Cloud Pricing: The Spot Market Strikes Back

Dierks, Ludwig; Seuken, Sven

doi:10.2139/ssrn.3383420

Cited by 11 publications

(9 citation statements)

References 32 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Both on-demand and reservation models are available with the latter typically offering a lower price. Several works [1], [5], [8], [31] have studied the pricing strategies for platform providers, as well as delay modeling and cost evaluation for the users. b) Stochastic scheduling and checkpointing: Many prior works have considered stochastic scheduling for jobs with execution time uncertainty.…”

Section: Related Workmentioning

confidence: 99%

Reservation and Checkpointing Strategies for Stochastic Jobs

Gainaru

Goglin

Honoré

et al. 2020

2020 IEEE International Parallel and Distributed Processing Symposium (IPDPS)

View full text Add to dashboard Cite

In this paper, we are interested in scheduling and checkpointing stochastic jobs on a reservation-based platform, whose cost depends both (i) on the reservation made, and (ii) on the actual execution time of the job. Stochastic jobs are jobs whose execution time cannot be determined easily. They arise from the heterogeneous, dynamic and data-intensive requirements of new emerging fields such as neuroscience. In this study, we assume that jobs can be interrupted at any time to take a checkpoint, and that job execution times follow a known probability distribution. Based on past experience, the user has to determine a sequence of fixed-length reservation requests, and to decide whether the state of the execution should be checkpointed at the end of each request. The objective is to minimize the expected cost of a successful execution of the jobs. We provide an optimal strategy for discrete probability distributions of job execution times, and we design fully polynomial-time approximation strategies for continuous distributions with bounded support. These strategies are then experimentally evaluated and compared to standard approaches such as periodic-length reservations and simple checkpointing strategies (either checkpoint all reservations, or none). The impact of an imprecise knowledge of checkpoint and restart costs is also assessed experimentally.

show abstract

Section: Related Workmentioning

confidence: 99%

Reservation and Checkpointing Strategies for Stochastic Jobs

Gainaru

Goglin

Honoré

et al. 2020

2020 IEEE International Parallel and Distributed Processing Symposium (IPDPS)

View full text Add to dashboard Cite

show abstract

“…There is a large literature on market design challenges in the context of the cloud [Kash and Key 2016]. Existing work has studied both queueing models where decisions are made online with no consideration of the future [Abhishek et al 2012;Dierks and Seuken 2019] and reservation models which assume very strong information about the future [Azar et al 2015;Babaioff et al 2017]. Our work sits in an interesting intermediate position where users may have rough information about the types of their deployments.…”

Section: Related Workmentioning

confidence: 99%

“…All of a deployment's processes are memoryless (i.e., they only depend on the current state). This is common in the literature whenever arrival and departure processes are modelled (e.g., in queuing theory), and has been used in previous models of cloud computing [Abhishek et al 2012;Dierks and Seuken 2019]. Memorylessness is reasonable at cloud scale, but it is not essential for our approach and policies.…”

Section: Formal Modelmentioning

confidence: 99%

On the cluster admission problem for cloud computing

Dierks¹,

Kash

Seuken³

2019

Proceedings of the 14th Workshop on the Economics of Networks, Systems and Computation

Self Cite

View full text Add to dashboard Cite

Cloud computing providers face the problem of matching heterogeneous customer workloads to resources that will serve them. This is particularly challenging if customers, who are already running a job on a cluster, scale their resource usage up and down over time.The provider therefore has to continuously decide whether she can add additional workloads to a given cluster or if doing so would impact existing workloads' ability to scale. Currently, this is often done using simple threshold policies to reserve large parts of each cluster, which leads to low utilization. We propose more sophisticated policies for controlling admission to a cluster and demonstrate that they significantly increase cluster utilization. We first introduce the cluster admission problem and formalize it as a constrained Partially Observable Markov Decision Process (POMDP). As it is infeasible to solve the POMDP optimally, we then systematically design admission policies that estimate moments of each workload's distribution of future resource usage. Via extensive simulations grounded in a trace from Microsoft Azure, we show that our admission policies lead to a substantial improvement over the simple threshold policy. We then show that substantial further gains are possible if high-quality information is available about arriving workloads. Based on this, we propose an information elicitation approach to incentivize users to provide this information and simulate its effects.

show abstract

“…Other work in this space has dealt with comparing pricing mechanisms such as the on-demand market and the spot market [Abhishek et al, 2012, Dierks and Seuken, 2016, Hoy et al, 2016, achieving fairness in job allocation [Friedman et al, 2014], and studying models of real-time pricing with budget constraints [Friedman et al, 2015]. gave a survey of the current state of research in economics and computer science with respect to cloud pricing.…”

Section: Related Workmentioning

confidence: 99%

Simple pricing schemes for the cloud

Kash

Key

Suksompong

2017

Proceedings of the 12th Workshop on the Economics of Networks, Systems and Computation

View full text Add to dashboard Cite

e problem of pricing the cloud has a racted much recent a ention due to the widespread use of cloud computing and cloud services. From a theoretical perspective, several mechanisms that provide strong e ciency or fairness guarantees and desirable incentive properties have been designed. However, these mechanisms o en rely on a rigid model, with several parameters needing to be precisely known in order for the guarantees to hold. In this paper, we consider a stochastic model and show that it is possible to obtain good welfare and revenue guarantees with simple mechanisms that do not make use of the information on some of these parameters. In particular, we prove that a mechanism that sets the same price per time step for jobs of any length achieves at least 50% of the welfare and revenue obtained by a mechanism that can set di erent prices for jobs of di erent lengths, and the ratio can be improved if we have more speci c knowledge of some parameters. Similarly, a mechanism that sets the same price for all servers even though the servers may receive di erent kinds of jobs can provide a reasonable welfare and revenue approximation compared to a mechanism that is allowed to set di erent prices for di erent servers.

show abstract

Cloud Pricing: The Spot Market Strikes Back

Cited by 11 publications

References 32 publications

Reservation and Checkpointing Strategies for Stochastic Jobs

Reservation and Checkpointing Strategies for Stochastic Jobs

On the cluster admission problem for cloud computing

Simple pricing schemes for the cloud

Contact Info

Product

Resources

About