conducted in expression using the real agricultural settings, the study gives a visible proof that food traceability, quality assurance, resource management and environment conservation all gain the significant improvements. By relying on IoT sensors utilizing the latest 5G connectivity, stakeholders can have real-time data analytics and communication to take well-informed decisions, optimize resource usage, and manage the environmental risks. The analytics showed a noticeable boost in product traceability with real-time tracking, temperature monitoring during transportation, and automatic data logging rendering success rates above 95%. Resource management benefits are demonstrated in H2O and fertilizer reduction, reaching 30%, and up to 20% higher yields of crops than is traditional. Ecological conservation results exhibit the improved ecosystem health and the possibility to detect environmental pollution in real-time allows operational conservation measures. The research emphasizes on the prospect of 5G IoT whose benefits might transform agricultural practices and guarantee the authenticity and sustainability of food in the face of fast-changing environmental conditions.

Role of Industry 5.0 for driving sustainability in the manufacturing sector: an emerging research agenda

PurposeWith ever-increasing global concerns over environmental degradation and resource scarcity, the need for sustainable manufacturing (SM) practices has become paramount. Industry 5.0 (I5.0), the latest paradigm in the industrial revolution, emphasizes the integration of advanced technologies with human capabilities to achieve sustainable and socially responsible production systems. This paper aims to provide a comprehensive analysis of the role of I5.0 in enabling SM. Furthermore, the review discusses the integration of sustainable practices into the core of I5.0.Design/methodology/approachThe systematic literature review (SLR) method is adopted to: explore the understanding of I5.0 and SM; understand the role of I5.0 in addressing sustainability challenges, including resource optimization, waste reduction, energy efficiency and ethical considerations and propose a framework for effective implementation of the I5.0 concept in manufacturing enterprises.FindingsThe concept of I5.0 represents a progressive step forward from previous industrial revolutions, emphasizing the integration of advanced technologies with a focus on sustainability. I5.0 offers opportunities to optimize resource usage and minimize environmental impact. Through the integration of automation, artificial intelligence (AI) and big data analytics (BDA), manufacturers can enhance process efficiency, reduce waste and implement proactive sustainability measures. By embracing I5.0 and incorporating SM practices, industries can move towards a more resource-efficient, environmentally friendly and socially responsible manufacturing paradigm.Research limitations/implicationsThe findings presented in this article have several implications including the changing role of the workforce, skills requirements and the need for ethical considerations for SM, highlighting the need for interdisciplinary collaborations, policy support and stakeholder engagement to realize its full potential.Originality/valueThis article aims to stand on an unbiased assessment to ascertain the landscape occupied by the role of I5.0 in driving sustainability in the manufacturing sector. In addition, the proposed framework will serve as a basis for the effective implementation of I5.0 for SM.

Exploiting Graphics Processing Units to Speed Up Subgraph Enumeration for Efficient Graph Pattern Mining GraphDuMato

Subgraph selection involves searching an input graph for subgraphs with a certain attribute. Graph pattern mining (GPM) relies on this technique, despite its computing complexity and rapid growth. Problems with uncoalesced access to memory, separation, and strain unbalance make efficient subgraph identification on GPUs a huge problem, given GPUs' success in speeding up operations across multiple industries. It is surprising that these challenges have not received sufficient attention in earlier research. This work presents new methods for effectively building and running subgraph enumeration on GPUs. Optimization of computational resource usage is achieved by combining a warp-centric design with a Depth first search (DFS) style approach. Memory efficiency, execution divergence, and GPU activity parallelization could all be enhanced by integrating these two methods. An affordable load balancing level is also incorporated for the purpose of dispersing work among thread warps, which further decreases GPU idleness. The GraphDuMato system facilitates the utilization of GPM methods with its intuitive application programming interface (API). Testing proves that GraphDuMato can outperform state-of-the-art GPM algorithms on a regular basis and may mine subgraphs with up to twelve trees.

