Abstract
Virtualization has become one of the backbone technologies for constructing cloud servers and data centers, in that the technology supports to partition the physical computing resources of the system into multiple logical ones for concurrently running multiple VMs (virtual machines) in a single physical server. In constructing multiple VM environments, the qcow2 format is widely used as a virtual disk layout due to its attractive features such as high storage space efficiency and high reliability. Unfortunately, according to the policy of the qcow2 format, the hypervisor calls file sync operations excessively during allocating data clusters. Moreover, significant performance degradation may occur due to the CoW (Copy-on-Write) operations that are unconditionally performed by the hypervisor whenever new data clusters are allocated on the virtual disk. In this paper, we introduce a transactional hypervisor that mitigates the performance overhead caused by the data cluster allocation onto the qcow2-based virtual disk. To achieve this, we adopt transactional support to the existing hypervisor and propose a new type of file sync operation, called gsync, which flushes the modified metadata in a bundle. Moreover, we introduce an optimization technique on the CoW mechanism that performs CoW operations selectively according to the amount of data that will be updated on each data cluster. The experimental results clearly show that the proposed hypervisor outperforms the conventional hypervisor in terms of write IOPS (Input/Output Per Second) by up to 78.4%, by reducing the number of file sync calls and that of CoW operations by up to 50.6% and 52.6%, respectively.
Highlights
Nowadays, several types of computing systems adopt virtualization technology into their systems to fully utilize their own computing resources or to provide diverse computing platforms on a physical machine [1]
The proposed hypervisor is designed to answer the simple question: how to mitigate the performance overhead caused by the data cluster allocation of the qcow2-based virtual disk? To resolve this problem, we propose a transactional hypervisor which improves the performance of the qcow2-based virtual disk by reducing excessive file sync calls and unnecessary CoW operations in the process of data cluster allocation
Group-sync: Through several analyses, which will be discussed in detail in Section III, we found that the conventional hypervisor provides strong consistency for the qcow2-based virtual disk by frequently incurring file sync calls during the process of data cluster allocation
Summary
Several types of computing systems adopt virtualization technology into their systems to fully utilize their own computing resources or to provide diverse computing platforms on a physical machine [1]. We propose a transactional hypervisor which improves the performance of the qcow2-based virtual disk by reducing excessive file sync calls and unnecessary CoW operations in the process of data cluster allocation. Transactional support: We found that the hypervisor acquires qemu_lock to handle a write request transferred from a VM It splits the write request in units of data clusters and reflects them one by one into the qcow2-based virtual disk with the file sync calls. Group-sync (gsync): Through several analyses, which will be discussed in detail, we found that the conventional hypervisor provides strong consistency for the qcow2-based virtual disk by frequently incurring file sync calls during the process of data cluster allocation.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
