MTD-DS: an SLA-aware Decision Support Benchmark for Multi-tenant Parallel DBMSs - Optimisation Dynamique de Requêtes Réparties à grande échelle Access content directly
Reports Year : 2023

MTD-DS: an SLA-aware Decision Support Benchmark for Multi-tenant Parallel DBMSs

Abstract

Multi-tenant DBMSs are used by Cloud providers for their DBaaS (Database-as-a-Service) products. They could be Single-node RDBMSs installed in VMs, SQL-on-Hadoop systems running on a cluster, or parallel RDBMSs with a shared-nothing or shareddisk architecture. From a Cloud provider's point of view, it is interesting to measure these systems' capability of dealing with multi-tenant workloads. From a tenant's point of view, having the above information on different providers could be helpful in choosing the most suitable one (or several for a multi-cloud deployment). In this paper, we present MTD-DS benchmark (with MTD for Multi-Tenant parallel DBMSs and DS for Decision Support), which extends TPC-DS by adding a multitenant query workload generator, a performance SLO (Service Level Objective) generator, configurable DBaaS pricing models, and new metrics to measure the potential capability of a multitenant parallel DBMS in obtaining the best trade-off between the provider's benefit and the tenants' satisfaction. Example experimental results have been produced to show the relevance and the feasibility of the MTD-DS benchmark.
Fichier principal
Vignette du fichier
An SLA-aware Decision Support Benchmark for Multi-tenant DBMSs.pdf (691.71 Ko) Télécharger le fichier
Origin Files produced by the author(s)
licence
Copyright

Dates and versions

hal-04312262 , version 1 (28-11-2023)

Licence

Copyright

Identifiers

  • HAL Id : hal-04312262 , version 1

Cite

Shaoyi Yin, Franck Morvan, Jorge Martinez-Gil, Abdelkader Hameurlain. MTD-DS: an SLA-aware Decision Support Benchmark for Multi-tenant Parallel DBMSs. IRIT/RR--2023--05--FR, IRIT - Institut de Recherche en Informatique de Toulouse. 2023. ⟨hal-04312262⟩
137 View
40 Download

Share

Gmail Mastodon Facebook X LinkedIn More