際際滷

際際滷Share a Scribd company logo

Duganiperfn43 120911020533-phpapp02

Jamil Jamil
Jamil Jamil

This document discusses using Iperf, a tool for active measurements of the maximum achievable bandwidth on IP networks. It describes how Iperf can test TCP and UDP throughput, and notes some limitations in TCP measurements. The document provides examples of Iperf commands and options to measure bandwidth, including adjusting buffers, parallel transfers, and reporting intervals. It also describes how to use Iperf to generate very high bandwidth streams, with cautions about unintended network impacts.

1 of 14
Download to read offline
1 Using Iperf Jon M. Dugan <jdugan@es.net> NANOG 43, Brooklyn, NY June 1, 2008 Energy Sciences Network Lawrence Berkeley National Laboratory Networking for the Future of Science
2 Outline TCP Measurements UDP Measurements Useful tricks
3 Iperfs notion of clients and servers Client is the sender Server is the receiver (discard server)
4 TCP Measurements Measures TCP Achievable Bandwidth Measurement includes the end system Sometimes called memory-to-memory tests Limits of what we can measure TCP is a largely a black box Many things can limit TCP throughput Loss Congestion Buffer Starvation Out of order delivery
5 Example Iperf TCP Invocation Server (receiver): $ iperf -s ------------------------------------------------------------ Server listening on TCP port 5001 TCP window size: 85.3 KByte (default) ------------------------------------------------------------ [ 4] local 10.0.1.5 port 5001 connected with 10.0.1.10 port 60830 [ 4] 0.0-10.0 sec 1.09 GBytes 933 Mbits/sec [ 4] local 10.0.1.5 port 5001 connected with 10.0.1.10 port 60831 [ 4] 0.0-10.0 sec 1.08 GBytes 931 Mbits/sec Client (sender): $ iperf -c 10.0.1.5 ------------------------------------------------------------ Client connecting to 10.0.1.5, TCP port 5001 TCP window size: 129 KByte (default) ------------------------------------------------------------ [ 3] local 10.0.1.10 port 60830 connected with 10.0.1.5 port 5001 [ ID] Interval Transfer Bandwidth [ 3] 0.0-10.2 sec 1.09 GBytes 913 Mbits/sec
6 Bandwidth Delay Product The amount of in flight data allowed for a TCP connection BDP = bandwidth * round trip time Example: 1Gb/s cross country, ~100ms 1,000,000,000 b/s * .1 s = 100,000,000 bits 100,000,000 / 8 = 12,500,000 bytes 12,500,000 bytes / (1024*1024) ~ 12MB To get full TCP performance the TCP window needs to be large enough to accommodate the Bandwidth Delay Product
7 UDP Measurements UDP provides greater transparency We can directly measure some additional things: Loss Jitter Out of order delivery
8 Example Iperf UDP Invocation Server (receiver): $ iperf -u -s ------------------------------------------------------------ Server listening on UDP port 5001 Receiving 1470 byte datagrams UDP buffer size: 107 KByte (default) ------------------------------------------------------------ [ 3] local 10.0.1.5 port 5001 connected with 10.0.1.10 port 65299 [ 3] 0.0-10.0 sec 1.25 MBytes 1.05 Mbits/sec 0.008 ms 0/ 893 (0%) Client (sender): $ iperf -u -c 10.0.1.5 -b 1M ------------------------------------------------------------ Client connecting to 10.0.1.5, UDP port 5001 Sending 1470 byte datagrams UDP buffer size: 9.00 KByte (default) ------------------------------------------------------------ [ 3] local 10.0.1.10 port 65300 connected with 10.0.1.5 port 5001 [ ID] Interval Transfer Bandwidth [ 3] 0.0-10.0 sec 1.25 MBytes 1.05 Mbits/sec [ 3] Server Report: [ 3] 0.0-10.0 sec 1.25 MBytes 1.05 Mbits/sec 0.003 ms 0/ 893 (0%) [ 3] Sent 893 datagrams
9 Adjusting Iperf for performance The w option for Iperf can be used to request a particular buffer size. This sets both send and receive buffer size. The OS may need to be tweaked to allow buffers of sufficient size. See http://dsd.lbl.gov/TCP-tuning/ and http://www.psc.edu/networking/perf_tune.html Parallel transfers may help as well, the P option can be used for this
10 Useful Iperf Invocations UDP and TCP: -i n report status every n seconds -d do bidirectional test simultaneously -r do bidirectional test one after another
11 Using Iperf to generate high rate streams UDP doesnt require a receiver If you have good counters on your switches & routers those can be used to measure Turns out UDP reception can be very resource intensive resulting in drops at the NIC at high rates (8-9 Gb/s)
12 Never do this Need to generate 10 Gb/s but only have a 1 Gb/s host? Iperf UDP 1 Gb/s Destined for 10.1.1.1 Use the T option to Iperf to control the number of times the traffic loops Can also use firewall filters to discard a certain TTL range. Other filters may be prudent as well. 10.0.1.1 10.0.1.2 Static route: 10.1.1.1/32 10.0.1.2 Static route: 10.1.1.1/32 10.0.1.1
13 Iperf Development Primarily in maintenance mode Accepting and apply patches Fixing bugs and documentation as time allows Future Directions libiperf
14 More Information http://iperf.sourceforge.net iperf-users@lists.sourceforge.net You can reach me at: Jon Dugan <jdugan@es.net>

