Learn-as-you-go with Megh: Efficient Live Migration of Virtual Machines

Basu, Debabrota; Wang, Xiayang; Hong, Yang; Chen, Haibo; Bressan, Stéphane

doi:10.1109/tpds.2019.2893648

Cited by 66 publications

(40 citation statements)

References 37 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…, h m }. We assume the paradigm of discrete time control, where we take scheduling and migration decisions at periodic intervals [2], [30], [31]. We assume a scheduler is already present in the broker.…”

Section: Methodology a System Model And Problem Formulationmentioning

confidence: 99%

PreGAN: Preemptive Migration Prediction Network for Proactive Fault-Tolerant Edge Computing

Tuli¹,

Casale²,

Jennings³

2021

Preprint

View full text Add to dashboard Cite

Building a fault-tolerant edge system that can quickly react to node overloads or failures is challenging due to the unreliability of edge devices and the strict service deadlines of modern applications. Moreover, unnecessary task migrations can stress the system network, giving rise to the need for a smart and parsimonious failure recovery scheme. Prior approaches often fail to adapt to highly volatile workloads or accurately detect and diagnose faults for optimal remediation. There is thus a need for a robust and proactive fault-tolerance mechanism to meet service level objectives. In this work, we propose PreGAN, a composite AI model using a Generative Adversarial Network (GAN) to predict preemptive migration decisions for proactive fault-tolerance in containerized edge deployments. PreGAN uses co-simulations in tandem with a GAN to learn a few-shot anomaly classifier and proactively predict migration decisions for reliable computing. Extensive experiments on a Raspberry-Pi based edge environment show that PreGAN can outperform state-of-the-art baseline methods in fault-detection, diagnosis and classification, thus achieving high quality of service. PreGAN accomplishes this by 5.1% more accurate fault detection, higher diagnosis scores and 23.8% lower overheads compared to the best method among the considered baselines.

show abstract

Section: Methodology a System Model And Problem Formulationmentioning

confidence: 99%

PreGAN: Preemptive Migration Prediction Network for Proactive Fault-Tolerant Edge Computing

Tuli¹,

Casale²,

Jennings³

2021

Preprint

View full text Add to dashboard Cite

show abstract

“…Relevant scholars proposed a soft decision tree algorithm (soft decision tree), which defines a complete set of tree building and pruning processes and improves the applicability of decision trees through subassembly and reorganization [14]. Relevant scholars proposed the C-fuzzy decision tree (C-fuzzy decision tree) algorithm based on the fuzzy clustering algorithm [15]. e algorithm can consider multiple attributes at the same time when building a tree.…”

Section: Related Workmentioning

confidence: 99%

System Optimization of Talent Life Cycle Management Platform Based on Decision Tree Model

Han¹

2022

Journal of Mathematics

View full text Add to dashboard Cite

Decision tree algorithm is a widely used classification and prediction method. Because it generates a tree-like classifier, it has a simple structure and is extensively used by people. Regardless of the decision tree algorithm, the decision attributes are classified according to the condition attributes. The judgment process is from the root node to the leaf node. Each branch of the tree takes the form of selecting the best split attribute. However, this classification method of decision tree makes it rely too much on training data. If the data are more complicated, there are noisy data, incomplete data, etc. The decision tree will often have overfitting problems. This study mainly analyzes the random forest algorithm model and the CART algorithm and applies the CART algorithm to the model according to the random forest model. Aiming at the algorithm’s shortcomings in solving big data, this study will improve the algorithm through the MapReduce programming model to achieve parallelization of the process and construction of the function. Combining the construction goals and principles of the talent supply chain management system, this study constructs the overall framework and operational process of the enterprise talent supply chain management system based on the decision tree model from the overall level and the operational level. Aiming at the enterprise’s talent management problems, it focuses on designing integrated management, flexible management, talent information integrated management, and evaluation and optimization management models to ensure that the constructed system is operable and measurable and can achieve dynamic optimization. Based on the current situation of talent management in a company, this study analyzes the enterprise talent supply chain management model based on the decision tree model proposed in this study and constructs the overall framework and core model of a company’s talent supply chain management system. The current situation of the company puts forward the safeguard measures for the implementation of the management system to assure that the established management system can be effectively implemented.

show abstract

“…However, their algorithm faces the curse of dimensionality due to the infinite combination of state space. To tackle this problem, Basu et al [11] used the Deep Q-learning Network (DQN) and proposed Megh to optimize energy efficiency. In their model, a state is a mapping between PMs and VMs with certain workloads.…”

Section: Related Workmentioning

confidence: 99%

“…Therefore, it is necessary to strike a trade-off between energy consumption and SLA for a good VM consolidation strategy. Considering this compromise, a number of studies [4][5][6][7][8][9][10][11] have been conducted to design an effective VM consolidation from three aspects: minimizing energy consumption, avoiding SLA violation, and reducing VM migrations. However, most of these studies are based on simulation experiments rather than verification in real data centers, and their results show that the performance of different methods significantly varies when applied on systems with different characteristics.…”

Section: Introductionmentioning

confidence: 99%

I-Neat: An intelligent framework for adaptive virtual machine consolidation

Liu

Zhao

Dong

et al. 2022

Tsinghua Sci. Technol.

View full text Add to dashboard Cite

With the increasing use of cloud computing, high energy consumption has become one of the major challenges in cloud data centers. Virtual Machine (VM) consolidation has been proven to be an efficient way to optimize energy consumption in data centers, and many research works have proposed to optimize VM consolidation.However, the performance of different algorithms is related with the characteristics of the workload and system status; some algorithms are suitable for Central Processing Unit (CPU)-intensive workload and some for web application workload. Therefore, an adaptive VM consolidation framework is necessary to fully explore the potential of these algorithms. Neat is an open-source dynamic VM consolidation framework, which is well integrated into OpenStack.However, it cannot conduct dynamic algorithm scheduling, and VM consolidation algorithms in Neat are few and basic, which results in low performance for energy saving and Service-Level Agreement (SLA) avoidance. 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 scheduler can select appropriate algorithms for the local manager from an algorithm library with many load detection algorithms. The algorithm library is designed based on a template, and in addition to the algorithms of Neat, I-Neat adds six new algorithms to the algorithm library. Furthermore, the framework manager helps users add self-defined algorithms to I-Neat without modifying the source code. Our experimental results indicate that the intelligent scheduler and these novel algorithms can effectively reduce energy consumption with SLA assurance.

show abstract

Learn-as-you-go with Megh: Efficient Live Migration of Virtual Machines

Cited by 66 publications

References 37 publications

PreGAN: Preemptive Migration Prediction Network for Proactive Fault-Tolerant Edge Computing

PreGAN: Preemptive Migration Prediction Network for Proactive Fault-Tolerant Edge Computing

System Optimization of Talent Life Cycle Management Platform Based on Decision Tree Model

I-Neat: An intelligent framework for adaptive virtual machine consolidation

Contact Info

Product

Resources

About