CPU Sharing Techniques for Performance Isolation in Multi-tenant Relational Database-as-a-Service

Sudipto Das, Vivek Narasayya, Feng Li, and Manoj Syamala

Abstract

Multi-tenancy and resource sharing are essential to make a Database-as-a-Service (DaaS) cost-effective. However, one major consequence of resource sharing is that the performance of one tenant's workload can be significantly affected by the resource demands of co-located tenants. The lack of performance isolation in a shared environment can make DaaS less attractive to performance-sensitive tenants. Our approach to performance isolation in a DaaS is to isolate the key resources needed by the tenants' workload. In this paper, we focus on the problem of effectively sharing and isolating CPU among co-located tenants in a multi-tenant DaaS. We show that traditional CPU sharing abstractions and algorithms are inadequate to support several key new requirements that arise in DaaS: (a) absolute and fine-grained CPU reservations without static allocation; (b) support elasticity by dynamically adapting to bursty resource demands; and (c) enable the DaaS provider to suitably tradeoff revenue with fairness. We implemented these new scheduling algorithms in a commercial DaaS prototype and extensive experiments demonstrate the effectiveness of our techniques.

Details

Publication typeInproceedings
Published inProceedings of the VLDB Endowment
Pages12
Volume7
Number1
PublisherVery Large Data Bases Endowment Inc.
> Publications > CPU Sharing Techniques for Performance Isolation in Multi-tenant Relational Database-as-a-Service