1IPERF(1) User Manuals IPERF(1)
2
6 iperf - perform network traffic tests using network sockets. Metrics
7 include throughput and latency.
8
10 iperf -s [options]
11 iperf -c server [options]
13 iperf -u -s [options]
15 iperf -u -c server [options]
17
20 iperf 2 is a testing tool which performs network traffic measurements
21 using network sockets. The performance metrics supported include
22 throughput and latency. Iperf can use both TCP and UDP sockets (or pro‐
23 tocols.) It supports unidirectional, full duplex (same socket) and
24 bidirectional traffic, and supports multiple, simultaneous traffic
25 streams. It supports multicast traffic including source specific multi‐
26 cast (SSM) joins. Its multi-threaded design allows for peak perfor‐
27 mance. Metrics displayed help to characterize host to host network per‐
28 formance. Note: Setting the enhanced (-e) option provides all available
29 metrics.
30 The user must establish both a both a server (to receive traffic) and a
32 client (to generate and send traffic) for a test to occur. The client
33 and server typically are on different hosts or computers but need not
34 be.
35
37 -b, --bandwidth
38 set the target bandwidth and optional standard deviation per
39 <mean>,[<stdev>] (See NOTES for suffixes)
40 -e, --enhanced
42 Display enhanced output in reports otherwise use legacy report
43 (ver 2.0.5) formatting (see notes)
44 -f, --format [abkmgBKMG]
46 format to report: adaptive, bits, Bytes, Kbits, Mbits, Gbits,
47 KBytes, MBytes, GBytes (see NOTES for more)
48 -h, --help
50 print a help synopsis
51 -i, --interval < n[p] | f >
53 sample or display interval reports every n seconds (default) or
54 n packets (per optional p suffix.) If f is used then the inter‐
55 val will be each frame or burst. The frame interval reporting is
56 experimental. Also suggest a compile with fast-sampling, i.e.
57 ./configure --enable-fastsampling
58 -l, --len n[kmKM]
60 set read/write buffer size (TCP) or length (UDP) to n (TCP de‐
61 fault 128K, UDP default 1470)
62 --l2checks
64 perform layer 2 length checks on received UDP packets (requires
65 systems that support packet sockets, e.g. Linux)
66 -m, --print_mss
68 print TCP maximum segment size (MTU - TCP/IP header)
69 --NUM_REPORT_STRUCTS <count>
71 Override the default shared memory size between the traffic
72 thread(s) and reporter thread in order to mitigate mutex lock
73 contentions. The default value of 5000 should be sufficient for
74 1Gb/s networks. Increase this upon seeing the Warning message of
75 reporter thread too slow. If the Warning message isn't seen,
76 then increasing this won't have any significant effect (other
77 than to use some additional memory.)
78 -o, --output filename
80 output the report or error message to this specified file
81 --permit-key [=<value>]
83 Set a key value that must match for the server to accept traffic
84 on a connection. If the option is given without a value on the
85 server a key value will be autogenerated and displayed in its
86 initial settings report. The value is required on clients. The
87 value will also be used as part of the transfer id in reports.
88 The option set on the client but not the server will also cause
89 the server to reject the client's traffic. TCP only, no UDP sup‐
90 port. (suggested is to use with -1 or --single-client, -P 1 on
91 the server, and -t on the server which is very short)
92 -p, --port n
94 set server port to listen on/connect to to n (default 5001)
95 --sum-only
97 set the output to sum reports only. Useful for -P at large val‐
98 ues
99 -u, --udp
101 use UDP rather than TCP
102 -w, --window n[kmKM]
104 TCP window size (socket buffer size)
105 -z, --realtime
107 Request real-time scheduler, if supported.
108 -B, --bind host[:port][%dev]
110 bind to host, ip address or multicast address, optional port or
111 device (see NOTES)
112 -C, --compatibility
114 for use with older versions does not sent extra msgs
115 -M, --mss n
117 set TCP maximum segment size (MTU - 40 bytes)
118 -N, --nodelay
120 set TCP no delay, disabling Nagle's Algorithm
121 -v, --version
123 print version information and quit
124 -x, --reportexclude [CDMSV]
126 exclude C(connection) D(data) M(multicast) S(settings) V(server)
127 reports
128 -y, --reportstyle C|c
130 if set to C or c report results as CSV (comma separated values)
131 -Z, --tcp-congestion
133 Set the default congestion-control algorithm to be used for new
134 connections. Platforms must support setsockopt's TCP_CONGESTION.
135 (Notes: See sysctl and tcp_allowed_congestion_control for avail‐
136 able options. May require root privileges.)
137
139 -1, --singleclient
140 set the server to process only one client at a time
141 -b, --bandwidth n[kmgKMG]
143 set target read rate to n bits/sec. TCP only for the server.