Hierarchical Resource Orchestration Framework for Real-time Containers

Container-based virtualization is a promising deployment model in fog and edge computing applications, because it allows a seamless co-existence of virtualized applications in a heterogeneous environment without introducing significant overhead. Certain application domains (e.g., industrial automation, automotive, or aerospace) mandate that applications exhibit a certain degree of temporal predictability. Container-based virtualization cannot be easily used for such applications, since the technology is not designed to support real-time properties and handle temporal disturbances. This article proposes a framework consisting of a static offline and a dynamic online phase for resource allocation and adaptive re-dimensioning of real-time containers. In the offline phase, the optimal initial deployment and dimensioning of containers are decided based on ideal system models. Additionally, to adapt to dynamic variations caused by changing workloads or interferences, the online phase adapts the CPU usage and limits of real-time containers at runtime to improve the real-time behavior of the real-time containerized applications while optimizing resource usage. We implement the framework in a real Linux-based system and show through a series of experiments that the proposed framework is able to adjust and re-distribute computing resources between containers to improve the real-time behavior of containerized applications in the presence of temporal disturbances while optimizing resource usage.

Quick-Response Model for Pre- and Post-Disaster Evacuation and Aid Distribution: The Case of the Tula River Flood Event

Background: In the context of humanitarian logistics, efficiently evacuating people from disaster-stricken areas is a complex challenge. This study focuses on the Tula River region in Hidalgo, Mexico, exploring the evacuation and support of individuals in temporary shelters. Despite the fact that the topic has been addressed in the literature, it is necessary to have quick response methods that can be used by decision-makers to adapt and utilize existing spaces as temporary shelters, in addition to knowing how to evacuate people. Methods: Addressing this void, a methodology to minimize evacuation and aid distribution costs is introduced. Leveraging existing algorithms, particularly Integer Linear Programming, the model determines shelter activation and utilizes the Vehicle Routing Problem to assess aid delivery strategies. Results: The research identifies optimal evacuation routes from 13 affected areas to 34 shelters and analyzes aid distribution costs under various demand scenarios: original, increased, and decreased by 10%, based on the number of transport units allocated and Google Maps distances. It also evaluates the costs associated with humanitarian aid distribution under varying collection strategies, involving state and municipal governments. Conclusion: This approach provides a decision-making foundation and can be adapted for similar analyses in other communities during extreme events.

A Practical Approach to Increase Crop Production Using Wireless Sensor Technology

Introduction; The global demand for food production continues to rise due to the growing population and changing consumption patterns. Traditional agricultural practices often fail to meet this demand efficiently, leading to the exploration of innovative technologies to enhance crop productivity. Wireless sensor technology (WST) has emerged as a promising tool to monitor and optimize agricultural practices, providing real-time data on various environmental parameters crucial for crop growth. Objective; This study aims to evaluate the effectiveness of wireless sensor technology in increasing crop production. By integrating WST into conventional farming practices, we seek to optimize resource usage, reduce waste, and improve crop yields. Methods; We have proposed an IoT-enabled soil nutrient classification and crop recommendation model to recommend crops. By incorporating machine learning, artificial intelligence (AI), the cloud, sensors, and other automated equipment into the decision-assisting system, farmers will be able to take decisive actions without relying entirely on regional farming offices. Results; The analysis showed that the plot using wireless sensor technology exhibited a significant increase in crop yield compared to the traditional plot. Soil moisture levels were maintained within optimal ranges, leading to better water usage efficiency. Additionally, the automated system adjusted fertilizer application based on real-time soil nutrient data, resulting in improved plant health and productivity. Conclusions; The integration of wireless sensor technology in agriculture presents a practical and effective approach to increase crop production. This technology enables precise monitoring and management of critical growth parameters, resulting in higher yields and more efficient resource use. Adopting WST can significantly contribute to meeting the global food demand while promoting sustainable farming practices.

