Towards performance isolation in multi-tenant SaaS applications

Abstract

Multi-tenancy has shown promising results in achieving high operational cost efficiency by sharing hardware and software resources among multiple customer organisations, called tenants. In the context of cloud computing, this paradigm enables cloud providers to reduce operational costs by dividing resources and to simplify application management and maintenance. Maximum cost efficiency is achieved with application-level multi-tenancy. However, this high level of resource sharing complicates performance isolation between the different tenants, i.e. ensuring compliance with the SLAs of the different tenants and ensuring that the behaviour of one tenant cannot adversely affect the performance of the other tenants.This paper explores the challenges of performance isolation in the context of multi-tenant SaaS applications. In addition, we propose a middleware architecture to enforce performance isolation based on the tenant-specific SLAs, using a tenant-aware profiler and a scheduler. Our prototype reveals promising initial results.

Implementing the Required Degree of Multitenancy Isolation: A Case Study of Cloud-Hosted Bug Tracking System

“…Bauer and Adams [5] discussed how to use virtualization to ensure that the failure of one tenants' instance does not cascade to other tenants' instances. The work of Walraven et al [13] is closely related to ours. By using a multitenant implementation of a hotel booking application deployed on top of a cluster, the authors implemented a middleware framework for enforcing performance isolation.…”

“…; and (iv) creating a heavy load burst using the Synchronous Timer so that a certain number of request are fired at the same time. This type of configuration can be likened to unpredictable (i.e., sudden increase) workload [4] and aggressive load [13].…”

Evaluating Degrees of Multitenancy Isolation: A Case Study of Cloud-Hosted GSD Tools

“…We also added the Synchronous Timer to the Samplers and then reduce the ramp-up period by one-tenth so that all the requests are sent ten times faster. The scenario is similar to unpredictable (i.e., sudden increase) workload [1] and aggressive load [18]. Scenario 2 -Variation in request arrival rate: This scenario represents a case where there is variation in the frequency with which code changes are committed to the source code to trigger a build process.…”

“…Bauer and Adams [2] discusses how to use virtualization to ensure that the failure of one tenants instance does not cascade to other tenant instances. A closely related work to ours is that of Walraven et al [18] where the authors implemented a middleware framework for enforcing performance isolation. They used a multitenant implementation of a hotel booking application deployed on top of a cluster for illustration.…”

Evaluating Degrees of Isolation between Tenants enabled by Multitenancy Patterns for Cloud-hosted Version Control Systems (VCS)

“…Guo et al [11] evaluated different isolation capabilities related to authentication, information protection, faults, administration etc. A closely related work to ours is that of Walraven et al [12] where they implemented a middleware framework for enforcing performance isolation. The authros used a multitenant implementation of a hotel booking application deployed on top of a cluster for illustration.…”