On Cloud-based Oversubscription

Householder, Rachel A.; Arnold, Scott; Green, Robert C.

doi:10.14445/22315381/ijett-v8p273

Cited by 31 publications

(16 citation statements)

References 16 publications

(36 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The overbooking concept is not new and has been studied and applied in various fields since long ago [25]. Examples of this diversity are bandwidth allocation [20], airline yield management [26] and parallel computer scheduling [21].…”

Section: Overbookingmentioning

confidence: 99%

An Autonomic Approach to Risk-Aware Data Center Overbooking

Tomás

Tordsson

2014

IEEE Trans. Cloud Comput.

View full text Add to dashboard Cite

Abstract-Elasticity is a key characteristic of cloud computing that increases the flexibility for cloud consumers, allowing them to adapt the amount of physical resources associated to their services over time in an on-demand basis. However, elasticity creates problems for cloud providers as it may lead to poor resource utilization, specially in combination with other factors, such as user overestimations and pre-defined VM sizes. Admission control mechanisms are thus needed to increase the number of services accepted, raising the utilization without affecting services performance. This work focuses on implementing an autonomic risk-aware overbooking architecture capable of increasing the resource utilization of cloud data centers by accepting more virtual machines than physical available resources. Fuzzy logic functions are used to estimate the associated risk to each overbooking decision. By using a distributed PID controller approach, the system is capable of self-adapting over time -changing the acceptable level of risk -depending on the current status of the cloud data center. The suggested approach is extensively evaluated using a combination of simulations and experiments executing real cloud applications with real-life available workloads. Our results show a 50% increment at both resource utilization and capacity allocated with acceptable performance degradation and more stable resource utilization over time.

show abstract

Section: Overbookingmentioning

confidence: 99%

An Autonomic Approach to Risk-Aware Data Center Overbooking

Tomás

Tordsson

2014

IEEE Trans. Cloud Comput.

View full text Add to dashboard Cite

show abstract

“…Resource overbooking has been studied as a solution to mitigate this resource utilization problem [2]. However, deciding what is the suitable amount of overbooking, given the current set of applications, is a really difficult task, and is mostly approached by predictions.…”

Section: Background and Related Workmentioning

confidence: 99%

“…Resource overbooking has been proven as a suitable solution to increase resource utilization ratios [2], specially when users overprovision their resource requests and/or datacenters suffer from the Virtual Machine (VM) sprawl phenomenon. Resource overbooking is a well known technique, previously applied e.g., in network bandwidth multiplexing.…”

Section: Introductionmentioning

confidence: 99%

Reducing Noisy-Neighbor Impact with a Fuzzy Affinity-Aware Scheduler

Tomás

Vázquez²,

Tordsson³

et al. 2015

2015 International Conference on Cloud and Autonomic Computing

View full text Add to dashboard Cite

Abstract-Overbooking techniques have been proven efficient to increase overall utilization of cloud datacenters. However, overbooking may also degrade applications performance as (at least) some applications need to share physical resources such as CPU or memory. Consequently, interference may increase among the virtual machines that share resources, the so called noisy neighbors effect. We present an affinity-aware scheduler to reduce the impact of such interference. A fuzzy logic engine accounts for the uncertainty in these environments and estimates which CPU cores are currently more suitable for each incoming application. This helps the scheduler make virtual machine to physical resource mapping decisions, also known as vcpu pinning. An experimental evaluation based on a combination of interactive services and batch applications confirms that our affinity-aware fuzzy scheduler reduces the interference among applications, enabling more predictable performance and consequently safer overbooking.

show abstract

“…As the authors of [60] state, much of the research conducted thus far has focused on managing oversubscription of a single physical machine though this narrow focus is rather limiting. While other projects successfully manage memory overcommitment at a host level, we have not found any previous work that automatically manages oversubscription in an on-premises Cloud.…”

Section: Related Workmentioning

confidence: 99%

“…This situation, at the Cloud infrastructure level, is called oversubscription [60], which is a technique that can lead to an increase in the number of VMs per physical host though it can have an impact on the Quality of Service and probably violate the Service Level Agreement established by the Cloud provider. However, oversubscription can enable Cloud providers to better use the available memory in their physical systems if the appropriate countermeasures are introduced.…”

Section: Introductionmentioning

confidence: 99%

Efficient and elastic management of computing infrastructures

Laguna¹,

de²

View full text Add to dashboard Cite

AcknowledgementsEn primer lugar quiero dar las gracias a mis padres Lola y Ramón, por haberme animado siempre a continuar trabajando. Gracias a ellos he ido avanzando en mi vida y he llegado hasta aquí. Están tan orgullosos, que sólo por oírles hablar de mí, merece la pena seguir haciendo cosas. Muchas gracias papás, espero seguir consiguiendo que os sintáis orgullosos de mí.Quiero dar las gracias a Mar porque siempre está a mi lado y por haberme apoyado tanto para terminar este trabajo. Ha estado todo el tiempo ahí animándome, diciéndome lo poco que me iba a costar terminarlo, y que no me preocupara, porque iba a quedar bien. Muchas gracias mi amor, espero que sigas siempre junto a mí.También tengo que agradecer la labor de Miguel Caballer y Fernando Alvarruíz, porque con ellos empezó esta tesis y porque han aportado mucho trabajo a ella. Y agradecer a Nacho Blanquer y a Germán Moltó, que se hayan ido incorporando al desarrollo del trabajo, y se hayan implicado de tal forma que han acabado siendo mis directores. Muchas gracias a todos, espero poder continuar trabajando con vosotros.Y finalmente, quiero dar las gracias a Vicente Hernández, porqueél es el artífice de que hace años empezase con esto de la investigación y sé que, aunque no me haya podido acompañar hasta aquí, se habría sentido orgulloso al ver que por fin he terminado. Muchas gracias, Vicente.iii AbstractModern data centers integrate a lot of computer and electronic devices. However, some reports state that the mean usage of a typical data center is around 50% of its peak capacity, and the mean usage of each server is between 10% and 50%. A lot of energy is destined to power on computer hardware that most of the time remains idle. Therefore, it would be possible to save energy simply by powering off those parts from the data center that are not actually used, and powering them on again as they are needed.Most data centers have computing clusters that are used for intensive computing, recently evolving towards an on-premises Cloud service model. Despite the use of low consuming components, higher energy savings can be achieved by dynamically adapting the system to the actual workload. The main approach in this case is the usage of energy saving criteria for scheduling the jobs or the virtual machines into the working nodes. The aim is to power off idle servers automatically. But it is necessary to schedule the power management of the servers in order to minimize the impact on the end users and their applications.The objective of this thesis is the elastic and efficient management of cluster infrastructures, with the aim of reducing the costs associated to idle components. This objective is addressed by automating the power management of the working nodes in a computing cluster, and also proactive stimulating the load distribution to achieve idle resources that could be powered off by means of memory overcommitment and live migration of virtual machines. Moreover, this automation is of interest for virtual clusters, as they also suffer from the same ...

show abstract

On Cloud-based Oversubscription

Cited by 31 publications

References 16 publications

An Autonomic Approach to Risk-Aware Data Center Overbooking

An Autonomic Approach to Risk-Aware Data Center Overbooking

Reducing Noisy-Neighbor Impact with a Fuzzy Affinity-Aware Scheduler

Efficient and elastic management of computing infrastructures

Contact Info

Product

Resources

About