Revolutionizing Agriculture: Harnessing the Power of Artificial Intelligence for Sustainable Farming Practices

Artificial Intelligence (AI) has emerged as a groundbreaking technology with the potential to revolutionize various sectors, and agriculture stands as one of its prime beneficiaries. This paper delves into the transformative impact of AI in agriculture, exploring its multifaceted applications, benefits, and implications.AI-powered technologies, such as machine learning, robotics, and data analytics, are increasingly being integrated into agricultural practices, offering innovative solutions to age-old challenges. One of the pivotal areas where AI is making significant strides is in precision agriculture. Through the utilization of drones, sensors, and satellite imagery, AI enables farmers to collect real-time data on soil quality, crop health, and weather conditions. By leveraging this data, farmers can make informed decisions regarding irrigation, fertilization, and pest management, thereby optimizing resource usage and enhancing crop yields.Furthermore, AI-driven predictive analytics empower farmers to anticipate crop diseases, pests, and adverse weather events, allowing for proactive mitigation strategies. This predictive capability not only minimizes yield losses but also reduces reliance on chemical inputs, fostering sustainable agricultural practices. Moreover, AI facilitates autonomous farming machinery equipped with advanced sensors and actuators, enabling precise and efficient field operations. This not only enhances productivity but also alleviates the labor-intensive nature of traditional farming, addressing the challenge of labor scarcity in many agricultural regions.The impact of AI in agriculture extends beyond farm-level operations to supply chain management and market forecasting. AI algorithms analyze vast volumes of data encompassing market trends, consumer preferences, and climatic patterns to provide insights into optimal planting schedules, pricing strategies, and distribution logistics. By optimizing supply chain processes, AI enhances efficiency, minimizes wastage, and ensures food security.However, the widespread adoption of AI in agriculture raises pertinent concerns regarding data privacy, cybersecurity, and socioeconomic disparities. The collection and utilization of vast amounts of agricultural data necessitate robust data governance frameworks to safeguard farmer privacy and prevent unauthorized access. Additionally, the digital divide between large-scale commercial farms and smallholder farmers must be addressed to ensure equitable access to AI technologies. Governments, policymakers, and industry stakeholders need to collaborate to develop inclusive policies and infrastructure that foster the equitable deployment of AI in agriculture.In conclusion, the integration of AI in agriculture represents a paradigm shift towards more sustainable, efficient, and resilient food production systems. By harnessing the power of AI-driven technologies, farmers can mitigate risks, optimize resource utilization, and improve livelihoods. However, realizing the full potential of AI in agriculture requires concerted efforts to address regulatory, ethical, and socioeconomic challenges. Through strategic collaboration and innovation, AI has the potential to transform agriculture, ensuring food security and sustainability for future generations.

Performance evaluation with scalability varying on energy consumption in S-DYMO, M-DYMO and FB-B-DYMO routing algorithm in MANETs

Mobile Ad Hoc Networks (MANETs) are dynamic and self-configuring networks characterized by node mobility and decentralized operations. Efficient routing protocols play a crucial role in ensuring reliable communication while optimizing resource usage, particularly energy consumption. This paper presents a comparative performance evaluation of three prominent routing algorithms, such as S-DYMO, M-DYMO, and FB-B-DYMO, focusing on their scalability and energy efficiency in MANET environments. This study utilizes simulation-based experiments to assess the performance metrics of S-DYMO, M-DYMO, and FB-B-DYMO. Key parameters evaluated include packet delivery ratio (PDR), end-to-end delay, routing overhead, and energy consumption under varying network sizes and mobility scenarios. Each algorithm's ability to scale with network size and maintain performance under dynamic conditions is analyzed to understand its suitability for real-world MANET deployments. The results indicate that S-DYMO exhibits robust scalability and efficient energy utilization in small to medium-sized networks, leveraging its proactive route discovery approach. Conversely, M-DYMO demonstrates competitive performance in larger networks due to its reactive route maintenance strategy, which balances energy efficiency with scalability challenges. FB-DYMO, integrating feedbackbased mechanisms, shows promising results in mitigating routing overhead and energy consumption but exhibits sensitivity to network dynamics. The findings provide insights into the trade-offs between scalability and energy consumption inherent in S-DYMO, M-DYMO, and FB-B-DYMO routing algorithms, offering guidance for selecting appropriate protocols based on specific MANET deployment scenarios. Future research directions include enhancing protocol adaptability to dynamic network conditions and optimizing energy-efficient routing strategies tailored for evolving MANET environments.