More Related Content

Duganiperfn43 120911020533-phpapp02

  • 1. 1 Using Iperf Jon M. Dugan <jdugan@es.net> NANOG 43, Brooklyn, NY June 1, 2008 Energy Sciences Network Lawrence Berkeley National Laboratory Networking for the Future of Science
  • 2. 2 Outline TCP Measurements UDP Measurements Useful tricks
  • 3. 3 Iperfs notion of clients and servers Client is the sender Server is the receiver (discard server)
  • 4. 4 TCP Measurements Measures TCP Achievable Bandwidth Measurement includes the end system Sometimes called memory-to-memory tests Limits of what we can measure TCP is a largely a black box Many things can limit TCP throughput Loss Congestion Buffer Starvation Out of order delivery
  • 5. 5 Example Iperf TCP Invocation Server (receiver): $ iperf -s ------------------------------------------------------------ Server listening on TCP port 5001 TCP window size: 85.3 KByte (default) ------------------------------------------------------------ [ 4] local 10.0.1.5 port 5001 connected with 10.0.1.10 port 60830 [ 4] 0.0-10.0 sec 1.09 GBytes 933 Mbits/sec [ 4] local 10.0.1.5 port 5001 connected with 10.0.1.10 port 60831 [ 4] 0.0-10.0 sec 1.08 GBytes 931 Mbits/sec Client (sender): $ iperf -c 10.0.1.5 ------------------------------------------------------------ Client connecting to 10.0.1.5, TCP port 5001 TCP window size: 129 KByte (default) ------------------------------------------------------------ [ 3] local 10.0.1.10 port 60830 connected with 10.0.1.5 port 5001 [ ID] Interval Transfer Bandwidth [ 3] 0.0-10.2 sec 1.09 GBytes 913 Mbits/sec
  • 6. 6 Bandwidth Delay Product The amount of in flight data allowed for a TCP connection BDP = bandwidth * round trip time Example: 1Gb/s cross country, ~100ms 1,000,000,000 b/s * .1 s = 100,000,000 bits 100,000,000 / 8 = 12,500,000 bytes 12,500,000 bytes / (1024*1024) ~ 12MB To get full TCP performance the TCP window needs to be large enough to accommodate the Bandwidth Delay Product
  • 7. 7 UDP Measurements UDP provides greater transparency We can directly measure some additional things: Loss Jitter Out of order delivery
  • 8. 8 Example Iperf UDP Invocation Server (receiver): $ iperf -u -s ------------------------------------------------------------ Server listening on UDP port 5001 Receiving 1470 byte datagrams UDP buffer size: 107 KByte (default) ------------------------------------------------------------ [ 3] local 10.0.1.5 port 5001 connected with 10.0.1.10 port 65299 [ 3] 0.0-10.0 sec 1.25 MBytes 1.05 Mbits/sec 0.008 ms 0/ 893 (0%) Client (sender): $ iperf -u -c 10.0.1.5 -b 1M ------------------------------------------------------------ Client connecting to 10.0.1.5, UDP port 5001 Sending 1470 byte datagrams UDP buffer size: 9.00 KByte (default) ------------------------------------------------------------ [ 3] local 10.0.1.10 port 65300 connected with 10.0.1.5 port 5001 [ ID] Interval Transfer Bandwidth [ 3] 0.0-10.0 sec 1.25 MBytes 1.05 Mbits/sec [ 3] Server Report: [ 3] 0.0-10.0 sec 1.25 MBytes 1.05 Mbits/sec 0.003 ms 0/ 893 (0%) [ 3] Sent 893 datagrams
  • 9. 9 Adjusting Iperf for performance The w option for Iperf can be used to request a particular buffer size. This sets both send and receive buffer size. The OS may need to be tweaked to allow buffers of sufficient size. See http://dsd.lbl.gov/TCP-tuning/ and http://www.psc.edu/networking/perf_tune.html Parallel transfers may help as well, the P option can be used for this
  • 10. 10 Useful Iperf Invocations UDP and TCP: -i n report status every n seconds -d do bidirectional test simultaneously -r do bidirectional test one after another
  • 11. 11 Using Iperf to generate high rate streams UDP doesnt require a receiver If you have good counters on your switches & routers those can be used to measure Turns out UDP reception can be very resource intensive resulting in drops at the NIC at high rates (8-9 Gb/s)
  • 12. 12 Never do this Need to generate 10 Gb/s but only have a 1 Gb/s host? Iperf UDP 1 Gb/s Destined for 10.1.1.1 Use the T option to Iperf to control the number of times the traffic loops Can also use firewall filters to discard a certain TTL range. Other filters may be prudent as well. 10.0.1.1 10.0.1.2 Static route: 10.1.1.1/32 10.0.1.2 Static route: 10.1.1.1/32 10.0.1.1
  • 13. 13 Iperf Development Primarily in maintenance mode Accepting and apply patches Fixing bugs and documentation as time allows Future Directions libiperf