144 -s, --server
146 run in server mode
147 --histograms[=binwidth[u],bincount,[lowerci],[upperci]]
149 enable latency histograms for udp packets (-u), for tcp writes
150 (with --trip-times), or for either udp or tcp with --isochronous
151 clients. The binning can be modified. Bin widths (default 1 mil‐
152 lisecond, append u for microseconds,) bincount is total bins
153 (default 1000), ci is confidence interval between 0-100% (de‐
154 fault lower 5%, upper 95%, 3 stdev 99.7%)
155 -B, --bind ip | ip%device
157 bind src ip addr and optional src device for receiving
158 -D, --daemon
160 run the server as a daemon. On Windows this will run the speci‐
161 fied command-line under the IPerfService, installing the service
162 if necessary. Note the service is not configured to auto-start
163 or restart - if you need a self-starting service you will need
164 to create an init script or use Windows "sc" commands.
165 -H, --ssm-host host
167 Set the source host (ip addr) per SSM multicast, i.e. the S of
168 the S,G
169 -R, --remove
171 remove the IPerfService (Windows only).
172 -U, --single_udp
174 run in single threaded UDP mode
175 -V, --ipv6_domain
177 Enable IPv6 reception by setting the domain and socket to
178 AF_INET6 (Can receive on both IPv4 and IPv6)
179
181 -b, --bandwidth n[kmgKMG][,n[kmgKMG]] | n[kmgKMG]pps
182 set target bandwidth to n bits/sec (default 1 Mbit/sec) or n
183 packets per sec. This may be used with TCP or UDP. Optionally,
184 for variable loads, use format of mean,standard deviation
185 -c, --client host | host%device
187 run in client mode, connecting to host where the optional %dev
188 will SO_BINDTODEVICE that output interface (requires root and
189 see NOTES)
190 --connect-only[=n]
192 only perform a TCP connect (or 3WHS) without any data transfer,
193 useful to measure TCP connect() times. Optional value of n is
194 the total number of connects to do (zero is run forever.) Note
195 that -i will rate limit the connects where -P will create bursts
196 and -t will end the client and hence end its connect attempts.
197 --connect-retries n]
199 number of times to retry a TCP connect at the application level.
200 See operating system information on the details of TCP connect
201 related settings.
202 -d, --dualtest
204 Do a bidirectional test simultaneous test using two unidirec‐
205 tional sockets
206 --fq-rate n[kmgKMG]
208 Set a rate to be used with fair-queueing based socket-level pac‐
209 ing, in bytes or bits per second. Only available on platforms
210 supporting the SO_MAX_PACING_RATE socket option. (Note: Here the
211 suffixes indicate bytes/sec or bits/sec per use of uppercase or
212 lowercase, respectively)
213 --full-duplex
215 run a full duplex test, i.e. traffic in both transmit and re‐
216 ceive directions using the same socket
217 --incr-dstip
219 increment the destination ip address when using the parallel
220 (-P) option
221 --ipg n
223 set the inter-packet gap to n (units of seconds) for packets or
224 within a frame/burst when --isochronous is set
225 --isochronous[=fps:mean,stdev]
227 send isochronous traffic with frequency frames per second and
228 load defined by mean and standard deviation using a log normal
229 distribution, defaults to 60:20m,0. (Note: Here the suffixes in‐
230 dicate bytes/sec or bits/sec per use of uppercase or lowercase,
231 respectively. Also the p suffix is supported to set the burst
232 size in packets, e.g. isochronous=2:25p will send two 25 packet
233 bursts every second, or one 25 packet burst every 0.5 seconds.)
234 --local-only[=1|0]
236 Set 1 to limit traffic to the local network only (through the
237 use of SO_DONTROUTE) set to zero otherwise with optional over‐
238 ride of compile time default (see configure --default-localonly)
239 --near-congestion[=n]
241 Enable TCP write rate limiting per the sampled RTT. The delay is
242 applied after the -l number of bytes have completed. The op‐
243 tional value is the multiplier to the RTT and defines the time
244 delay. This value defaults to 0.5 if it is not set. Values less
245 than 1 are supported but the value cannot be negative. This is
246 an experimental feature. It is not likely stable on live net‐
247 works. Suggested use is over controlled test networks.
248 --no-connect-sync
251 By default, parallel traffic threads (per -P greater than 1)
252 will synchronize after their TCP connects and prior to each
253 sending traffic, i.e. all the threads first complete (or error)
254 the TCP 3WHS before any traffic thread will start sending. This
255 option disables that synchronization such that each traffic
256 thread will start sending immediately after completing its suc‐
257 cessful connect.
258 --no-udp-fin
260 Don't perform the UDP final server to client exchange which
261 means there won't be a final server report displayed on the
262 client. All packets per the test will be from the client to the
263 server and no packets should be sent in the other direction.
264 It's highly suggested that -t be set on the server if this op‐
265 tion is being used. This is because there will be only one
266 trigger ending packet sent from client to server and if it's
267 lost then the server will continue to run. (Requires ver 2.0.14
268 or better)
269 -n, --num n[kmKM]
271 number of bytes to transmit (instead of -t)
272 --permit-key-timeout <value>
274 Set the lifetime of the permit key in seconds. Defaults to 20
275 seconds if not set. A value of zero will disable the timer.