Machine learning approaches for efficient energy utilization in cloud data centers

Cloud computing technology provides access on demand to virtualized resources, services, and applications via a distributed network. In cloud data centers effective energy utilization is a critical concern in today’s technology-driven world. Cloud data centers (CDC) are massive facilities that host and manage an enormous amount of data and computing resources. This article addresses the growing significance and energy-intensive nature of data centers. Due to the rapid growth of cloud computing, it offers on-demand access to resources globally and leads to substantial power consumption and carbon impact on the environment. They consume substantial amounts of energy, and optimizing their energy utilization is essential for reducing operational costs, minimizing environmental impact, and ensuring sustainable growth. To combat this, efficient energy-saving approaches using machine learning methods have been researched. ML methods hold great potential for enhancing energy efficiency in CDCs by analysing data, detecting patterns, and optimizing resource usage. The focus areas include CPU usage prediction, overload finding, underload estimation, selection, migration, and relocation of VMs to attain resources and improve energy utilization. The paper compares energy-saving results achieved by different machine learning techniques in data centers to minimize energy usage and meet service level agreements (SLA). The machine learning approaches reduce energy consumption from 1.6% to 88.5% compared to the benchmark approach mentioned, considering various settings and parameters.

Enhanced desalination with polyamide thin-film membranes using ensemble ML chemometric methods and SHAP analysis.

Addressing global freshwater scarcity requires innovative technological solutions, among which desalination through thin-film composite polyamide membranes stands out. The performance of these membranes plays a vital role in desalination, necessitating advanced predictive modeling for optimization. This study harnesses machine learning (ML) algorithms, including support vector machine (SVM), neural networks (NN), linear regression (LR), and multivariate linear regression (MLR), alongside their ensemble techniques to predict and enhance average water flux (AWF) and average salt rejection (ASR) essential metrics of desalination efficiency. To ensure model interpretability and feature importance analysis, SHapley Additive exPlanations (SHAP) were employed, providing both global and local insights into feature contributions. Initially, the individual models were validated, with NN demonstrating superior performance for both AWF and ASR, achieving the lowest mean absolute error (MAE = 0.001) and root mean squared error (RMSE = 0.0111) for AWF and an MAE = 0.0107 and RMSE = 0.0982 for ASR. The accuracy of predictions improved significantly with ensemble models, as evidenced by the near-perfect Nash-Sutcliffe efficiency (NSE) values. Specifically, the NN ensemble (NN-E) and Linear Regression ensemble (LR-E) reached an MAE and RMSE of 0.001 and 0.0111, respectively, for AWF. For ASR, NN-E reduced the MAE to 0.0013 and the RMSE to 0.0089, while LR-E maintained competitive performance with an MAE of 0.0133 and an RMSE of 0.0936. SHAP analysis revealed that features such as MDP and TMC were critical drivers of performance, with MDP showing the most significant positive impact on ASR. These findings demonstrate the dominance of ensemble methods over individual algorithms in predicting key desalination parameters. The enhanced precision in estimating AWF and ASR offered by these neuro-intelligent ensembles, combined with the interpretability provided by SHAP analysis, can lead to significant environmental and operational improvements in membrane performance, optimizing resource usage and minimizing ecological impacts. This study paves the way for integrating intelligent ML ensembles and SHAP-based interpretability into the practical field of membrane technology, marking a step forward toward sustainable and efficient desalination processes.

Scheduling Molds Manufacturing Processes through Process Mining

The molds industry, being a highly competitive sector, is constantly driven by organizations looking to optimize their business processes to gain a competitive edge. To access advanced management tools, there has been a strong push for integrating Industry 4.0 technologies, which in turn has resulted in a significant increase in the amount of manufacturing process data, enabling more accurate and real-time analysis of processes. However, challenges persist due to the unpredictable and ad-hoc nature of this industry’s business processes. As a result, effective resource management has become a crucial aspect that organizations are actively addressing. In this study, our objective is to leverage process mining algorithms and complement them with a genetic algorithm to generate efficient scheduling solutions to produce mold parts on factory floors. By analyzing existing data and identifying the workflows through process mining, we seek to gain a more realistic perspective on the mold manufacturing business process. Through this analysis, our aim is to uncover meaningful workflows and sequences for producing mold parts, ensuring that critical activities are appropriately incorporated while minimizing unnecessary repetitions. By identifying these purposeful sequences, we can provide the factory floor with schedules that optimize resource usage and enhance the overall efficiency of mold manufacturing processes.

