Automated performance control in a virtual distributed storage system

Storage demand in large organizations has grown rapidly in the last few decades. In this paper, we describe Storage@desk, a new virtual distributed storage system that utilizes a large number of distributed machines to provide storage services with quality of service guarantees. We present the Storage@desk architecture and core components. The machines on which Storage@desk relies exist in an environment that is unreliable and dynamic. We utilize replication to provide high levels of availability and reliability. Our evaluation reveals that Storage@desk achieves better read and write performance compared to CIFS. As Storage@desk serves clients and applications on shared storage resources, it is crucial to ensure predictable storage access even when the workloads are unknown a priori. To this end, we take a control theoretic approach for automated performance control in Storage@desk. Given a reference value, the feedback controller is able to effectively regulate service requests to virtual storage resources under various scenarios.