276 -r, --tradeoff
278 Do a bidirectional test individually - client-to-server, fol‐
279 lowed by a reversed test, server-to-client
280 -t, --time n
282 time in seconds to listen for new traffic connections, receive
283 traffic or transmit traffic (Defaults: transmit is 10 secs while
284 listen and receive are indefinite)
285 --trip-times
287 enable the measurement of end to end write to read latencies
288 (client and server clocks must be synchronized)
289 --txdelay-time
291 time in seconds to hold back or delay after the TCP connect and
292 prior to the socket writes. For UDP it's the delay between the
293 traffic thread starting and the first write.
294 --txstart-time n.n
296 set the txstart-time to n.n using unix or epoch time format
297 (supports microsecond resolution, e.g 1536014418.123456) An ex‐
298 ample to delay one second using command substitution is iperf -c
299 192.168.1.10 --txstart-time $(expr $(date +%s) + 1).$(date +%N)
300 -B, --bind ip | ip:port | ipv6 -V | [ipv6]:port -V
302 bind src ip addr and optional port as the source of traffic (see
303 notes)
304 -F, --fileinput name
306 input the data to be transmitted from a file
307 -I, --stdin
309 input the data to be transmitted from stdin
310 -L, --listenport n
312 port to receive bidirectional tests back on
313 -P, --parallel n
315 number of parallel client threads to run
316 -R, --reverse
318 reverse the traffic flow (useful for testing through firewalls,
319 see NOTES)
320 -S, --tos
322 set the socket's IP_TOS (byte) field
323 -T, --ttl n
325 time-to-live, for multicast (default 1) -V, --ipv6_domain Set
326 the domain to IPv6 (send packets over IPv6)
327 -X, --peerdetect
329 run peer version detection prior to traffic.
330 -Z, --linux-congestion algo
332 set TCP congestion control algorithm (Linux only)
333
335 TCP tests (client)
336 iperf -c <host> -e -i 1
338 ------------------------------------------------------------
339 Client connecting to <host>, TCP port 5001 with pid 5149
340 Write buffer size: 128 KByte
341 TCP window size: 340 KByte (default)
342 ------------------------------------------------------------
343 [ 3] local 45.56.85.133 port 49960 connected with 45.33.58.123 port
344 5001 (ct=3.23 ms)
345 [ ID] Interval Transfer Bandwidth Write/Err Rtry
346 Cwnd/RTT NetPwr
347 [ 3] 0.00-1.00 sec 126 MBytes 1.05 Gbits/sec 1006/0 0
348 56K/626 us 210636.47
349 [ 3] 1.00-2.00 sec 138 MBytes 1.15 Gbits/sec 1100/0 299
350 483K/3884 us 37121.32
351 [ 3] 2.00-3.00 sec 137 MBytes 1.15 Gbits/sec 1093/0 24
352 657K/5087 us 28162.31
353 [ 3] 3.00-4.00 sec 126 MBytes 1.06 Gbits/sec 1010/0 284
354 294K/2528 us 52366.58
355 [ 3] 4.00-5.00 sec 117 MBytes 980 Mbits/sec 935/0 373
356 487K/2025 us 60519.66
357 [ 3] 5.00-6.00 sec 144 MBytes 1.20 Gbits/sec 1149/0 2
358 644K/3570 us 42185.36
359 [ 3] 6.00-7.00 sec 126 MBytes 1.06 Gbits/sec 1011/0 112
360 582K/5281 us 25092.56
361 [ 3] 7.00-8.00 sec 110 MBytes 922 Mbits/sec 879/0 56
362 279K/1957 us 58871.89
363 [ 3] 8.00-9.00 sec 127 MBytes 1.06 Gbits/sec 1014/0 46
364 483K/3372 us 39414.89
365 [ 3] 9.00-10.00 sec 132 MBytes 1.11 Gbits/sec 1054/0 0
366 654K/3380 us 40872.75
367 [ 3] 0.00-10.00 sec 1.25 GBytes 1.07 Gbits/sec 10251/0 1196
368 -1K/3170 us 42382.03
369 where (per -e,)
372 ct= TCP connect time (or three way handshake time 3WHS)
373 Write/Err Total number of successful socket writes. Total number
374 of non-fatal socket write errors
375 Rtry Total number of TCP retries
376 Cwnd/RTT (*nix only) TCP congestion window and round trip time
377 (sampled where NA indicates no value)
378 NetPwr (*nix only) Network power defined as (throughput / RTT)
379 TCP tests (server)
382 iperf -s -e -i 1 -l 8K
384 ------------------------------------------------------------
385 Server listening on TCP port 5001 with pid 13430
386 Read buffer size: 8.00 KByte
387 TCP window size: 85.3 KByte (default)
388 ------------------------------------------------------------
389 [ 4] local 45.33.58.123 port 5001 connected with 45.56.85.133 port
390 49960
