Performance Modelling and Verification of Cloud-Based Auto-Scaling Policies

Evangelidis, Alexandros; Parker, David; Bahsoon, Rami

doi:10.1109/ccgrid.2017.39

Cited by 11 publications

(3 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Evangledis et al [9] addressed performance modeling and formal verification of auto-scaling policies in PaaS and IaaS to provide performance guarantees to reduce SLAs violations, where two cloud services providers Amazon EC2 and Azure have been considered. The authors considered rule-based auto-scaling policies, where upper and/or lower bound on performance metrics such as CPU are expressed.…”

Section: Related Workmentioning

confidence: 99%

Modeling and Performance Analysis of Resource Provisioning in Cloud Computing using Probabilistic Model Checking

Debbi

2021

IJCAI

View full text Add to dashboard Cite

Cloud computing consists of an advanced set of technologies that allow cloud providers to offer computing resources such as infrastructure, platforms and applications to be accessible over the Internet as services. Cloud computing relies on virtualization of resources in the cloud data centers, where a set of Virtual Machines (VMs) are deployed on Physical Machines (PMs) to provision and serve user requests. Due to the dynamic nature of cloud environments and complexity of resources virtualization, as well as the diversity of user's requests, developing effective techniques to evaluate and analyze the performance of cloud centers has become highly required. In this paper, we propose the use of probabilistic model checking as an effective framework for the evaluation and the performance analysis of resource provisioning in the cloud. Based on an analytical model for resource provisioning in Infrastructure-as-a-Service (IaaS) cloud, we build a stochastic model using the probabilistic model checker PRISM and analyze it against a useful set of probabilistic and reward properties that help to measure and analyze cloud performance in an efficient way.Povzetek: Analizirane so razne komponente računanja v oblaku, npr. modeliranje in performance virov.

show abstract

Section: Related Workmentioning

confidence: 99%

Modeling and Performance Analysis of Resource Provisioning in Cloud Computing using Probabilistic Model Checking

Debbi

2021

IJCAI

View full text Add to dashboard Cite

show abstract

“…Evangelidis et al propose a probabilistic verification scheme aimed at dynamically evaluating auto-scaling policies of IaaS and PaaS virtual machines in Amazon EC2 and Microsoft Azure [59]. For that, it applies a Markov model implemented in the PRISM model checker [60].…”

Section: Verification Of Cloud Administration Concernsmentioning

confidence: 99%

A component-based framework for certification of components in a cloud of HPC services

Dantas

Carvalho

Barbosa

2020

Science of Computer Programming

View full text Add to dashboard Cite

HPC Shelf is a proposal of a cloud computing platform to provide component-oriented services for High Performance Computing (HPC) applications. This paper presents a Verification-as-a-Service (VaaS) framework for component certification on HPC Shelf. Certification is aimed at providing higher confidence that components of parallel computing systems of HPC Shelf behave as expected according to one or more requirements expressed in their contracts. To this end, new abstractions are introduced, starting with certifier components. They are designed to inspect other components and verify them for different types of functional, non-functional and behavioral requirements. The certification framework is naturally based on parallel computing techniques to speed up verification tasks.

show abstract

“…Also observe that this policy will converge to a stable deployment when the ratio of tasks per instance is between T /2 and T , so in the end it will produce a result similar to a tracking policy, though expectedly much more slowly. We therefore propose to evaluate a more aggressive set of rules, P aggr , in which the number of new resources is in proportion both to already running instances and distance to the target (some examples of this approach can also be found in the literature [24,16]). Finally, it should be noted that rules #11 and #12 are present in all the policies to keep the number of instances bounded (in general, lower bounds are chosen to satisfy a minimum desired throughput, while upper bounds are given by cost).…”

Section: Scalability Designmentioning

confidence: 99%

A self-scalable distributed network simulation environment based on cloud computing

et al. 2018

View full text Add to dashboard Cite

While parameter sweep simulations can help undergraduate students and researchers to understand computer networks, their usage in the academia is hindered by the significant computational load they convey. This paper proposes DNSE3, a service oriented computer network simulator that, deployed in a cloud computing infrastructure, leverages its elasticity and pay-per-use features to compute parameter sweeps. The performance and cost of using this application is evaluated in several experiments applying different scalability policies, with results that meet the demands of users in educational institutions. Additionally, the usability of the application has been measured following industry standards with real students, yielding a very satisfactory user experience.

show abstract

Performance Modelling and Verification of Cloud-Based Auto-Scaling Policies

Cited by 11 publications

References 18 publications

Modeling and Performance Analysis of Resource Provisioning in Cloud Computing using Probabilistic Model Checking

Modeling and Performance Analysis of Resource Provisioning in Cloud Computing using Probabilistic Model Checking

A component-based framework for certification of components in a cloud of HPC services

A self-scalable distributed network simulation environment based on cloud computing

Contact Info

Product

Resources

About