Exploring Blockchain Applications in the Transition to a Circular Economy: A Comprehensive Sector-Focused Review

Industry 4.0 technologies have become one of the main enablers for implementing Circular Economy e↵orts, creating smart and connected products, processes, and systems that can optimize resource usage, reduce waste, and improve efficiency, with a significant environmental impact. Blockchain is a technology that was initially centered on authentication and security applications. Currently, its utility has extended to solving multiple problems across industries. This paper explores the di↵erent applications of the Blockchain related to the circular economy. A systematic literature review method was applied to analyze the main objectives, advantages, disadvantages, opportunities, and challenges of implementing Blockchain in various industrial sectors. Results show a growing trend in applications with favorable impacts, especially in sectors like Manufacturing, Energy, and Automotive.

Enabling Sustainable Learning: A Machine Learning Approach for an Eco-friendly Multi-factor Adaptive E-Learning System

Adaptive learning seeks to create personalized learning experiences by considering various cognitive and affective factors. However, conventional adaptive models often fall short in meeting diverse learner needs, relying heavily on single factors like learning style. In response, we present a comprehensive framework that integrates an AI-based adaptive learning model, which not only accounts for multiple factors such as prior performance, leisure interests, and learning style but also aligns with principles of green smart education. Leveraging the k-means clustering algorithm, our approach brings together learners with similar leisure interests. Predicting student performance involves a Gradient Boosting Regressor, with demographic data and past performance contributing to a performance metric. Additionally, our system incorporates sustainability practices, optimizing resource usage in computation and data storage to promote eco-friendliness in education. Artificial neural networks predict individual learning styles, and a decision tree algorithm personalizes educational content delivery to align with preferences. Our objective is twofold: to enhance the overall performance of the proposed model and to champion sustainability in education, fostering a greener and more adaptive learning ecosystem.

Enhancing efficiency and adaptability in mixed model line balancing through the fusion of learning effects and worker prerequisites

This research introduces a comprehensive scheme to tackle the Mixed-Model Assembly Line Balancing Problem (MALBPLW) within manufacturing contexts. The primary aim is to optimize assembly line task assignments by integrating both the learning effect and worker prerequisites. The learning effect recognizes the enhanced efficiency of workers over time due to learning and experience. A novel mathematical model and solution approach are proposed, encompassing factors like cycle time, task interdependencies, worker classifications, and the learning effect. The model endeavors to minimize the overall costs related to both workers and workstations while simultaneously maximizing production efficiency. Experimental assessments are conducted to evaluate the efficacy of this proposed approach. Diverse manufacturing scenarios are inspected, comparing and analyzing cost reductions and production efficiency. The outcomes highlight the effectiveness of this approach in achieving enhanced cost-effectiveness and resource utilization in contrast to conventional methods. This study contributes significantly to advancing assembly line balancing and production planning techniques by presenting a pragmatic framework for optimizing resource usage and reducing costs in manufacturing environments. The knowledge extracted from these discoveries can significantly assist professionals in the industry seeking to improve manufacturing processes and strengthen competitiveness.

Study on advancement in bioinformatics for crop improvement: integrating genomics, phenomics, and machine learning

Bioinformatics has emerged as a vital tool in agriculture, performing an important role in crop improvement and precision agriculture. This review study explores the applications of bioinformatics in these areas, highlighting the importance of next-generation sequencing (NGS) and third-generation sequencing technologies like PacBio and ONT. The integration of bioinformatics and genomics has revolutionized crop improvement strategies, facilitating the creation of crop varieties that are more suitable for evolving climates. The review emphasizes the creation of integrated crop databases for managing large-scale genotype and phenotype data and discusses the role of bioinformatics in precision agriculture for optimizing resource usage and minimizing environmental impact. Multiomics technologies are highlighted for gaining insights into plant biology and facilitating informed decisions in crop improvement. Machine learning algorithms play a prominent role in interpreting diverse and complex datasets generated by imaging or sequencing, aiding in efficient plant phenotyping and the identification of associations between DNA sequences and traits. The review concludes by emphasizing the importance of collaboration, data integration, and the potential of bioinformatics in advancing agriculture.

