Brief announcement: application-sensitive QoS scheduling in storage servers

The growing popularity of multi-tenant, cloud-based computing platforms is driving research into new QoS models that permit flexible sharing of the underlying infrastructure. In this paper, we re-examine the use of the commonly-used proportional-share model for resource allocation, in the context of modern heterogeneous, multi-tiered storage systems. We highlight the limitations of a conventional proportional sharing approach to resource allocation, and describe a new allocation model that provides strong isolation between clients. This improves the performance characteristics from the viewpoints of both the clients and the service provider.