1 .TH NETEM 8 "25 November 2011" "iproute2" "Linux"
3 NetEm \- Network Emulator
10 .IR OPTIONS " := [ " LIMIT " ] [ " DELAY " ] [ " LOSS \
11 " ] [ " CORRUPT " ] [ " DUPLICATION " ] [ " REORDERING " ][ " RATE " ]"
19 .IR TIME " [ " JITTER " [ " CORRELATION " ]]]"
22 .BR distribution " { "uniform " | " normal " | " pareto " | " paretonormal " } ]"
27 .IR PERCENT " [ " CORRELATION " ] |"
30 .IR p13 " [ " p31 " [ " p32 " [ " p23 " [ " p14 "]]]] |"
33 .IR p " [ " r " [ " 1-h " [ " 1-k " ]]]"
38 .IR PERCENT " [ " CORRELATION " ]]"
40 .IR DUPLICATION " := "
42 .IR PERCENT " [ " CORRELATION " ]]"
46 .IR PERCENT " [ " CORRELATION " ] [ "
52 .IR RATE " [ " PACKETOVERHEAD " [ " CELLSIZE " [ " CELLOVERHEAD " ]]]]"
56 NetEm is an enhancement of the Linux traffic control facilities
57 that allow to add delay, packet loss, duplication and more other
58 characteristics to packets outgoing from a selected network
59 interface. NetEm is built using the existing Quality Of Service (QOS)
60 and Differentiated Services (diffserv) facilities in the Linux
64 netem has the following options:
68 limits the effect of selected options to the indicated number of next packets.
71 adds the chosen delay to the packets outgoing to chosen network interface. The
72 optional parameters allows to introduce a delay variation and a correlation.
73 Delay and jitter values are expressed in ms while correlation is percentage.
76 allow the user to choose the delay distribution. If not specified, the default
77 distribution is Normal. Additional parameters allow to consider situations in
78 which network has variable delays depending on traffic flows concurring on the
79 same path, that causes several delay peaks and a tail.
82 adds an independent loss probability to the packets outgoing from the chosen
83 network interface. It is also possible to add a correlation, but this option
84 is now deprecated due to the noticed bad behavior.
87 adds packet losses according to the 4-state Markov using the transition
88 probabilities as input parameters. The parameter p13 is mandatory and if used
89 alone corresponds to the Bernoulli model. The optional parameters allows to
90 extend the model to 2-state (p31), 3-state (p23 and p32) and 4-state (p14).
91 State 1 corresponds to good reception, State 4 to independent losses, State 3
92 to burst losses and State 2 to good reception within a burst.
95 adds packet losses according to the Gilbert-Elliot loss model or its special
96 cases (Gilbert, Simple Gilbert and Bernoulli). To use the Bernoulli model, the
97 only needed parameter is p while the others will be set to the default
98 values r=1-p, 1-h=1 and 1-k=0. The parameters needed for the Simple Gilbert
99 model are two (p and r), while three parameters (p, r, 1-h) are needed for the
100 Gilbert model and four (p, r, 1-h and 1-k) are needed for the Gilbert-Elliot
101 model. As known, p and r are the transition probabilities between the bad and
102 the good states, 1-h is the loss probability in the bad state and 1-k is the
103 loss probability in the good state.
106 allows the emulation of random noise introducing an error in a random position
107 for a chosen percent of packets. It is also possible to add a correlation
108 through the proper parameter.
111 using this option the chosen percent of packets is duplicated before queuing
112 them. It is also possible to add a correlation through the proper parameter.
115 there are two ways to use this option:
120 in this first example every 5th (10th, 15th) packet is sent immediately while
121 other packets are delayed by 10 ms
126 in this second example 25% of packets are sent immediately (with correlation of
127 50%) while the other are delayed by 10 ms.
130 delay packets based on packet size and is a replacement for
133 specified in common units (e.g. 100kbit). Optional
135 (in bytes) specify an per packet overhead and can be negative. A positive value can be
136 used to simulate additional link layer headers. A negative value can be used to
137 artificial strip the Ethernet header (e.g. -14) and/or simulate a link layer
138 header compression scheme. The third parameter - an unsigned value - specify
139 the cellsize. Cellsize can be used to simulate link layer schemes. ATM for
140 example has an payload cellsize of 48 bytes and 5 byte per cell header. If a
141 packet is 50 byte then ATM must use two cells: 2 * 48 bytes payload including 2
142 * 5 byte header, thus consume 106 byte on the wire. The last optional value
144 can be used to specify per cell overhead - for our ATM example 5.
146 can be negative, but use negative values with caution.
148 Note that rate throttling is limited by several factors: the kernel clock
149 granularity avoid a perfect shaping at a specific level. This will show up in
150 an artificial packet compression (bursts). Another influence factor are network
151 adapter buffers which can also add artificial delay.
154 The main known limitation of Netem are related to timer granularity, since
155 Linux is not a real-time operating system.
159 tc qdisc add dev eth0 root netem rate 5kbit 20 100 5
161 delay all outgoing packets on device eth0 with a rate of 5kbit, a per packet
162 overhead of 20 byte, a cellsize of 100 byte and a per celloverhead of 5 byte:
167 Hemminger S. , "Network Emulation with NetEm", Open Source Development Lab,
169 (http://devresources.linux-foundation.org/shemminger/netem/LCA2005_paper.pdf)
172 Netem page from Linux foundation, (http://www.linuxfoundation.org/en/Net:Netem)
175 Salsano S., Ludovici F., Ordine A., "Definition of a general and intuitive loss
176 model for packet networks and its implementation in the Netem module in the
177 Linux kernel", available at http://netgroup.uniroma2.it/NetemCLG
184 Netem was written by Stephen Hemminger at Linux foundation and is based on NISTnet.
185 This manpage was created by Fabio Ludovici <fabio.ludovici at yahoo dot it> and
186 Hagen Paul Pfeifer <hagen@jauu.net>