391 [ ID] Interval Transfer Bandwidth Reads
392 Dist(bin=1.0K)
393 [ 4] 0.00-1.00 sec 124 MBytes 1.04 Gbits/sec 22249
394 798:2637:2061:767:2165:1563:589:11669
395 [ 4] 1.00-2.00 sec 136 MBytes 1.14 Gbits/sec 24780
396 946:3227:2227:790:2427:1888:641:12634
397 [ 4] 2.00-3.00 sec 137 MBytes 1.15 Gbits/sec 24484
398 1047:2686:2218:810:2195:1819:728:12981
399 [ 4] 3.00-4.00 sec 126 MBytes 1.06 Gbits/sec 20812
400 863:1353:1546:614:1712:1298:547:12879
401 [ 4] 4.00-5.00 sec 117 MBytes 984 Mbits/sec 20266
402 769:1886:1828:589:1866:1350:476:11502
403 [ 4] 5.00-6.00 sec 143 MBytes 1.20 Gbits/sec 24603
404 1066:1925:2139:822:2237:1827:744:13843
405 [ 4] 6.00-7.00 sec 126 MBytes 1.06 Gbits/sec 22635
406 834:2464:2249:724:2269:1646:608:11841
407 [ 4] 7.00-8.00 sec 110 MBytes 921 Mbits/sec 21107
408 842:2437:2747:592:2871:1903:496:9219
409 [ 4] 8.00-9.00 sec 126 MBytes 1.06 Gbits/sec 22804
410 1038:1784:2639:656:2738:1927:573:11449
411 [ 4] 9.00-10.00 sec 133 MBytes 1.11 Gbits/sec 23091
412 1088:1654:2105:710:2333:1928:723:12550
413 [ 4] 0.00-10.02 sec 1.25 GBytes 1.07 Gbits/sec 227306
414 9316:22088:21792:7096:22893:17193:6138:120790
415 where (per -e,)
417 Reads Total number of socket reads
418 Dist(bin=size) Eight bin histogram of the socket reads returned
419 byte count. Bin width is set per size. Bins are separated by a
420 colon. In the example, the bins are 0-1K, 1K-2K, .., 7K-8K.
421 TCP tests (server with --trip-times on client) iperf -s -i 1 -w 4M
424 ------------------------------------------------------------
425 Server listening on TCP port 5001
426 TCP window size: 8.00 MByte (WARNING: requested 4.00 MByte)
427 ------------------------------------------------------------
428 [ 4] local 192.168.1.4%eth0 port 5001 connected with 192.168.1.7 port
429 44798 (trip-times) (MSS=1448) (peer 2.0.14-alpha)
430 [ ID] Interval Transfer Bandwidth Burst Latency
431 avg/min/max/stdev (cnt/size) inP NetPwr Reads=Dist
432 [ 4] 0.00-1.00 sec 19.0 MBytes 159 Mbits/sec
433 52.314/10.238/117.155/19.779 ms (151/131717) 1.05 MByte 380.19
434 781=306:253:129:48:18:15:8:4
435 [ 4] 1.00-2.00 sec 20.0 MBytes 168 Mbits/sec
436 53.863/21.264/79.252/12.277 ms (160/131080) 1.08 MByte 389.38
437 771=294:236:126:60:18:24:10:3
438 [ 4] 2.00-3.00 sec 18.2 MBytes 153 Mbits/sec
439 58.718/22.000/137.944/20.397 ms (146/130964) 1.06 MByte 325.64
440 732=299:231:98:52:18:19:10:5
441 [ 4] 3.00-4.00 sec 19.7 MBytes 165 Mbits/sec 50.448/
442 8.921/82.728/14.627 ms (158/130588) 997 KByte 409.00
443 780=300:255:121:58:15:18:7:6
444 [ 4] 4.00-5.00 sec 18.8 MBytes 158 Mbits/sec
445 53.826/11.169/115.316/15.541 ms (150/131420) 1.02 MByte 366.24
446 761=302:226:134:52:22:17:7:1
447 [ 4] 5.00-6.00 sec 19.5 MBytes 164 Mbits/sec
448 50.943/11.922/76.134/14.053 ms (156/131276) 1.03 MByte 402.00
449 759=273:246:149:45:16:18:4:8
450 [ 4] 6.00-7.00 sec 18.5 MBytes 155 Mbits/sec
451 57.643/10.039/127.850/18.950 ms (148/130926) 1.05 MByte 336.16
452 710=262:228:133:37:16:20:8:6
453 [ 4] 7.00-8.00 sec 19.6 MBytes 165 Mbits/sec
454 52.498/12.900/77.045/12.979 ms (157/131003) 1.00 MByte 391.78
455 742=288:200:135:68:16:23:4:8
456 [ 4] 8.00-9.00 sec 18.0 MBytes 151 Mbits/sec 58.370/
457 8.026/150.243/21.445 ms (144/131255) 1.06 MByte 323.81
458 716=268:241:108:51:20:17:8:3
459 [ 4] 9.00-10.00 sec 18.4 MBytes 154 Mbits/sec
460 56.112/12.419/79.790/13.668 ms (147/131194) 1.05 MByte 343.70
461 822=330:303:120:26:16:14:9:4
462 [ 4] 10.00-10.06 sec 1.03 MBytes 146 Mbits/sec
463 69.880/45.175/78.754/10.823 ms (9/119632) 1.74 MByte 260.40
464 62=26:30:5:1:0:0:0:0
465 [ 4] 0.00-10.06 sec 191 MBytes 159 Mbits/sec 54.183/
466 8.026/150.243/16.781 ms (1526/131072) 1.03 MByte 366.98
467 7636=2948:2449:1258:498:175:185:75:48
468 where (per -e,)
470 Burst Latency One way TCP write() to read() latency in mean/min‐
471 imum/maximum/standard deviation format (Note: requires the
472 client's and server's system clocks to be synchronized to a com‐
473 mon reference, e.g. using precision time protocol PTP. A GPS
474 disciplined OCXO is a recommended reference.)
