1 .\" netsniff-ng - the packet sniffing beast
2 .\" Copyright 2013 Daniel Borkmann.
3 .\" Subject to the GPL, version 2.
4 .TH IFPPS 8 "03 March 2013" "Linux" "netsniff-ng toolkit"
6 ifpps \- top-like networking and system statistics
10 \fBifpps\fR { [\fIoptions\fR] | [\fIdevice\fR] }
14 ifpps is a small utility which periodically provides top-like networking
15 and system statistics from the kernel. ifpps gathers its data directly
16 from procfs files and does not make use of any user space monitoring
17 libraries which would falsify statistics under high load.
19 For instance, consider the following scenario: two directly connected
20 Linux machines with Intel Core 2 Quad Q6600 2.40GHz CPUs, 4 GB RAM, and
21 an Intel 82566DC-2 Gigabit Ethernet NIC are used for performance evaluation.
22 One machine generates 64 byte network packets by using the kernel space
23 packet generator pktgen with a maximum possible packet rate. The other
24 machine displays statistics about incoming network packets by using i)
28 iptraf which incorporates
30 shows an average packet rate of 246,000 pps while on the other hand ifpps shows
31 an average packet rate of 1,378,000 pps. Hence, due to packet copies and
32 deferring statistics creation into user space, a measurement error of
33 approximately 460 percent occurs. Tools like iptraf might display much more
34 information such as TCP per flow statistics (hence the use of the pcap library).
35 This is not possible with ifpps, because overall networking statistics are its
36 focus; statistics, which are also fairly reliable under high packet load.
38 ifpps also periodically displays CPU load, interrupt, software interrupt
39 data per sample interval as well as total interrupts, all per CPU. In case
40 the number of CPUs exceeds 5 or the number specified by the user with the
41 \fB-n\fP command line option, ifpps will only display this number top
42 heavy hitters. The topmost heavy hitter CPU will be marked with \[lq]+\[rq].
43 The least heavy hitter will always be displayed and is marked with
44 \[lq]-\[rq]. In addition, the average for all the above per-CPU data is
45 shown. Optionally the median values can be displayed using the \fB-m\fP
48 ifpps also supports directly the
50 data sample format. This facilitates creation of gnuplot figures from ifpps time
55 .B -d <netdev>, --dev <netdev>
56 Networking device to fetch statistics from, for example eth0, wlan0.
59 Set maximum number of top hitter CPUs (in terms of time spent in system/user
60 mode) to display in ncurses mode, default is 10.
62 .B -t <time>, --interval <time>
63 Statistics refresh interval in milliseconds, default is 1000ms.
66 Output (once) the ncurses data to the terminal as
71 Continuously output the terminal data after a refresh interval. This option
72 is only available if option \fB-c\fP is given. For \fB-l\fP it is
73 usually recommended to redirect the output into a file that is to be processed
78 Show median values across all CPUs for CPU load, interrupts (per interval and
79 absolute) and software interrupts.
82 Omit printing the CSV header. This option is only available if \fB-c\fP is given.
85 Turn on promiscuous mode for the given networking device.
88 Show percentage of current throughput in relation to theoretical line rate.
91 Suppress possible warnings in the ncurses output, e.g. about a too low sampling
92 interval that could cause performance regression.
95 Show version information.
103 Default ncurses output for the eth0 device.
106 Ncurses output for the eth0 device in promiscuous mode.
108 .B ifpps -lpcd wlan0 > plot.dat
109 Continuous terminal output for the wlan0 device in promiscuous mode.
112 On 10Gbit/s cards or higher, receive and transmit statistics are usually
113 accumulated at a higher duration interval than 1 second. Thus, it might
114 be advisable to alter the timing to a higher accumulation interval for such cards.
117 ifpps is licensed under the GNU GPL version 2.0.
121 was originally written for the netsniff-ng toolkit by Daniel Borkmann. It
122 is currently maintained by Tobias Klauser <tklauser@distanz.ch> and Daniel
123 Borkmann <dborkma@tik.ee.ethz.ch>.
131 .BR astraceroute (8),
135 Manpage was written by Daniel Borkmann.
138 This page is part of the Linux netsniff-ng toolkit project. A description of the project,
139 and information about reporting bugs, can be found at http://netsniff-ng.org/.