1. Lecture 7: Performance Issues with Virtualization Xiaowei Yang (Duke University)

2. Recap What’s cloud computing Security issues

3. Today How virtualization impacts performance (measurement) – The Impact of Virtualization on Network Performance of Amazon EC2 Data Center Guohui Wang, T. S. Eugene Ng – Performance Profiling in a Virtualized Environment Jiaqing Du, Nipun Sehrawat, and Willy Zwaenepoel

4. Virtualization Impacts Network Performance Cloud computing uses virtualization to provide resource sharing – Time sharing – Processors, NIC, harddisk, performance monitoring unit Xen Dom0 Guest0 Guest1 Driver domain NIC

5. Measure the impact of virtualization Processor sharing Packet round-trip delay TCP/UDP throughput Packet loss

6. Processor sharing Measurement methodology – CPUTest: loops for a million times – Gettimeofday() at the beginning and end of the loop – Saves timestamps in memory – Dump the saves timestamps to disks

7. Packet round-trip delay 10 ping probes per second, 5000 round- trip delay measurements

8. TCP/UDP throughput UDPTest – The program itself must be able to read and send data as fast as possible 128KB UDPSend Dom0 Xen NIC UDPRecv Dom0 Xen NIC

9. TCP/UDP throughput TCPTest – Send and Receive window set to 256KB – 0.5ms –  4Gbps 128KB TCPSend Dom0 Xen NIC TCPRecv Dom0 Xen NIC

10. Packet loss Badabing – Packet loss estimation – Active probes and statistical estimations

11. Experiments setup: spatial Across different instances, availability zone For each availability zone (3 total) – 250 pairs of small instance – 50 pairs of medium instances – TCP/UDP May congest the network Limit to 800MB data – Badabing One minute

12. Experiments setup: temporal Over a period of time Two small instance pairs and one medium in each of the availability zone – Processor sharing, network performance, and packet loss – Round by round for 150 hours – Each round separate for 10 minutes

13. Results

14. Processor sharing

15. How to compute CPU share Loop > 1ms  off Loop ~ 3us  on

16. Distribution of CPU share

17. Bandwidth measurement results

19. A closer look

20. Medium instances’ performance

21. Why? TCP/UDP sender is scheduled off UDP > TCP – Sender can send when receiver is scheduled off

22. End-to-end delays

23. Why Queuing? – Unlikely, rare packet loss – <500Mb/s TCP throughput  no congestion Receiver scheduled off –  no reply until it’s on again

24. Packet loss estimation

25. Why? Measurement tool not considering virtualization does not fit – Badabing detects loss episodes – Loss episode A packet loss within some time of now One way delay is larger than a threshold  Large variation leads to a high loss estimate

26. Badabing overestimates loss

27. Implications Delay = queuing + transmission + propagation + scheduled off Affect the fidelity of network testing if using small instances May cause performance problems for some applications –  yet to be validated

28. Discussion Easy fix: using medium or large instances I/O sharing Better way to share networks? Ways to detect whether you are sharing with another instance Try it on other instances?