475 cnt Number of completed bursts received and used for the burst
476 latency calculations
477 size Average burst size in bytes (computed average and estimate
478 only)
479 inP inP, short for in progress, is the average number of bytes
480 in progress or in flight. This is taken from the application
481 level write to read perspective. Note this is a mean value. The
482 parenthesis value is the standard deviation from the mean. (Re‐
483 quires --trip-times on client. See Little's law in NOTES.)
484 NetPwr Network power defined as (throughput / one way latency)
485 UDP tests (client)
488 iperf -c <host> -e -i 1 -u -b 10m
490 ------------------------------------------------------------
491 Client connecting to <host>, UDP port 5001 with pid 5169
492 Sending 1470 byte datagrams, IPG target: 1176.00 us (kalman adjust)
493 UDP buffer size: 208 KByte (default)
494 ------------------------------------------------------------
495 [ 3] local 45.56.85.133 port 32943 connected with 45.33.58.123 port
496 5001
497 [ ID] Interval Transfer Bandwidth Write/Err PPS
498 [ 3] 0.00-1.00 sec 1.19 MBytes 10.0 Mbits/sec 852/0 851 pps
499 [ 3] 1.00-2.00 sec 1.19 MBytes 10.0 Mbits/sec 850/0 850 pps
500 [ 3] 2.00-3.00 sec 1.19 MBytes 10.0 Mbits/sec 850/0 850 pps
501 [ 3] 3.00-4.00 sec 1.19 MBytes 10.0 Mbits/sec 851/0 850 pps
502 [ 3] 4.00-5.00 sec 1.19 MBytes 10.0 Mbits/sec 850/0 850 pps
503 [ 3] 5.00-6.00 sec 1.19 MBytes 10.0 Mbits/sec 850/0 850 pps
504 [ 3] 6.00-7.00 sec 1.19 MBytes 10.0 Mbits/sec 851/0 850 pps
505 [ 3] 7.00-8.00 sec 1.19 MBytes 10.0 Mbits/sec 850/0 850 pps
506 [ 3] 8.00-9.00 sec 1.19 MBytes 10.0 Mbits/sec 851/0 850 pps
507 [ 3] 0.00-10.00 sec 11.9 MBytes 10.0 Mbits/sec 8504/0 850 pps
508 [ 3] Sent 8504 datagrams
509 [ 3] Server Report:
510 [ 3] 0.00-10.00 sec 11.9 MBytes 10.0 Mbits/sec 0.047 ms 0/ 8504
511 (0%) 0.537/ 0.392/23.657/ 0.497 ms 850 pps 2329.37
512 where (per -e,)
514 Write/Err Total number of successful socket writes. Total number
515 of non-fatal socket write errors
516 PPS Transmit packet rate in packets per second
517 UDP tests (server) iperf -s -i 1 -w 4M -u
520 ------------------------------------------------------------
521 Server listening on UDP port 5001
522 Receiving 1470 byte datagrams
523 UDP buffer size: 8.00 MByte (WARNING: requested 4.00 MByte)
524 ------------------------------------------------------------
525 [ 3] local 192.168.1.4 port 5001 connected with 192.168.1.1 port 60027
526 (WARN: winsize=8.00 MByte req=4.00 MByte) (trip-times) (0.0) (peer
527 2.0.14-alpha)
528 [ ID] Interval Transfer Bandwidth Jitter Lost/Total
529 Latency avg/min/max/stdev PPS inP NetPwr
530 [ 3] 0.00-1.00 sec 44.5 MBytes 373 Mbits/sec 0.071 ms 52198/83938
531 (62%) 75.185/ 2.367/85.189/14.430 ms 31854 pps 3.64 MByte 620.58
532 [ 3] 1.00-2.00 sec 44.8 MBytes 376 Mbits/sec 0.015 ms
533 59549/143701 (41%) 79.609/75.603/85.757/ 1.454 ms 31954 pps 3.56 MByte
534 590.04
535 [ 3] 2.00-3.00 sec 44.5 MBytes 373 Mbits/sec 0.017 ms
536 59494/202975 (29%) 80.006/75.951/88.198/ 1.638 ms 31733 pps 3.56 MByte
537 583.07
538 [ 3] 3.00-4.00 sec 44.5 MBytes 373 Mbits/sec 0.019 ms
539 59586/262562 (23%) 79.939/75.667/83.857/ 1.145 ms 31767 pps 3.56 MByte
540 583.57
541 [ 3] 4.00-5.00 sec 44.5 MBytes 373 Mbits/sec 0.081 ms
542 59612/322196 (19%) 79.882/75.400/86.618/ 1.666 ms 31755 pps 3.55 MByte
543 584.40
544 [ 3] 5.00-6.00 sec 44.7 MBytes 375 Mbits/sec 0.064 ms
545 59571/381918 (16%) 79.767/75.571/85.339/ 1.556 ms 31879 pps 3.56 MByte
546 588.02
547 [ 3] 6.00-7.00 sec 44.6 MBytes 374 Mbits/sec 0.041 ms
548 58990/440820 (13%) 79.722/75.662/85.938/ 1.087 ms 31820 pps 3.58 MByte
549 586.73
550 [ 3] 7.00-8.00 sec 44.7 MBytes 375 Mbits/sec 0.027 ms
551 59679/500548 (12%) 79.745/75.704/84.731/ 1.094 ms 31869 pps 3.55 MByte
552 587.46
