diff options
| -rw-r--r-- | ifpps.8 | 44 |
1 files changed, 22 insertions, 22 deletions
@@ -15,24 +15,24 @@ ifpps \- top-like networking and system statistics ifpps is a small utility which periodically provides top-like networking and system statistics from the kernel. ifpps gathers its data directly from procfs files and does not apply any user space monitoring libraries -which would falsify statistics on high load. +which could falsify statistics on high load. For instance, consider the following scenario: two directly connected -Linux machines with Intel Core 2 Quad Q6600 2.40GHz CPUs, 4 GB RAM, and -an Intel 82566DC-2 Gigabit Ethernet NIC are used for performance evaluation. -One machine generates 64 byte network packets by using the kernel space -packet generator pktgen with a maximum possible packet rate. The other -machine displays statistics about incoming network packets by using i) -iptraf(8) and ii) ifpps. - -iptraf that incorporates pcap(3) shows an average packet rate of -246,000 pps while on the other hand ifpps shows an average packet rate of -1,378,000 pps. Hence, due to copying packets and deferring statistics -creation into user space, measurement error of approx. 460 per cent +Linux machines, each with an Intel Core 2 Quad Q6600 2.40GHz CPUs, 4 GB +RAM, and an Intel 82566DC-2 Gigabit Ethernet NIC, are used for performance +evaluation. One machine generates 64 byte network packets by using the +kernel space packet generator, pktgen, with a maximum possible packet rate. +The other machine displays statistics about incoming network packets by +using i) iptraf(8) and ii) ifpps. + +iptraf which incorporates pcap(3) shows an average packet rate of +246,000 pps, while on the other hand, ifpps shows an average packet rate +of 1,378,000 pps. Hence, due to packet copies and deferring statistical +calculations to user space, a measurement error of approx. 460 per cent occurs. Tools like iptraf might display much more information such as TCP per flow statistics (therefore the use of the pcap library), which -is not implemented in ifpps, because overall networking statistics are -in our focus; statistics, which are also fairly reliable under high packet +is not implemented in ifpps, because overall networking statistics is +our focus; statistics, which are also fairly reliable under high packet load. .SH OPTIONS @@ -41,7 +41,7 @@ load. Networking device to fetch statistics from, e.g. eth0, wlan0. .SS -t <time>, --interval <time> -Statistics refresh interval in milliseconds, default is 1000 ms. +Statistics refresh interval in milliseconds, default is 1000ms(1 sec). .SS -p, --promisc Turn on promiscuous mode for the given networking device. @@ -51,12 +51,12 @@ Output (once) the ncurses data to the terminal as gnuplot(1)-ready data. .SS -l, --loop Continuously output the terminal data after a refresh interval. This option -only is available, if option ``-c'' is given. For ``-l'' it is usually -recommended to redirect the output into a file that is later being processed +is available only if option ``-c'' is given. For ``-l'' it is usually +recommended to redirect the output into a file that will be processed with gnuplot(1). .SS -v, --version -Show versioning information. +Show version information. .SS -h, --help Show user help. @@ -74,12 +74,12 @@ Continuous terminal output for the wlan0 device in promiscuous mode. .SH NOTE On 10Gbit/s cards or higher, receive and transmit statistics are usually -accumulated each > 1sec. Thus, it might be advised to alter the timing -option to a higher accumulation interval for such cards. +accumulated at a higher duration interval than 1 second. Thus, it might +be advised to alter the timing option to a longer interval for such cards. .SH BUGS -Systems with a failry high number of cores (> 32) are currently not -supported. This should however not be a big deal to fix that. The only +Systems with a fairly high number of cores (> 32) are currently not +supported. This should, however, not be a big deal to fix. The only challenge would be to display the presented information in a sane way, probably by selectively hiding uninteresting statistics. |