Article
Computer Science, Information Systems
Yanxin Liu, Yao Zhao, Jian Dong, Lianpeng Li, Chunpei Wang, Decheng Zuo
Summary: In this paper, an Intelligent Neat framework (I-Neat) is proposed, which adds an intelligent scheduler using reinforcement learning and a framework manager to improve the usability of the system. The experimental results indicate that the intelligent scheduler and these novel algorithms can effectively reduce energy consumption with SLA assurance.
TSINGHUA SCIENCE AND TECHNOLOGY
(2022)
Article
Computer Science, Information Systems
Pengcheng Wei, Yushan Zeng, Bei Yan, Jiahui Zhou, Elaheh Nikougoftar
Summary: Virtualization technology represented by Virtual Machines (VMs) plays a crucial role in cloud computing infrastructure, and Virtual Machine Placement (VMP) poses challenges for energy efficiency in cloud data centers. This paper proposes VMP-A3C, a Deep Reinforcement Learning (DRL) based strategy, to optimize the allocation of VMs to Host Machines (HMs) in order to achieve load balancing and reduce energy consumption, while avoiding Service Level Agreement Violations (SLAV). The effectiveness of VMP-A3C is evaluated in terms of deployment rate, energy consumption, SLAV, and the numbers of shutdown HMs and migrated VMs. Notably, VMP-A3C outperforms the best existing method by reducing energy consumption by 2.54% and requiring 7.14% fewer HMs.
JOURNAL OF KING SAUD UNIVERSITY-COMPUTER AND INFORMATION SCIENCES
(2023)
Article
Computer Science, Theory & Methods
Wenbin Yao, Zhuqing Wang, Yingying Hou, Xikang Zhu, Xiaoyong Li, Yamei Xia
Summary: In the cloud computing environment, unbalanced utilization of multi-dimensional resources in physical servers leads to inefficient resource utilization and energy wastage in data centers. We propose a load balancing strategy based on virtual machine consolidation (LBVMC) to reduce energy consumption and service level agreement (SLA) violations. We present algorithms for load state classification, migratable VM selection based on multi-dimensional resource utilization, and VM placement based on resource fitness and load correlation. Experimental results demonstrate the superior performance of LBVMC in reducing energy consumption and SLA violations.
FUTURE GENERATION COMPUTER SYSTEMS-THE INTERNATIONAL JOURNAL OF ESCIENCE
(2023)
Article
Computer Science, Hardware & Architecture
Kashav Ajmera, Tribhuwan Kumar Tewari
Summary: Cloud data center serves tremendous workload demand due to the ever-increasing usage of internet services. Scheduling these workloads over physical servers is a combinatorial problem similar to an NP-complete problem. Dynamic workload changes result in high power consumption and SLA violation. This paper proposes a novel algorithm, ICSA-ROPE, which finds optimal VM schedules at each scheduling interval to minimize power consumption and ensure SLA. The implementation on a CloudSim simulator shows significant performance efficiency improvement compared to other algorithms.
Article
Computer Science, Information Systems
Seyyed Meysam Rozehkhani, Farnaz Mahan
Summary: With the continuous development of cloud computing as a successful business and the increasing interest of organizations in utilizing its environment, there is a high demand for provided services. However, the high maintenance costs and limited resources in the cloud computing environment make it essential to find an appropriate approach for controlling and managing resources. Existing exploratory methods have their own distinct features and may not evaluate all quality of service (QoS) criteria. Therefore, this study proposes a granular model that incorporates various computation criteria and utilizes past data to derive membership functions and inference rules. Experimental analysis demonstrates that this method outperforms other related methods and consistently delivers satisfactory performance across all QoS criteria.
INFORMATION SCIENCES
(2022)
Article
Computer Science, Information Systems
Minhaj Ahmad Khan
Summary: The proposed normalization-based VM consolidation strategy aims to minimize energy consumption, SLA violations, and the number of VM migrations in cloud environments. The strategy utilizes resource parameters to identify overloaded hosts and compares the capacity of virtual machines and hosts. Experimental results demonstrate that this approach outperforms other well-known methods in reducing energy consumption, SLA violations, and the number of VM migrations.
