Towards Multi-tenant Performance SLOs

Lang, Willis; Shankar, Srinath; Patel, Jignesh M.; Kalhan, Ajay

doi:10.1109/icde.2012.101

Cited by 68 publications

(42 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Methods to guarantee performance SLOs under transactional and analytical multi-tenant workloads have been proposed in [23] and [6]. Both works consider profiling-based analyses, but rather focus on non-in-memory hardware configurations.…”

Section: Related Workmentioning

confidence: 99%

Experiments or simulation? A characterization of evaluation methods for in-memory databases

Molka

Casale

2015

2015 11th International Conference on Network and Service Management (CNSM)

View full text Add to dashboard Cite

Abstract-The recent growth of interest for in-memory databases poses the question on whether established prediction methods such as response surfaces and simulation are effective to describe the performance of these systems. In particular, the limited dependence of in-memory technologies on the disk makes methods such as simulation more appealing than in the past, since disks are difficult to simulate. To answer this question, we study an in-memory commercial solution, SAP HANA, deployed on a high-end server with 120 physical cores. First, we apply experimental design methods to generate response surfaces that describe database performance as a function of workload and hardware parameters. Next, we develop a class-switching queueing network model to predict in-memory database performance under similar scenarios. By comparing the applicability of the two approaches to modeling multi-tenancy, we find that both queueing and response surface models yield mean prediction errors in the range 5%-22% with respect to mean memory occupancy and response times, but the accuracy for the latter deteriorates in response surfaces as the number of experiments are reduced, whereas simulation is effective in all cases. This suggests that simulation can be very effective in performance prediction for in-memory database management.

show abstract

Section: Related Workmentioning

confidence: 99%

Experiments or simulation? A characterization of evaluation methods for in-memory databases

Molka

Casale

2015

2015 11th International Conference on Network and Service Management (CNSM)

View full text Add to dashboard Cite

show abstract

“…The problem of handling many (often in 1000s) small and independent databases on a multi-tenant database Platformas-a-Service (usually called PaaS or DBaaS) has received considerable attention in recent years [48,37,32,36,16]. That is not the problem we focus on in this paper.…”

Section: Introductionmentioning

confidence: 79%

Tempo

Tan

Babu

2016

Proc. VLDB Endow.

View full text Add to dashboard Cite

Multi-tenant database systems have a component called the Resource Manager, or RM that is responsible for allocating resources to tenants. RMs today do not provide direct support for performance objectives such as: "Average job response time of tenant A must be less than two minutes", or "No more than 5% of tenant B's jobs can miss the deadline of 1 hour." Thus, DBAs have to tinker with the RM's lowlevel configuration settings to meet such objectives. We propose a framework called Tempo that brings simplicity, selftuning, and robustness to existing RMs. Tempo provides a simple interface for DBAs to specify performance objectives declaratively, and optimizes the RM configuration settings to meet these objectives. Tempo has a solid theoretical foundation which gives key robustness guarantees. We report experiments done on Tempo using production traces of data-processing workloads from companies such as Facebook and Cloudera. These experiments demonstrate significant improvements in meeting desired performance objectives over RM configuration settings specified by human experts.

show abstract

“…It is obvious that increasing the number of tenants decreases the per tenant performance but reduces the overall operating cost of the provider [24].…”

Section: A Provider's Revenuementioning

confidence: 99%

A Performance and Profit Oriented Data Replication Strategy for Cloud Systems

Tos

Mokadem

Hameurlain

et al. 2016

2016 Intl IEEE Conferences on Ubiquitous Intelligence &Amp; Computing, Advanced and Trusted Computing, Scalable Computing and C

View full text Add to dashboard Cite

OATAO is an open access repository that collects the work of Toulouse researchers and makes it freely available over the web where possible. This is an author-deposited version published in : http://oatao.univ-toulouse.fr/ Eprints ID : 18792The contribution was presented at CBDCom 2016 :https://cbdcom2016.sciencesconf.org/ Abstract-In today's world, tenants of cloud systems expect timely responses to queries that process ever-increasing sizes of data. However, most cloud providers offer their services without any performance guarantees to their tenants. In this paper we propose a data replication strategy that aims to satisfy performance guarantees for the tenant while ensuring profitability of the cloud provider. Our strategy estimates the response time of the queries, as well as the expenditures that affect the profitability of the cloud provider. The decision of whether to perform replication is determined by the fulfillment of these two criteria. Validity of the proposed strategy is provided by means of a simulation study.

show abstract

Towards Multi-tenant Performance SLOs

Cited by 68 publications

References 27 publications

Experiments or simulation? A characterization of evaluation methods for in-memory databases

Experiments or simulation? A characterization of evaluation methods for in-memory databases

Tempo

A Performance and Profit Oriented Data Replication Strategy for Cloud Systems

Contact Info

Product

Resources

About