Red Hat Enterprise Linux KVM hypervisor I-O.pdf

(890 KB) Pobierz
Red Hat Enterprise Linux
KVM Hypervisor I/O
Red Hat and IBM Demonstrate Top Virtualization
Performance with KVM on IBM's x3850X5
Khoa Huynh, Ph.D. - Linux Technology Center, IBM
Andrew Theurer - Linux Technology Center, IBM
Dor Laor - Senior Engineering Manager, Red Hat
Version 1.0
August 2012
1801 Varsity Drive™
Raleigh NC 27606-2072 USA
Phone: +1 919 754 3700
Phone: 888 733 4281
Fax: +1 919 754 3701
PO Box 13588
Research Triangle Park NC 27709 USA
Linux is a registered trademark of Linus Torvalds.
Red Hat, Red Hat Enterprise Linux and the Red Hat "Shadowman" logo are registered trademarks of
Red Hat, Inc. in the United States and other countries.
IBM, the IBM logo, ibm.com, ServeRAID, System x, X-Architecture, and all other IBM products and
services mentioned herein are trademarks of International Business Machines Corporation in the
United States and other countries
UNIX is a registered trademark of The Open Group.
Intel, the Intel logo and Xeon are registered trademarks of Intel Corporation or its subsidiaries in the
United States and other countries.
All other trademarks referenced herein are the property of their respective owners.
© 2012 by Red Hat, Inc and IBM Corp. This material may be distributed only subject to the terms and
conditions set forth in the Open Publication License, V1.0 or later (the latest version is presently
available at
http://www.opencontent.org/openpub/).
The information contained herein is subject to change without notice. Red Hat, Inc. shall not be liable
for technical or editorial errors or omissions contained herein.
Distribution of modified versions of this document is prohibited without the explicit permission of Red
Hat Inc.
Distribution of this work or derivative of this work in any standard (paper) book form for commercial
purposes is prohibited unless prior permission is obtained from Red Hat Inc.
The GPG fingerprint of the
security@redhat.com
key is:
CA 20 86 86 2B D6 9D FC 65 F6 EC C4 21 91 80 CD DB 42 A6 0E
Send feedback to
refarch-feedback@redhat.com
www.redhat.com
ii
refarch-feedback@redhat.com
Table of Contents
1 Executive Summary........................................................................................ 1
2 Introduction...................................................................................................... 2
2.1 The KVM hypervisor......................................................................................................... 2
2.2 Red Hat
®
Enterprise Linux
®
6 Virtualization.................................................................... 3
2.3 IBM
®
System x
®
Servers.................................................................................................. 3
2.4 Motivation......................................................................................................................... 4
3 Test Setup....................................................................................................... 5
3.1 Test Hardware.................................................................................................................. 5
3.2 Workload........................................................................................................................... 5
3.3 KVM Configuration........................................................................................................... 7
4 Results............................................................................................................ 8
4.1 Single VM......................................................................................................................... 8
4.2 Multiple VMs................................................................................................................... 11
5 Summary....................................................................................................... 15
Appendix A: References................................................................................... 16
refarch-feedback@redhat.com
iii
www.redhat.com
1 Executive Summary
The Kernel-based Virtual Machine (KVM) hypervisor has earned a reputation as the highest
performing hypervisor in virtualization benchmarks, holding the top seven results in
SPECvirt_sc2010 [1] and recently achieving a leadership spot among x86-virtualized results
on SAP's 2-tier SD benchmark [2]. One of the key ingredients to this success is the KVM
hypervisor's ability to handle the high I/O rates required by enterprise workloads, such as
databases, ERP (Enterprise Resource Planning) systems, and low-latency financial trading
applications, that are running in virtual machines.
This paper describes the test environment which is comprised of an IBM
®
System x3850 X5
host server with QLogic
®
QLE 256x Host Bus Adapters (HBAs), Red Hat
®
Enterprise Linux
®
6.3 hypervisor, and Red Hat Enterprise Linux 6.3 guests. The test results validate that a
single KVM guest can handle more than 800,000 I/Os per second (IOPS) and four KVM
guests running on the same host can support an aggregate I/O rate of more than 1.4 million
IOPS – hitting the “bare-metal” limit of the test storage setup.
These
results show that KVM
can achieve the highest storage I/O performance in a virtualized environment and is ready for
the most demanding enterprise workloads.
refarch-feedback@redhat.com
1
4
www.redhat.com
2 Introduction
2.1 The KVM hypervisor
A hypervisor is a specialized operating system that runs virtual machines instead of physical
applications. The Kernel-based Virtual Machine (KVM) project represents the next generation
in open-source virtualization. KVM is fully integrated into the Linux operating system both as
a host and a guest. Unlike other hypervisors, KVM does not make any distinction between
running in either host or hypervisor mode. This duality in design has helped KVM to rapidly
mature into a stable, high-performing hypervisor, positioned to outperform any other
hypervisor available on the market today.
The first design principle includes the following:
Leverage all hardware-assisted virtualization capabilities provided by Intel
®
Virtualization Technology (VT) and AMD
®
Secure Virtual Machine (SVM).
Feature the latest hardware virtualization extensions, including:
Hardware nested paging (EPT/NPT)
Pause loop filtering
I/O off-load features, such as secure PCI pass-through using Intel VT-D or AMD
I/O Memory Management Unit (IOMMU)
Exploit hardware capabilities while keeping the KVM virtualization overhead to the
absolute minimum
The second design principle includes the following:
Leverage the Linux operating system
Fulfill the many components required by a hypervisor, such as memory management,
scheduler, I/O stack, and device drivers by reusing optimized, off-the-shelf Linux
implementations
The Linux kernel, with its 20 years of development, is the industry leader in terms of
performance and availability. The Linux process scheduler, for example, provides completely
fair scheduling (CFS) that is optimized to manage complex workloads and NUMA systems,
while offering low latency, high performance determinism, and fine-grained Service Level
Agreement (SLA) for applications. By placing the KVM hypervisor directly into the Linux
kernel, all of these services and advantages have a direct impact on the hypervisor
performance.
www.redhat.com
2
refarch-feedback@redhat.com
Zgłoś jeśli naruszono regulamin