On the Use of Fuzzy Modeling in Virtualized Data Center Management

Xu, Jing; Zhao, Ming; Fortes, José A. B.; Carpenter, Robert J.; Yousif, Mazin

doi:10.1109/icac.2007.28

Cited by 127 publications

(61 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The notion of fuzzy inference for resource modeling appears in [18] in a different setting. The work is aimed at realizing a two-level resource management system with autonomic resource allocation.…”

Section: Related Workmentioning

confidence: 99%

A Novel Approach for Performance Characterization of IaaS Clouds

Ács

Németh

Gergely

2014

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Abstract. Infrastructure-as-a-Service (IaaS) clouds are widely used today, however there are no standardized or commonly used performance evaluation methods and metrics that can be used to compare the services of the different providers. Performance evaluation tools and benchmarks are able to grasp some aspects or details of performance but for various reasons are not capable to characterize cloud performance. Our aim is to collect these elementary or primitive facets of performance and derive high-level aggregated and qualitative performance characterization semantically far above the output of tools and benchmarks. We designed and implemented a framework that collects low-level raw performance data (in terms of CPU, disk, memory and network) of cloud providers based on standard benchmark tools and these data are aggregated and evaluated using a hierarchical fuzzy system. In this process performance characteristics are associated with symbolic values and fuzzy inference is applied to produce the normalized qualitative comparable and readable performance metrics. In this paper, we discuss the issues of cloud performance analysis, present the concept and implementation of our method, illustrate the proposed solution by comparing -in terms of performancethe general purpose medium instance type of the Amazon EC2 cloud (in Ireland) and the standard instance type of the OpenNebula installation at MTA SZTAKI.

show abstract

Section: Related Workmentioning

confidence: 99%

A Novel Approach for Performance Characterization of IaaS Clouds

Ács

Németh

Gergely

2014

Lecture Notes in Computer Science

View full text Add to dashboard Cite

show abstract

“…The work proposed in [23] designs different SISO (Single-input single-output) and MIMO controllers to determine the CPU allocation of VMs, relying on Kalman filters. [32] and [33] apply an adaptive fuzzy controller to the application, and estimate the required CPU load for the input workload. The problem of the control theory based approaches is that they are difficult to set up and are essentially reactive, which can cause problems during the allocation of resources, since these take minutes until they are ready for use.…”

Section: Related Workmentioning

confidence: 99%

Combining time series prediction models using genetic algorithm to autoscaling Web applications hosted in the cloud infrastructure

Messias¹,

Estrella²,

Ehlers³

et al. 2015

Neural Comput & Applic

View full text Add to dashboard Cite

In a cloud computing environment, companies have the ability to allocate resources according to demand. However, there is a delay that may take minutes between the request for a new resource and it is ready for using. This causes the reactive techniques, which request a new resource only when the system reaches a certain load threshold, are not suitable for the resource allocation process. To address this problem, it is necessary to predict requests that arrive at the system in the next period of time to allocate the necessary resources, before the system becomes overloaded. There are several time-series forecasting models to calculate the workload predictions based on history of monitoring data. However, it is difficult to know which is the best time series forecasting model to be used in each case. The work becomes even more complicated when the user does not have much historical data to be analyzed. Most related work, considers only single methods to evaluate the results of the forecast. Other work propose an approach that selects suitable forecasting methods for a given context. But in this case, it is necessary to have a significant amount of data to train the classifier. Moreover, the best solution may not be a specific model, but rather a combination of models. In this paper we propose an adaptive prediction method using genetic algorithms to combine time-series forecasting models. Our method does not require a previous phase of training, because it constantly adapts the extent to which the data is coming. To evaluate our proposal we use three logs extracted from real web servers. The results show that our proposal often brings the best result, and is generic enough to adapt to various types of time series.

show abstract

“…In [16], the authors propose a two-level optimization scheme to allocate CPU shares to VMs processing two enterprise applications on a single host. Controllers, local to each VM, use fuzzylogic models to estimate CPU shares for the current workload intensity, and make CPU-share requests to a global controller.…”

Section: Related Workmentioning

confidence: 99%

Power and Performance Management of Virtualized Computing Environments Via Lookahead Control

Kusic

Kephart

Hanson³

et al. 2008

2008 International Conference on Autonomic Computing

228

View full text Add to dashboard Cite

There is growing incentive to reduce the power consumed by large-scale data centers that host online services such as banking, retail commerce, and gaming. Virtualization is a promising approach to consolidating multiple online services onto a smaller number of computing resources. A virtualized server environment allows computing resources to be shared among multiple performance-isolated platforms called virtual machines. By dynamically provisioning virtual machines, consolidating the workload, and turning servers on and off as needed, data center operators can maintain the desired quality-of-service (QoS) while achieving higher server utilization and energy efficiency. We implement and validate a dynamic resource provisioning framework for virtualized server environments wherein the provisioning problem is posed as one of sequential optimization under uncertainty and solved using a lookahead control scheme. The proposed approach accounts for the switching costs incurred while provisioning virtual machines and explicitly encodes the corresponding risk in the optimization problem. Experiments using the Trade6 enterprise application show that a server cluster managed by the controller conserves, on average, 22% of the power required by a system without dynamic control while still maintaining QoS goals. Finally, we use trace-based simulations to analyze controller performance on server clusters larger than our testbed, and show how concepts from approximation theory can be used to further reduce the computational burden of controlling large systems.

show abstract

On the Use of Fuzzy Modeling in Virtualized Data Center Management

Abstract: One of the most important goals of data-center management is to

Cited by 127 publications

References 14 publications

A Novel Approach for Performance Characterization of IaaS Clouds

A Novel Approach for Performance Characterization of IaaS Clouds

Combining time series prediction models using genetic algorithm to autoscaling Web applications hosted in the cloud infrastructure

Power and Performance Management of Virtualized Computing Environments Via Lookahead Control

Contact Info

Product

Resources

About