CLUSTER COMPUTING-THE JOURNAL OF NETWORKS SOFTWARE TOOLS AND APPLICATIONS
(2021)
Article
Computer Science, Artificial Intelligence
Mohamed A. Elshabka, Hanan A. Hassan, Walaa M. Sheta, Hany M. Harb
Summary: This study introduces a Security-aware Dynamic VM Consolidation (SDVMC) approach that minimizes overall security risks increase by using Risk Increase Threshold (RITH) while improving security without negative impact on energy consumption or Quality of Service (QoS).
EGYPTIAN INFORMATICS JOURNAL
(2021)
Article
Computer Science, Information Systems
Sahul Goyal, Lalit Kumar Awasthi
Summary: The demand for cloud-based computation is increasing exponentially, but it also brings issues such as energy consumption and SLA violations. Therefore, it is crucial to improve resource utilization in cloud data centers. This research proposes an adaptive VM placement approach based on a meta-heuristic algorithm, which combines energy conservation, resource utilization, and required quality of services. Experimental results show that this approach significantly reduces energy consumption and decreases SLA violations and VM migrations.
CLUSTER COMPUTING-THE JOURNAL OF NETWORKS SOFTWARE TOOLS AND APPLICATIONS
(2023)
Article
Multidisciplinary Sciences
Muhammad Ibrahim, Muhammad Imran, Faisal Jamil, Yun-Jung Lee, Do-Hyeun Kim
Summary: The proposed Efficient Adaptive Migration Algorithm (EAMA) effectively reduces the number of migrations and SLA violation, increases resource utilization, and decreases energy consumption compared to the PACPA and RUAEE algorithms.
Article
Computer Science, Artificial Intelligence
Mohamed Ghetas
Summary: The growing demand for cloud computing adoption poses challenges for researchers to make cloud computing more efficient and affordable. Server consolidation is a strategy to improve data center energy efficiency and resource utilization, and saving energy consumption can significantly reduce overall cloud management costs.
NEURAL COMPUTING & APPLICATIONS
(2021)
Article
Computer Science, Hardware & Architecture
Jean Pepe Buanga Mapetu, Lingfu Kong, Zhen Chen
Summary: This paper investigates the trade-off between energy consumption, SLA violations, and VM migration quantity in cloud data centers, proposing a dynamic VM consolidation approach-based load balancing. Through four different methods, the approach efficiently controls relevant metrics to tackle the NP-hard optimization problem in data centers.
JOURNAL OF SUPERCOMPUTING
(2021)
Article
Computer Science, Software Engineering
Ahmadreza Hassannezhad Najjari, Ali Asghar Pourhaji Kazem
Summary: This paper provides a systematic overview of live VM migration and investigates the relevant literature. There is a lack of comprehensive overview on live migration, and this paper fills this gap. Through categorizing and discussing 50 selected research reports, different results are obtained.
CONCURRENCY AND COMPUTATION-PRACTICE & EXPERIENCE
(2022)
Article
Chemistry, Analytical
Pingping Li, Jiuxin Cao
Summary: This study proposes a virtual machine consolidation approach based on dynamic load mean and multi-objective optimization to address the issues of energy consumption and resource utilization in cloud data centers. By considering multiple factors and utilizing optimization algorithms, it achieves reduced energy consumption, optimized resource utilization, and ensured quality of service.
Review
Computer Science, Artificial Intelligence
Arif Ullah, Nazri Mohd Nawi, Soukaina Ouhame
Summary: Cloud computing is a new technology that utilizes virtualization and virtual machines to improve data center performance and efficiency to meet user demands. Improvements in virtual machine parameters directly enhance cloud computing performance.
ARTIFICIAL INTELLIGENCE REVIEW
(2022)
Article
Computer Science, Information Systems
Priyanka Nehra, A. Nagaraju
Summary: This paper proposes a Support Vector Regression-based methodology to predict a host's future utilization using multiple resource's utilization history. Compared to existing approaches, the proposed method performs better in terms of root mean square error, mean absolute percentage error, mean square error, mean absolute error, and R2.
JOURNAL OF KING SAUD UNIVERSITY-COMPUTER AND INFORMATION SCIENCES
(2022)
Article
Computer Science, Information Systems
Satish Kumar, Tao Chen, Rami Bahsoon, Rajkumar Buyya
Summary: In this article, we propose DebtCom, a framework that determines whether to trigger recomposition based on the technical debt metaphor and time-series prediction of workload. Our core idea is that recomposition can be unnecessary if the under-/over-utilization only cause temporarily negative effects, and the current composition plan, although carries debt, can generate greater benefit in the long-term. The results confirm that, in contrast to the state-of-the-art, DebtCom achieves better utility while having lower cost and number of recompositions, rendering each composition plan more sustainable.
