A Stable Matching-Based Virtual Machine Allocation Mechanism for Cloud Data Centers

Wang, Jing V.; Fok, Kai-Yin; Cheng, Chi‐Tsun; Tse, Chi K.

doi:10.1109/services.2016.21

Cited by 19 publications

(10 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The reasoning behind the service level agreement violations (SLAV) time per active hosts is based on the observation that if there is an application that is managing the virtual machine migrations and it is busy with a host that has 100% utilization, it will not be able to address other hosts waiting for service provisioning. Therefore, virtual machines are deprived of the desired performance level causing SLA violations [33]. The mathematical definitions and formula are as follow,…”

Section: Simulationmentioning

confidence: 99%

Economic impact of energy saving techniques in cloud server

et al. 2019

View full text Add to dashboard Cite

In recent years, lot of research has been carried in the field of cloud computing and distributed systems to investigate and understand their performance. Economic impact of energy consumption is of major concern for major companies. Cloud Computing companies (Google, Yahoo, Gaikai, ONLIVE, Amazon and eBay) use large data centers which are comprised of virtual computers that are placed globally and require a lot of power cost to maintain. Demand for energy consumption is increasing day by day in IT firms. Therefore, Cloud Computing companies face challenges towards the economic impact in terms of power costs. Energy consumption is dependent upon several factors, e.g., service level agreement, virtual machine selection techniques, optimization policies, workload types etc. We address a solution for the energy saving problem by enabling dynamic voltage and frequency scaling technique for gaming data centers. The dynamic voltage and frequency scaling technique is compared against non-power aware and static threshold detection techniques. This helps service providers to meet the quality of service and quality of experience constraints by meeting service level agreements. The CloudSim platform is used for implementation of the scenario in which game traces are used as a workload for testing the technique. Selection of better techniques can help gaming servers to save energy cost and maintain a better quality of service for users placed globally. The novelty of the work provides an opportunity to investigate which technique behaves better, i.e., dynamic, static or non-power aware. The results demonstrate that less energy is consumed by implementing a dynamic voltage and frequency approach in comparison with static threshold consolidation or non-power aware technique. Therefore, more economical quality of services could be provided to the end users.

show abstract

Section: Simulationmentioning

confidence: 99%

Economic impact of energy saving techniques in cloud server

et al. 2019

View full text Add to dashboard Cite

show abstract

“…Wang et al [21] also address a matching-based VM consolidation mechanism using migration such that active servers can operate close to a desirable utilization threshold. ML is a recently used technique for energy-aware VM allocation in DCs.…”

Section: A Energy-aware Vm Allocationmentioning

confidence: 99%

Integrating Heuristic and Machine-Learning Methods for Efficient Virtual Machine Allocation in Data Centers

Pahlevan

Zapater

et al. 2018

IEEE Trans. Comput.-Aided Des. Integr. Circuits Syst.

View full text Add to dashboard Cite

Abstract-Modern cloud data centers (DCs) need to tackle efficiently the increasing demand for computing resources and address the energy efficiency challenge. Therefore, it is essential to develop resource provisioning policies that are aware of virtual machine (VM) characteristics, such as CPU utilization and data communication, and applicable in dynamic scenarios. Traditional approaches fall short in terms of flexibility and applicability for large-scale DC scenarios. In this paper we propose a heuristicand a machine learning (ML)-based VM allocation method and compare them in terms of energy, quality of service (QoS), network traffic, migrations, and scalability for various DC scenarios. Then, we present a novel hyper-heuristic algorithm that exploits the benefits of both methods by dynamically finding the best algorithm, according to a user-defined metric. For optimality assessment, we formulate an integer linear programming (ILP)-based VM allocation method to minimize energy consumption and data communication, which obtains optimal results, but is impractical at runtime. Our results demonstrate that the ML approach provides up to 24% server-to-server network traffic improvement and reduces execution time by up to 480x compared to conventional approaches, for large-scale scenarios. On the contrary, the heuristic outperforms the ML method in terms of energy and network traffic for reduced scenarios. We also show that the heuristic and ML approaches have up to 6% energy consumption overhead compared to ILP-based optimal solution. Our hyper-heuristic integrates the strengths of both the heuristic and the ML methods by selecting the best one during runtime.

show abstract

“…The data set we used in the simulations is obtained from real-world workload traces of PlanetLab [11]. We choose 10 days from the dataset as in [7] and average out the simulation results for comparisons. Two kinds of servers were emulated in the simulated data center: HP ProLiant G4 (2 cores × 1860 MIPS), and HP ProLiant G5 (2 cores × 2660 MIPS).…”

Section: A Simulation Setupmentioning

confidence: 99%

“…In an earlier work of the authors in this paper [6], host's temperature was considered as a migration criterion, which provides a better option in reflecting the utilization levels of hosts. In [7], the VM provisioning process was formulated as a stable matching problem that allows VMs and hosts to have different objectives in the allocation process.…”

Section: Introductionmentioning

confidence: 99%

A heuristics-based VM allocation mechanism for cloud data centers

Wang

Ganganath

Cheng

et al. 2017

2017 IEEE International Symposium on Circuits and Systems (ISCAS)

Self Cite

View full text Add to dashboard Cite

Abstract-Live migration, which allows virtual machines (VMs) to move across distinct physical hosts, is widely adopted for realizing energy saving and load balancing in modern Cloud data centers. However, putting VMs with high correlations on their CPU utilization patterns onto the same host is very likely to trigger overloading incidents and commit Service Level Agreements (SLA) violations, even when the host has not yet reached its critical limits. To avoid such SLA violations and to maintain a low energy consumption, this work proposes a heuristics-based VM allocation mechanism. Under the proposed mechanism, VMs showing high CPU utilization correlations to other co-located peers and hosts with extreme utilization values are both assigned with high heuristic values. VMs with high heuristic values are therefore less preferred to be co-located onto a host that has a high heuristic value. Performance of the proposed mechanism was evaluated using CloudSim with real-world workload data. Simulation results show that when comparing with other existing mechanisms, the proposed idea can further reduce both energy consumption and SLA violations.

show abstract

A Stable Matching-Based Virtual Machine Allocation Mechanism for Cloud Data Centers

Cited by 19 publications

References 7 publications

Economic impact of energy saving techniques in cloud server

Economic impact of energy saving techniques in cloud server

Integrating Heuristic and Machine-Learning Methods for Efficient Virtual Machine Allocation in Data Centers

A heuristics-based VM allocation mechanism for cloud data centers

Contact Info

Product

Resources

About