Improving the disk I/O performance is always a critical research issue in cloud computing platforms, especially for cloud computing platforms with distributed data-intensive workloads. To resolve the problem of poor disk I/O, in this paper, we propose a novel scheduling framework in the Xen-based hypervisor. In our framework, we employ a locality-aware scheduler with a deadline constraint and provide a basic QoS controller to guarantee the throughput of the cloud platform. The experimental results show that our solution increases the IOPS of Xen-based cloud platform by 9%-13% compared to the solution of early-deadline-first (EDF), deadline scheduling algorithm or Flubber. Moreover, according to the schedulability test, our solution can improve the miss deadline ratio significantly, compared to the EDF scheduling algorithm and Flubber.
2014 14th IEEE/ACM International Symposium on Cluster, Cloud and Grid Computing (CCGrid), pp.532-535