IEEE TRANSACTIONS ON SERVICES COMPUTING
(2023)
Article
Multidisciplinary Sciences
Deepika Saxena, Ashutosh Kumar Singh, Chung-Nan Lee, Rajkumar Buyya
Summary: The unsustainable demand and ineffective load management in Cloud Data Centres (CDCs) lead to high energy consumption, resource contention, excessive carbon emission, and security threats. A novel Sustainable and Secure Load Management (SaS-LM) Model is proposed to address these issues by dynamically adjusting the load to maximize security and sustainability. An evolutionary optimization algorithm called Dual-Phase Black Hole Optimization (DPBHO) is used to estimate resource usage and detect congestion. SaS-LM is evaluated using real-world Google Cluster VM traces and shows significant reductions in carbon emission and energy consumption, as well as improved resource utilization.
SCIENTIFIC REPORTS
(2023)
Article
Computer Science, Theory & Methods
Yogesh Sharma, Deval Bhamare, Nishanth Sastry, Bahman Javadi, Rajkumar Buyya
Summary: This article reviews how intent-driven service management systems manage and fulfill SLA requirements and proposes four intent management activities performed in a closed-loop manner. The article categorizes and compares existing SLA management techniques in IDSM systems and suggests future research directions.
ACM COMPUTING SURVEYS
(2023)
Article
Computer Science, Theory & Methods
Samodha Pallewatta, Vassilis Kostakos, Rajkumar Buyya
Summary: This article introduces the utilization of fog computing paradigm and microservice architecture in IoT applications, and their relationship. Efficient placement algorithms are required for microservices-based IoT applications to achieve diverse performance requirements and overcome the challenges introduced by the architecture.
ACM COMPUTING SURVEYS
(2023)
Article
Computer Science, Theory & Methods
Anupama Mampage, Shanika Karunasekera, Rajkumar Buyya
Summary: Serverless computing has gained much attention in recent years as it shifts the burden of resource management to cloud service providers. However, efficiently managing resources while maintaining function performance is challenging due to the dynamic and multi-tenant nature of serverless systems.
FUTURE GENERATION COMPUTER SYSTEMS-THE INTERNATIONAL JOURNAL OF ESCIENCE
(2023)
Article
Computer Science, Theory & Methods
Guangyao Zhou, Wenhong Tian, Rajkumar Buyya
Summary: Cloud computing is a large-scale distributed computing system that dynamically provides elastic services to users. Resource scheduling in Cloud computing, which aims to minimize makespan, is usually NP-Hard. This paper proposes multi-search-routes-based algorithms, integrating LPT and BFD as basic search routes with the OneStep neighborhood search algorithm, to optimize scheduling schemes for homogeneous and heterogeneous resources. Theoretical derivations and extensive experiments demonstrate the superiority of the proposed algorithms in minimizing makespan problems.
FUTURE GENERATION COMPUTER SYSTEMS-THE INTERNATIONAL JOURNAL OF ESCIENCE
(2023)
Article
Computer Science, Hardware & Architecture
Xiaogang Wang, Jian Cao, Rajkumar Buyya
Summary: This paper proposes an adaptive cloud bundle provisioning and multi-workflow scheduling model that dynamically scales resources on multi-type VM instances for the execution of complex workflows. The performance of the proposed algorithms is demonstrated to be superior to that of existing policies.
IEEE TRANSACTIONS ON COMPUTERS
(2023)
Article
Computer Science, Information Systems
Mohammad Goudarzi, Marimuthu Palaniswami, Rajkumar Buyya
Summary: Fog/Edge computing is a new computing paradigm that supports resource-constrained IoT devices by placing their tasks on edge and/or cloud servers. However, existing centralized Deep Reinforcement Learning (DRL)-based placement techniques lack generalizability and quick adaptability. To address this, we propose a distributed application placement technique based on IMPALA, which significantly improves the execution cost of IoT applications. Our technique utilizes recurrent layers to capture temporal behaviors and a replay buffer to enhance sample efficiency.
