Xen versus kvm_slides_20080623


Published on

We present initial results from and quantitative analysis of two leading open source hypervisors, Xen and KVM. This study focuses on the overall performance, performance isolation, and scalability of virtual machines running on these hypervisors. Our comparison was carried out using a benchmark suite that we developed to make the results easily repeatable. Our goals are to understand how the different architectural decisions taken by different hypervisor developers affect the resulting hypervisors, to help hypervisor developers realize areas of improvement for their hypervisors, and to help users make informed decisions
about their choice of hypervisor.

Published in: Technology
  • Be the first to comment

  • Be the first to like this

No Downloads
Total views
On SlideShare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide

Xen versus kvm_slides_20080623

  1. 1. Todd Deshane, Ph.D. Student, Clarkson UniversityXen Summit, June 23-24, 2008, Boston, MA, USA.
  2. 2. Xen and the Art of Virtualization (2003)◦ Reported remarkable performance resultsXen and the Art of Repeated Research (2004)◦ Validated performance resultsQuantifying the Performance IsolationProperties of Virtualization Systems (2007)◦ Isolation Benchmark Suite◦ Performance isolation testing methodology◦ Lots of attention from virtualization developers and industry Xen Summit Boston 2008
  3. 3. Understand architectural differences◦ Stand-alone versus integrated hypervisorHelp developers realize areas of improvement◦ Difficult for developers to test all cases◦ Overall performance is important, but not the only factorHelp users make informed decisions◦ Growing number of virtualization options to choose from◦ Different users have different virtualization needs◦ Hardware and software versions can make a big difference Xen Summit Boston 2008
  4. 4. Base machine◦ Ubuntu Linux 8.04 AMD64Software packages from Ubuntu repositories◦ Linux kernel 2.6.24-18◦ Xen 3.2.1+2.6.24-18-xen◦ KVM 62Guests◦ Ubuntu Linux 8.04 AMD64◦ Automated debootstrapHardware◦ 2.4 GHz Intel Core 2 CPU 6600, 4 GB of RAM, 250 GB disk Xen Summit Boston 2008
  5. 5. Overall performance◦ Performance of the whole system and components◦ Focused on macro-benchmarks◦ Standard benchmarks and repeatable methodsPerformance isolation◦ Protection from resource consumption from other guests◦ Representative workload with and without stress testsScalability◦ Ability to run more guests without loss of performance◦ Same workload on each guest, increase number of guests Xen Summit Boston 2008
  6. 6. Macro-benchmarks◦ CPU◦ Kernel compile◦ Disk I/OAutomated guest build◦ Benchvm virtualization benchmark suiteAutomated testing and reporting◦ Phoronix Test Suite Xen Summit Boston 2008
  7. 7. Xen Summit Boston 2008
  8. 8. Xen and KVM had similar CPU performance◦ Xen: 0.999, KVM: 0.993Xen was better than KVM on kernel compile◦ Xen: 0.487, KVM: 0.384KVM was better on disk I/O◦ Write – Xen: 0.855, KVM: 0.934◦ Read – Xen: 0.852, KVM: 0.994◦ Disk caching effects? Xen Summit Boston 2008
  9. 9. Isolation Benchmark Suite◦ Memory stress test: calloc()◦ Fork stress test: fork()◦ CPU stress test ◦ Mixed calculations in tight loop◦ Disk ◦ Threaded IOzone read and write◦ Network receiver ◦ Receive threaded UDP traffic from external host◦ Network sender ◦ Send threaded UDP traffic to external host Xen Summit Boston 2008
  10. 10. Xen Summit Boston 2008
  11. 11. Xen Summit Boston 2008
  12. 12. Xen was isolated on memory, fork, CPU, diskXen was slightly isolated on network senderXen showed no isolation on network receiver◦ Kernel bug(s) in Ubuntu?Xen had unexpectedly good disk performanceKVM was well-isolated on all stress testsKVM had unexpectedly good network senderperformanceKVM had unexpectedly poor disk and networkreceiver performance Xen Summit Boston 2008
  13. 13. Ran Apache compile in 1, 2, 4, 16, and 30guestsMeasured compile time and number of gueststhat ran to completion Xen Summit Boston 2008
  14. 14. Xen Summit Boston 2008
  15. 15. Xen scaled linearly with respect to number ofguestsKVM had many guest crashes◦ 4 guests: 1 crashed guest◦ 8 guests: 4 crashed guests◦ 16 guests: 7 crashed guests◦ 30 guests: system crashed during compile Xen Summit Boston 2008
  16. 16. Virtualization benchmarking is still difficultTesting on multiple categories is crucialAutomated testing is important and useful◦ Transparency in testing methods◦ Repeatability is needed, yet challenging◦ Always more cases to test◦ Challenging to adequately benchmark rapidly evolving technologies Xen Summit Boston 2008
  17. 17. Extend testing to include Xen HVM, and KVMwith paravirt I/OMore complete automation of testing processwith benchvmPort benchvm to PythonAdd support for more distros in benchvmUse Phoronix Test Suite-like functionality◦ Test profiles, test suites, batch benchmarking◦ Automated results parsing◦ Graphing/uploading of results◦ Automated system and test config collection and publishing Xen Summit Boston 2008
  18. 18. Benchmarking co-researchers◦ Zachary Shepherd, Jeanna Neefe Matthews, Muli Ben- Yehuda, Amit Shah, and Balaji RaoPerformance isolation and scalabilityresearchers◦ Wenjin Hu and Madhujith HapuarachchiEarly developers and testers of benchvm◦ Cyrus Katrak and Martin McDermottMembers of the Xen and KVM communities◦ Feedback and support Xen Summit Boston 2008
  19. 19. Xen Summit Boston 2008