553 [ 3] 8.00-9.00 sec 44.3 MBytes 371 Mbits/sec 0.078 ms
554 59230/559499 (11%) 80.346/75.514/94.293/ 2.858 ms 31590 pps 3.58 MByte
555 577.97
556 [ 3] 9.00-10.00 sec 44.4 MBytes 373 Mbits/sec 0.073 ms
557 58782/618394 (9.5%) 79.125/75.511/93.638/ 1.643 ms 31702 pps 3.55 MByte
558 588.99
559 [ 3] 10.00-10.08 sec 3.53 MBytes 367 Mbits/sec 0.129 ms
560 6026/595236 (1%) 94.967/80.709/99.685/ 3.560 ms 31107 pps 3.58 MByte
561 483.12
562 [ 3] 0.00-10.08 sec 449 MBytes 374 Mbits/sec 0.129 ms
563 592717/913046 (65%) 79.453/ 2.367/99.685/ 5.200 ms 31776 pps (null)
564 587.91
565 where (per -e,)
568 Latency End to end latency in mean/minimum/maximum/standard de‐
569 viation format (Note: requires the client's and server's system
570 clocks to be synchronized to a common reference, e.g. using pre‐
571 cision time protocol PTP. A GPS disciplined OCXO is a recom‐
572 mended reference.)
573 PPS Received packet rate in packets per second
574 inP inP, short for in progress, is the average number of bytes
575 in progress or in flight. This is taken from an application
576 write to read perspective. (Requires --trip-times on client. See
577 Little's law in NOTES.)
578 NetPwr Network power defined as (throughput / latency)
579 Isochronous UDP tests (client)
582 iperf -c 192.168.100.33 -u -e -i 1 --isochronous=60:100m,10m --realtime
584 ------------------------------------------------------------
585 Client connecting to 192.168.100.33, UDP port 5001 with pid 14971
586 UDP isochronous: 60 frames/sec mean= 100 Mbit/s, stddev=10.0 Mbit/s,
587 Period/IPG=16.67/0.005 ms
588 UDP buffer size: 208 KByte (default)
589 ------------------------------------------------------------
590 [ 3] local 192.168.100.76 port 42928 connected with 192.168.100.33
591 port 5001
592 [ ID] Interval Transfer Bandwidth Write/Err PPS
593 frames:tx/missed/slips
594 [ 3] 0.00-1.00 sec 12.0 MBytes 101 Mbits/sec 8615/0 8493 pps
595 62/0/0
596 [ 3] 1.00-2.00 sec 12.0 MBytes 100 Mbits/sec 8556/0 8557 pps
597 60/0/0
598 [ 3] 2.00-3.00 sec 12.0 MBytes 101 Mbits/sec 8586/0 8586 pps
599 60/0/0
600 [ 3] 3.00-4.00 sec 12.1 MBytes 102 Mbits/sec 8687/0 8687 pps
601 60/0/0
602 [ 3] 4.00-5.00 sec 11.8 MBytes 99.2 Mbits/sec 8468/0 8468 pps
603 60/0/0
604 [ 3] 5.00-6.00 sec 11.9 MBytes 99.8 Mbits/sec 8519/0 8520 pps
605 60/0/0
606 [ 3] 6.00-7.00 sec 12.1 MBytes 102 Mbits/sec 8694/0 8694 pps
607 60/0/0
608 [ 3] 7.00-8.00 sec 12.1 MBytes 102 Mbits/sec 8692/0 8692 pps
609 60/0/0
610 [ 3] 8.00-9.00 sec 11.9 MBytes 100 Mbits/sec 8537/0 8537 pps
611 60/0/0
612 [ 3] 9.00-10.00 sec 11.8 MBytes 99.0 Mbits/sec 8450/0 8450 pps
613 60/0/0
614 [ 3] 0.00-10.01 sec 120 MBytes 100 Mbits/sec 85867/0 8574 pps
615 602/0/0
616 [ 3] Sent 85867 datagrams
617 [ 3] Server Report:
618 [ 3] 0.00-9.98 sec 120 MBytes 101 Mbits/sec 0.009 ms 196/85867
619 (0.23%) 0.665/ 0.083/ 1.318/ 0.174 ms 8605 pps 18903.85
620 where (per -e,)
622 frames:tx/missed/slips Total number of isochronous frames or
623 bursts. Total number of frame ids not sent. Total number of
624 frame slips
625 Isochronous UDP tests (server)
628 iperf -s -e -u --udp-histogram=100u,2000 --realtime
630 ------------------------------------------------------------
631 Server listening on UDP port 5001 with pid 5175
632 Receiving 1470 byte datagrams
633 UDP buffer size: 208 KByte (default)
634 ------------------------------------------------------------
635 [ 3] local 192.168.100.33 port 5001 connected with 192.168.100.76 port
636 42928 isoch (peer 2.0.13-alpha)
637 [ ID] Interval Transfer Bandwidth Jitter Lost/Total
638 Latency avg/min/max/stdev PPS NetPwr Frames/Lost
639 [ 3] 0.00-9.98 sec 120 MBytes 101 Mbits/sec 0.010 ms 196/85867
640 (0.23%) 0.665/ 0.083/ 1.318/ 0.284 ms 8585 pps 18903.85 601/1
641 [ 3] 0.00-9.98 sec T8(f)-PDF:
642 bin(w=100us):cnt(85671)=1:2,2:844,3:10034,4:8493,5:8967,6:8733,7:8823,8:9023,9:8901,10:8816,11:7730,12:4563,13:741,14:1
643 (5.00/95.00%=3/12,Outliers=0,obl/obu=0/0)
644 [ 3] 0.00-9.98 sec F8(f)-PDF:
645 bin(w=100us):cnt(598)=15:2,16:1,17:27,18:68,19:125,20:136,21:103,22:83,23:22,24:23,25:5,26:3
646 (5.00/95.00%=17/24,Outliers=0,obl/obu=0/0)
647 where, Frames/lost Total number of frames (or bursts) received. Total
650 number of bursts lost or error-ed
651 T8-PDF(f) Latency histogram for packets
652 F8-PDF(f) Latency histogram for frames
653
657 Note: The environment variable option settings haven't been maintained
658 well. See the source code if these are of interest.
659
661 Numeric options: Some numeric options support format characters per
662 '<value>c' (e.g. 10M) where the c format characters are k,m,g,K,M,G.
663 Lowercase format characters are 10^3 based and uppercase are 2^n based,
664 e.g. 1k = 1000, 1K = 1024, 1m = 1,000,000 and 1M = 1,048,576
665 Rate limiting: The -b option supports read and write rate limiting at
667 the application level. The -b option on the client also supports vari‐
668 able offered loads through the <mean>,<standard deviation> format, e.g.
669 -b 100m,10m. The distribution used is log normal. Similar for the
670 isochronous option. The -b on the server rate limits the reads. Socket
671 based pacing is also supported using the --fq-rate long option. This
672 will work with the --reverse and --full-duplex options as well.
673 Synchronized clocks: The --trip-times option indicates that the
675 client's and server's clocks are synchronized to a common reference.
676 Network Time Protocol (NTP) or Precision Time Protocol (PTP) are com‐
677 monly used for this. The reference clock(s) error and the synchroniza‐
678 tion protocols will affect the accuracy of any end to end latency mea‐
679 surements.
680 Binding is done at the logical level (ip address or layer 3) using the
682 -B option and at the device (or layer 2) level using the percent (%)
683 separator for both the client and the server. On the client, the -B op‐
684 tion affects the bind(2) system call, and will set the source ip ad‐
685 dress and the source port, e.g. iperf -c <host> -B 192.168.100.2:6002.
686 This controls the packet's source values but not routing. These can be
687 confusing in that a route or device lookup may not be that of the de‐
688 vice with the configured source IP. So, for example, if the IP address
689 of eth0 is used for -B and the routing table for the destination IP ad‐
690 dress resolves the output interface to be eth1, then the host will send
691 the packet out device eth1 while using the source IP address of eth0 in
692 the packet. To affect the physical output interface (e.g. dual homed
693 systems) either use -c <host>%<dev> (requires root) which bypasses this
694 host route table lookup, or configure policy routing per each -B source
695 address and set the output interface appropriately in the policy
696 routes. On the server or receive, only packets destined to -B IP ad‐
697 dress will be received. It's also useful for multicast. For example,
698 iperf -s -B 224.0.0.1%eth0 will only accept ip multicast packets with
699 dest ip 224.0.0.1 that are received on the eth0 interface, while iperf
700 -s -B 224.0.0.1 will receive those packets on any interface, Finally,
701 the device specifier is required for v6 link-local, e.g. -c
702 [v6addr]%<dev> -V, to select the output interface.
703 Reverse, full-duplex, dualtest (-d) and tradeoff (-r): The --reverse
705 (-R) and --full-duplex options can be confusing when compared to the
706 older options of --dualtest (-d) and --tradeoff (-r). The newer options
707 of --reverse and --full-duplex only open one socket and read and write
708 to the same socket descriptor, i.e. use the socket in full duplex mode.