IEEE TRANSACTIONS ON MOBILE COMPUTING
(2023)
Article
Computer Science, Theory & Methods
Haiying Shen, Haoyu Wang, Jiechao Gao, Rajkumar Buyya
Summary: In this paper, a stable and reliable renewable energy allocation system is proposed to address the issues of increasing electricity cost, energy consumption, and harmful gas emissions in datacenters due to the instability of renewable energy supply. Using deep learning technique, the system predicts the probability of producing the desired amount of each renewable energy source and forecasts the energy demands of each physical machine group. By solving an optimization problem and utilizing reinforcement learning method, renewable energy resources with different instabilities are matched with different physical machine groups for supply. The amount of computing resource allocated to each job is adjusted based on job deadline and failure probability in each physical machine group, and a failure prediction based energy saving method is proposed. Real trace driven experiments demonstrate the effectiveness of the proposed methods in reducing SLO violations, total energy monetary cost, and total carbon emission compared to other methods.
IEEE TRANSACTIONS ON PARALLEL AND DISTRIBUTED SYSTEMS
(2023)
Article
Computer Science, Hardware & Architecture
Anwesha Mukherjee, Shreya Ghosh, Soumya K. Ghosh, Rajkumar Buyya
Summary: This paper proposes a delay-aware and secure service provisioning model for mission-critical applications, using a mobility-aware sensor-fog paradigm based on network coding and steganography. The proposed model outperforms the conventional sensor-cloud framework in terms of delay and power consumption, according to theoretical and simulation results.
JOURNAL OF SUPERCOMPUTING
(2023)
Article
Computer Science, Hardware & Architecture
Dawei Sun, Minghui Wu, Zhihong Yang, Atul Sajjanhar, Rajkumar Buyya
Summary: In this paper, a coordinated load balancing strategy (St-Stream) is proposed to address the problem of skewed data streams. The strategy performs migration pairing for resources at the task allocation stage, cuts and moves out tasks from high-load nodes in a hierarchical manner, and places the moved-out operators in the routing table. A two-tier coordination scheme is also designed to adjust the skewed load within nodes and dynamically restore balance between nodes. Experimental results show that the proposed strategy improves cluster load balance and enhances the performance of the stream processing system.
JOURNAL OF SUPERCOMPUTING
(2023)
Article
Computer Science, Hardware & Architecture
Somnath Bera, Tanushree Dey, Anwesha Mukherjee, Rajkumar Buyya
Summary: This paper proposes an IoT framework based on dew computing, edge computing, and federated learning for crop productivity prediction and recommendation. The framework analyzes soil parameters, environmental parameters, and weather data to predict crop productivity and recommend suitable crops. The proposed framework reduces delay and power consumption compared to both the conventional sensor-cloud framework and the edge-cloud framework. Four machine learning algorithms are compared based on their performance, and each obtains over 95% prediction accuracy. An Android application is also proposed for crop recommendation.
JOURNAL OF SUPERCOMPUTING
(2023)
Article
Computer Science, Information Systems
Prabhakar Krishnan, Kurunandan Jain, Amjad Aldweesh, P. Prabu, Rajkumar Buyya
Summary: This paper introduces a network data plane-based architecture that combines SDN, NFV and ML/AI to improve network management in OpenStack Clouds, ensuring predictability, reliability and security. The framework consists of lightweight monitoring, anomaly-detecting intelligent sensors, ML/AI-based threat analytics engine, and defensive actions deployed as VNFs, enabling high-speed threat detection and rapid response. The simulations and analysis show that this framework substantially secures and outperforms prior OpenStack solutions for Cloud architectures.
JOURNAL OF CLOUD COMPUTING-ADVANCES SYSTEMS AND APPLICATIONS
(2023)
Review
Computer Science, Hardware & Architecture
Shreshth Tuli, Fatemeh Mirhakimi, Samodha Pallewatta, Syed Zawad, Giuliano Casale, Bahman Javadi, Feng Yan, Rajkumar Buyya, Nicholas R. Jennings
Summary: In recent years, there has been a shift in computing paradigms towards decentralized systems like IoT, Edge, Fog, Cloud, and Serverless. This shift has been powered by the adoption of AI-driven autonomous systems for managing distributed computing resources. This survey explores the evolution of data-driven AI methods and their impact on computing systems, focusing on resource management and QoS optimization. It also discusses future research directions and the potential of AI-driven computing systems.
JOURNAL OF NETWORK AND COMPUTER APPLICATIONS
(2023)
Article
Computer Science, Information Systems
Shinu M. Rajagopal, M. Supriya, Rajkumar Buyya
Summary: The fog and edge computing paradigm offer a distributed architecture for smart healthcare systems driven by IoT applications, reducing latency and improving system efficiency.