Reserved or On-Demand Instances? A Revenue Maximization Model for Cloud Providers

Mazzucco, Michele; Dumas, Marlon

doi:10.1109/cloud.2011.25

Cited by 27 publications

(22 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…On the other hand, in multi-server queueing systems with limited waiting room, the distribution of the response times is not exponential, in spite of making Markovian assumptions [3]. However, we can exploit the fact that the observed average response time over an interval, which according to Equation (2) involves the sum of (tγ) response times, can be treated as being approximately normally distributed with mean W and variance V AR(W )/(tγ), with V AR(W ) being the variance of the response times. That approximation appeals to the central limit theorem and ignores the dependencies between individual response times.…”

Section: Policiesmentioning

confidence: 96%

See 1 more Smart Citation

Achieving Performance and Availability Guarantees with Spot Instances

Mazzucco

Dumas

2011

2011 IEEE International Conference on High Performance Computing and Communications

Self Cite

View full text Add to dashboard Cite

In the Infrastructure-as-a-Service (IaaS) cloud computing market, spot instances refer to virtual servers that are rented via an auction. Spot instances allow IaaS providers to sell spare capacity while enabling IaaS users to acquire virtual servers at a lower price than the regular market price (also called "on demand" price). Users bid for spot instances at their chosen limit price. Based on the bids and the available capacity, the IaaS provider sets a clearing price. A bidder acquires their requested spot instances if their bid is above the clearing price. However, these spot instances may be terminated by the IaaS provider impromptu if the auction's clearing price goes above the user's limit price. In this context, this paper addresses the following question: Can spot instances be used to run paid web services while achieving performance and availability guarantees? The paper examines the problem faced by a Software-asa-Service (SaaS) provider who rents spot instances from an IaaS provider and uses them to provide a web service on behalf of a paying customer. The SaaS provider incurs a monetary cost for renting computing resources from the IaaS provider, while charging its customer for executing web service transactions and paying penalties to the customer for failing to meet performance and availability objectives. To address this problem, the paper proposes a bidding scheme and server allocation policies designed to optimize the average revenue earned by the SaaS provider per time unit. Experimental results show that the proposed approach delivers higher revenue to the SaaS provider than an alternative approach where the SaaS provider runs the web service using "on demand" instances. The paper also shows that the server allocation policies seamlessly adapt to varying market conditions, traffic conditions, penalty levels and transaction fees.

show abstract

Section: Policiesmentioning

confidence: 96%

“…This option is attractive when utilization can be planned in advance. [1] and [2] have addressed the question of whether to acquire "on demand" or "reserved" instances.…”

Section: Introductionmentioning

confidence: 99%

Achieving Performance and Availability Guarantees with Spot Instances

Mazzucco

Dumas

2011

2011 IEEE International Conference on High Performance Computing and Communications

Self Cite

View full text Add to dashboard Cite

show abstract

“…The specified value of B, referred to as B spec , which is provided by the user, is the desired value of B maintained by the broker and is stored in the GoS component. Similar GoS criteria have been used by other researchers [30]. A previous work [31] by the authors also dealt with auto-scaling resources in a cloud environment while monitoring the value of B.…”

Section: System Overviewmentioning

confidence: 99%

A hybrid auto-scaling technique for clouds processing applications with service level agreements

et al. 2017

View full text Add to dashboard Cite

This research focuses on the automatic provisioning of cloud resources by an intermediary enterprise. This enterprise provides a virtual private cloud for a single client enterprise by using resources from a public cloud. The intermediary cloud provider is controlled by a broker that uses techniques to dynamically control the number of resources used by the client enterprise. The research presents a hybrid auto-scaling technique based on a combination of a reactive approach and a proactive approach to scale resources based on user demand. The primary goal of this auto-scaling technique is to achieve a profit for the intermediary enterprise while maintaining a desired grade of service for the client enterprise. The second goal of the technique is to reduce costs for the single client enterprise. The technique supports both on-demand requests and requests with service level agreements (SLAs). This paper describes the auto-scaling algorithms associated with the hybrid technique and includes a discussion of system design and implementation experience for a prototype system. A detailed performance analysis based on simulations and measurements made based on the prototype is presented.

show abstract

“…Mazucco et al [10] investigate the choice between on-demand and reserved instances from a provider's viewpoint, as opposed to the consumer's viewpoint that receives our focus. Optimising this choice is done by maximising a revenue function by optimal allocation of resources to both contract types.…”

Section: Related Workmentioning

confidence: 99%