Persistent Data Storage in KVM-based Clouds


One Line Summary

A discussion of implementation and performance issues for persistent data storage in KVM-based clouds


In this discussion, we will take a look at implemetation and performance issues for persistent data storage in clouds which use Kernel Virtual Monitor (KVM) as the hypervisor. We’ll discuss how persistent data storage could be implemented and how KVM plays a crucial role in that storage stack. As a case study, we will focus on a cloud infrastructure with IBM Scale-Out Network Attached Storage (SONAS) as the storage backend. We’ll examine how persistent data storage requests flow between cloud customers and SONAS. Finally we’ll discuss performance issues and how those issues could be addressed to improve end-to-end latency and throughput for persistent data storage.

In summary, we hope this presentation will give the audience a good sense of how persistent data storage could be implemented and optimized in a KVM-based cloud environment.


virtualization, kvm, Cloud computing, Storage Performance


  • Khoa


    Khoa Huynh joined IBM in 1989 where he first worked on OS/2 in the areas of kernel memory management, remote client management, and system performance. He joined IBM’s Linux Technology Center in 2001 where he led the Linux defect support team working with major Linux distributors, and later, joined the virtualization development team, contributing to XEN’s full virtualization support for Windows. Khoa is currently on the Linux Performance team, focusing on the performance of virtualization technologies, such as the IBM PowerVM, XEN, and KVM. Khoa also holds a doctorate in Computer Science. His areas of interest include virtualization, system architectures, and cloud computing. Most recently, Khoa presented his work in KVM performance at the 2010 Linux Plumbers Conference.

  • Biography

    Badari Pulavarty joined IBM’s Linux Technology Center (LTC) in 2001. Over the last 10 years, Badari contributed to various subsystems in Linux kernel. His interests are in performance and scalability of block IO, filesystems, DIO, AIO. Badari is currently on the KVM development team, focusing on storage scalability, performance and APIs for cloud offerings based on KVM.