709 The older -d and -r open second sockets in the opposite direction and
710 do not use a socket in full duplex mode. Note that full duplex applies
711 to the socket and not to the network devices and that full duplex sock‐
712 ets are supported by the operating systems regardless if an underlying
713 network supports full duplex transmission and reception. It's sug‐
714 gested to use --reverse if you want to test through a NAT firewall (or
715 -R on non-windows systems). This applies role reversal of the test af‐
716 ter opening the full duplex socket. (Note: Firewall piercing may be
717 required to use -d and -r if a NAT gateway is in the path.)
718 Also, the --reverse -b <rate> setting behaves differently for TCP and
720 UDP. For TCP it will rate limit the read side, i.e. the iperf client
721 (role reversed to act as a server) reading from the full duplex socket.
722 This will in turn flow control the reverse traffic per standard TCP
723 congestion control. The --reverse -b <rate> will be applied on transmit
724 (i.e. the server role reversed to act as a client) for UDP since there
725 is no flow control with UDP. There is no option to directly rate limit
726 the writes with TCP testing when using --reverse.
727 TCP Connect times: The TCP connect time (or three way handshake) can be
729 seen on the iperf client when the -e (--enhanced) option is set. Look
730 for the ct=<value> in the connected message, e.g.in '[ 3] local
731 192.168.1.4 port 48736 connected with 192.168.1.1 port 5001 (ct=1.84
732 ms)' shows the 3WHS took 1.84 milliseconds.
733 Packet per second (pps) calculation The packets per second calculation
735 is done as a derivative, i.e. number of packets divided by time. The
736 time is taken from the previous last packet to the current last packet.
737 It is not the sample interval time. The last packet can land at differ‐
738 ent times within an interval. This means that pps does not have to
739 match rx bytes divided by the sample interval. Also, with --trip-times
740 set, the packet time on receive is set by the sender's write time so
741 pps indicates the end to end pps with --trip-times. The RX pps calcula‐
742 tion is receive side only when -e is set and --trip-times is not set.
743 Little's Law in queuing theory is a theorem that determines the average
745 number of items (L) in a stationary queuing system based on the average
746 waiting time (W) of an item within a system and the average number of
747 items arriving at the system per unit of time (lambda). Mathematically,
748 it's L = lambda * W. As used here, the units are bytes. The arrival
749 rate is taken from the writes.
750 Network power: The network power (NetPwr) metric is experimental. It's
752 a convenience function defined as throughput/delay. For TCP transmits,
753 the delay is the sampled RTT times. For TCP receives, the delay is the
754 write to read latency. For UDP the delay is the end/end latency.
755 Don't confuse this with the physics definition of power (delta en‐
756 ergy/delta time) but more of a measure of a desirable property divided
757 by an undesirable property. Also note, one must use -i interval with
758 TCP to get this as that's what sets the RTT sampling rate. The metric
759 is scaled to assist with human readability.
760 Multicast: Iperf 2 supports multicast with a couple of caveats. First,
762 multicast streams cannot take advantage of the -P option. The server
763 will serialize multicast streams. Also, it's highly encouraged to use a
764 -t on a server that will be used for multicast clients. That is because
765 the single end of traffic packet sent from client to server may get
766 lost and there are no redundant end of traffic packets. Setting -t on
767 the server will kill the server thread in the event this packet is in‐
768 deed lost.
769 Fast Sampling: Use ./configure --enable-fastsampling and then compile
771 from source to enable four digit (e.g. 1.0000) precision in reports'
772 timestamps. Useful for sub-millisecond sampling.
773
775 Use ./configure --enable-thread-debug and then compile from source to
776 enable both asserts and advanced debugging of the tool itself.
777
779 See https://sourceforge.net/p/iperf2/tickets/
780
782 Iperf2, based from iperf (originally written by Mark Gates and Alex
783 Warshavsky), has a goal of maintenance with some feature enhancement.
784 Other contributions from Ajay Tirumala, Jim Ferguson, Jon Dugan <jdugan
785 at x1024 dot net>, Feng Qin, Kevin Gibbs, John Estabrook <jestabro at
786 ncsa.uiuc.edu>, Andrew Gallatin <gallatin at gmail.com>, Stephen Hem‐
787 minger <shemminger at linux-foundation.org>, Tim Auckland <tim.auckland
788 at gmail.com>, Robert J. McMahon <rjmcmahon at rjmcmahon.com>
789
791 accept(2),bind(2),close(2),connect(2),fcntl(2),getpeername(2),getsock‐
792 name(2),getsockopt(2),listen(2),read(2),recv(2),select(2),send(2),set‐
793 sockopt(2),shutdown(2),write(2),ip(7),socket(7),tcp(7),udp(7)
794 Source code at http://sourceforge.net/projects/iperf2/
796 "Unix Network Programming, Volume 1: The Sockets Networking API (3rd
798 Edition) 3rd Edition" by W. Richard Stevens (Author), Bill Fenner (Au‐
799 thor), Andrew M. Rudoff (Author)
800NLANR/DAST January 2021 IPERF(1)