On the Use of a Proportional-Share Market for Application SLO Support in Clouds

Costache, Stefania; Parlavantzas, Nikos; Morin, Christine; Kortas, Samuel

doi:10.1007/978-3-642-40047-6_35

Cited by 5 publications

(8 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…It is noteworthy that with large and complicated systems (i.e., composed of different physical hosts and many OSes), this initial benchmark should be performed one each kind of machines. Figures 10 and 11 show the comparison of boot time, with HDD and SSD machines under resources contentions, between the Grid5000 experiments, our model and a naive one (i.e., with a constant boot time of 30 seconds as used in [4]). We evaluated the model on three resource contention cases: CPU, mixed workload (using Pgbench) and I/O contention.…”

Section: Model Evaluationmentioning

confidence: 99%

“…However, one just has to give a look to the literature to understand that an accurate model is required. In Costache et al [4], the authors used a simple VM boot time model based on a small constant. More recently, a survey on cloud simulation tools [5] showed that current models of VM boot time are simply ignored because they assumed that the VM boot time duration is negligible.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Virtual Machine Boot Time Model

Nguyễn¹,

Lèbre²

2017

2017 25th Euromicro International Conference on Parallel, Distributed and Network-Based Processing (PDP)

View full text Add to dashboard Cite

International audience—Cloud computing by far brings a lot of undeniable advantages. Accordingly, many research works aim to evaluate the characteristics of cloud systems on many aspects such as performance, workload, cost, provisioning policies, and resources management. In order to setup a cloud system for running rigorous experiments, ones have to overcome a huge amount of challenges and obstacles to build, deploy, and manage systems and applications. Cloud simulation tools help researchers to focus only on the parts they are interested about without facing the aforementioned challenges. However, cloud simulators still do not provide accurate models for Virtual Machine (VM) operations. This leads to incorrect results in evaluating real cloud systems. Following previous works on live-migration, we present in this paper an experimental study we conducted in order to propose a first-class VM boot time model. Most cloud simulators often ignore the VM boot time or give a naive model to represent it. After studying the relationship between the VM boot time and different system parameters such as CPU utilization, memory usage, I/O and network bandwidth, we introduce a first boot time model that could be integrated into current cloud simulators. Through experiments, we also confirmed that our model correctly reproduced the boot time of a VM under different resources contention

show abstract

Section: Model Evaluationmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Virtual Machine Boot Time Model

Nguyễn¹,

Lèbre²

2017

2017 25th Euromicro International Conference on Parallel, Distributed and Network-Based Processing (PDP)

View full text Add to dashboard Cite

show abstract

“…Commercial systems such as the Amazon EC2 Spot Instances have adopted this strategy. S. Costache et al [35] proposes a market where the users bid for a VM with a certain amount of resources. To guarantee a steady amount of resources, their system migrates VMs between different nodes which has the potential to impose a significant performance penalty [36].…”

Section: Related Workmentioning

confidence: 99%

Partial Utility-Driven Scheduling for Flexible SLA and Pricing Arbitration in Clouds

Simão

Veiga

2016

IEEE Trans. Cloud Comput.

View full text Add to dashboard Cite

Cloud SLAs compensate customers with credits when average availability drops below certain levels. This is too inflexible because consumers lose non-measurable amounts of performance being only compensated later, in next charging cycles. We propose to schedule virtual machines (VMs), driven by range-based non-linear reductions of utility, different for classes of users and across different ranges of resource allocations: partial utility. This customer-defined metric, allows providers transferring resources between VMs in meaningful and economically efficient ways. We define a comprehensive cost model incorporating partial utility given by clients to a certain level of degradation, when VMs are allocated in overcommitted environments (Public, Private, Community Clouds). CloudSim was extended to support our scheduling model. Several simulation scenarios with synthetic and real workloads are presented, using datacenters with different dimensions regarding the number of servers and computational capacity. We show the partial utility-driven driven scheduling allows more VMs to be allocated. It brings benefits to providers, regarding revenue and resource utilization, allowing for more revenue per resource allocated and scaling well with the size of datacenters when comparing with an utility-oblivious redistribution of resources. Regarding clients, their workloads' execution time is also improved, by incorporating an SLA-based redistribution of their VM's computational power.

show abstract

“…Merkat is a market‐based PaaS system for private clouds, which uses an economic model to share resources among multiple applications. Similarly to our work, Merkat aims at maximizing the total value that applications obtain from a shared infrastructure.…”

Section: Related Workmentioning

confidence: 99%

“…To summarize, some of the reviewed systems have similar goals but do not employ the resource and execution management mechanisms proposed by our work . Other systems employ similar mechanisms or have flexible architectures, potentially enabling them to integrate such mechanisms .…”

Section: Related Workmentioning

confidence: 99%

Cloud resource management driven by profit augmentation

Freitas

Parlavantzas

Pazat

2016

Concurrency and Computation

Self Cite

View full text Add to dashboard Cite

International audienceCloud computing has become the infrastructure of choice for building and delivering software services. A key challenge for service providers is effectively managing cloud resources in order to increase profit while maintaining service-level agreements (SLAs) with customers. To address this challenge, this paper proposes a combination of automated mechanisms for resource and execution management. The resource management mechanisms, namely, under-provisioning and contract rescission, reduce resource allocation costs and minimize penalties incurred when performance objectives are violated. The execution management mechanisms, namely, crash recovery and delay recovery, minimize penalties incurred when reliability objectives are violated. The mechanisms are integrated in the Qu4DS framework and evaluated in the Grid'5000 testbed. The results show that the under-provisioning mechanism increases profit by 20–50%, the contract rescission mechanism increases profit up to four times, and the execution management mechanisms increase profit by up to 60%

show abstract

On the Use of a Proportional-Share Market for Application SLO Support in Clouds

Cited by 5 publications

References 15 publications

Virtual Machine Boot Time Model

Virtual Machine Boot Time Model

Partial Utility-Driven Scheduling for Flexible SLA and Pricing Arbitration in Clouds

Cloud resource management driven by profit augmentation

Contact Info

Product

Resources

About