Rapid 3D modelling: Clustering method based on dynamic load balancing strategy

AbstractThree‐dimensional (3D) reconstruction is a pivotal research area within computer vision and photogrammetry, offering a valuable foundation of data for the development of smart cities. However, existing methods for constructing 3D models from unmanned aerial vehicle (UAV) images often suffer from slow processing speeds and low central processing unit (CPU)/graphics processing unit (GPU) utilization rates. Furthermore, the utilization of cluster‐based distributed computing for 3D modelling frequently results in inefficient resource allocation across nodes. To address these challenges, this paper presents a novel approach to 3D modelling in clusters, incorporating a dynamic load‐balancing strategy. The method divides the 3D reconstruction process into multiple stages to lay the groundwork for distributing tasks across multiple nodes efficiently. Instead of traditional traversal‐based communication, this approach employs gossip communication techniques to reduce the network overhead. To boost the modelling efficiency, a dynamic load‐balancing strategy is introduced that prevents nodes from becoming overloaded, thus optimizing resource usage during the modelling process and alleviating resource waste issues in multidevice clusters. The experimental results indicate that in small‐scale data environments, this approach improves CPU/GPU utilization by 35.8%/23.4% compared with single‐machine utilization. In large‐scale data environments for cluster‐based 3D modelling tests, this method enhances the average efficiency by 61.4% compared with traditional 3D modelling software while maintaining a comparable model accuracy.

Optimized resource usage with hybrid auto-scaling system for knative serverless edge computing

In the most popular serverless platform - Knative, dynamic resource allocation is implemented using horizontal auto-scaling algorithms to create or delete service instances based on different monitored metrics. However, the assigned resources for each instance are fixed. Vertical scaling up or down assigned resources per instance is required to avoid over-provisioning resources which are limited at the edge. Hybrid (horizontal and vertical) auto-scaling solutions proposed by existing works have several limitations. These solutions are optimized for separated services and get degraded performance when applied in a normal environment with multiple concurrent services. Further, most methods make significant changes to the original Knative platform, and have not been considered to be adopted since then. In this article, instead of Knative modification, we developed separated Kubernetes operators and custom resources (CRs) that can assist the Knative auto-scaler with optimal hybrid auto-scaling configurations based on traffic prediction. First, we characterize each service with a profile of different assigned resource levels pairing with their optimal target Knative’s horizontal scaling request concurrency. Then, based on these profiles, we calculate the best-assigned resources level, target concurrency level, and the number of required instances corresponding to each future time step’s predicted traffic. Finally, these configurations are applied to Knative’s default auto-scaler and services’ CR. Experiments done on our testbed compared our solution with a Knative hybrid auto-scaler solution that does not consider the service’s target request concurrency, and the default Knative horizontal auto-scaler. The results show our solution improvements up to 14% and 20% in terms of resource usage, respectively.

Energy and Performance-Efficient Dynamic Consolidate VMs Using Deep-Q Neural Network

With cloud computing facing higher levels of big data than ever, the processor scale is rapidly expanding. Large clusters place a heavy burden on cloud service providers and the environment. High energy consumption decreases the economic benefits of cloud service providers, while enormous power demands pressure on the environment. The dynamic consolidation of virtual machines (VMs), which uses live migration technology to optimize resource usage and reduce energy consumption, is sufficient for saving energy while ensuring high performance with the desired level of quality of service (QoS) between cloud providers and users. In this paper, we propose a novel machine-learning algorithm called deep-Q neural network VM consolidation (DQNVMC) that combines the Q-leaning approach with deep learning neural network to find an approximately optimal solution. Furthermore, based on the real workload trace in the cloud environment, the experiments show that DQNVMC effectively reduces energy consumption while meeting the high performance of QoS requirements.