The Tcpdump Group git mirrors - tcpdump/blob - README.md

   1 # tcpdump
   2
   3 [![Build Status](https://travis-ci.org/the-tcpdump-group/tcpdump.svg?branch=master)](https://travis-ci.org/the-tcpdump-group/tcpdump)
   4
   5 [![Build Status](https://ci.appveyor.com/api/projects/status/github/the-tcpdump-group/tcpdump?branch=master&svg=true)](https://ci.appveyor.com/project/guyharris/tcpdump)
   6
   7 To report a security issue please send an e-mail to security@tcpdump.org.
   8
   9 To report bugs and other problems, contribute patches, request a
  10 feature, provide generic feedback etc please see the file
  11 CONTRIBUTING in the tcpdump source tree root.
  12
  13 TCPDUMP 4.x.y
  14 Now maintained by "The Tcpdump Group"
  15 See             https://www.tcpdump.org
  16
  17 Anonymous Git is available via:
  18
  19         git clone git://bpf.tcpdump.org/tcpdump
  20
  21 formerly from   Lawrence Berkeley National Laboratory
  22                 Network Research Group <tcpdump@ee.lbl.gov>
  23                 ftp://ftp.ee.lbl.gov/old/tcpdump.tar.Z (3.4)
  24
  25 This directory contains source code for tcpdump, a tool for network
  26 monitoring and data acquisition.  This software was originally
  27 developed by the Network Research Group at the Lawrence Berkeley
  28 National Laboratory.  The original distribution is available via
  29 anonymous ftp to `ftp.ee.lbl.gov`, in `tcpdump.tar.Z`.  More recent
  30 development is performed at tcpdump.org, https://www.tcpdump.org/.
  31
  32 Tcpdump uses libpcap, a system-independent interface for user-level
  33 packet capture.  Before building tcpdump, you must first retrieve and
  34 build libpcap, also originally from LBL and now being maintained by
  35 tcpdump.org; see https://www.tcpdump.org/.
  36
  37 Once libpcap is built (either install it or make sure it's in
  38 `../libpcap`), you can build tcpdump using the procedure in the `INSTALL.txt`
  39 file.
  40
  41 The program is loosely based on SMI's "etherfind" although none of the
  42 etherfind code remains.  It was originally written by Van Jacobson as
  43 part of an ongoing research project to investigate and improve tcp and
  44 internet gateway performance.  The parts of the program originally
  45 taken from Sun's etherfind were later re-written by Steven McCanne of
  46 LBL.  To insure that there would be no vestige of proprietary code in
  47 tcpdump, Steve wrote these pieces from the specification given by the
  48 manual entry, with no access to the source of tcpdump or etherfind.
  49
  50 Over the past few years, tcpdump has been steadily improved by the
  51 excellent contributions from the Internet community (just browse
  52 through the `CHANGES` file).  We are grateful for all the input.
  53
  54 Richard Stevens gives an excellent treatment of the Internet protocols
  55 in his book *"TCP/IP Illustrated, Volume 1"*. If you want to learn more
  56 about tcpdump and how to interpret its output, pick up this book.
  57
  58 Some tools for viewing and analyzing tcpdump trace files are available
  59 from the Internet Traffic Archive:
  60
  61 * http://ita.ee.lbl.gov/
  62
  63 Another tool that tcpdump users might find useful is tcpslice:
  64
  65 * https://github.com/the-tcpdump-group/tcpslice
  66
  67 It is a program that can be used to extract portions of tcpdump binary
  68 trace files. See the above distribution for further details and
  69 documentation.
  70
  71 Current versions can be found at https://www.tcpdump.org.
  72
  73  - The TCPdump group
  74
  75 original text by: Steve McCanne, Craig Leres, Van Jacobson
  76
  77 -------------------------------------
  78 ```
  79 This directory also contains some short awk programs intended as
  80 examples of ways to reduce tcpdump data when you're tracking
  81 particular network problems:
  82
  83 send-ack.awk
  84         Simplifies the tcpdump trace for an ftp (or other unidirectional
  85         tcp transfer).  Since we assume that one host only sends and
  86         the other only acks, all address information is left off and
  87         we just note if the packet is a "send" or an "ack".
  88
  89         There is one output line per line of the original trace.
  90         Field 1 is the packet time in decimal seconds, relative
  91         to the start of the conversation.  Field 2 is delta-time
  92         from last packet.  Field 3 is packet type/direction.
  93         "Send" means data going from sender to receiver, "ack"
  94         means an ack going from the receiver to the sender.  A
  95         preceding "*" indicates that the data is a retransmission.
  96         A preceding "-" indicates a hole in the sequence space
  97         (i.e., missing packet(s)), a "#" means an odd-size (not max
  98         seg size) packet.  Field 4 has the packet flags
  99         (same format as raw trace).  Field 5 is the sequence
 100         number (start seq. num for sender, next expected seq number
 101         for acks).  The number in parens following an ack is
 102         the delta-time from the first send of the packet to the
 103         ack.  A number in parens following a send is the
 104         delta-time from the first send of the packet to the
 105         current send (on duplicate packets only).  Duplicate
 106         sends or acks have a number in square brackets showing
 107         the number of duplicates so far.
 108
 109         Here is a short sample from near the start of an ftp:
 110                 3.00    0.20   send . 512
 111                 3.20    0.20    ack . 1024  (0.20)
 112                 3.20    0.00   send P 1024
 113                 3.40    0.20    ack . 1536  (0.20)
 114                 3.80    0.40 * send . 0  (3.80) [2]
 115                 3.82    0.02 *  ack . 1536  (0.62) [2]
 116         Three seconds into the conversation, bytes 512 through 1023
 117         were sent.  200ms later they were acked.  Shortly thereafter
 118         bytes 1024-1535 were sent and again acked after 200ms.
 119         Then, for no apparent reason, 0-511 is retransmitted, 3.8
 120         seconds after its initial send (the round trip time for this
 121         ftp was 1sec, +-500ms).  Since the receiver is expecting
 122         1536, 1536 is re-acked when 0 arrives.
 123
 124 packetdat.awk
 125         Computes chunk summary data for an ftp (or similar
 126         unidirectional tcp transfer). [A "chunk" refers to
 127         a chunk of the sequence space -- essentially the packet
 128         sequence number divided by the max segment size.]
 129
 130         A summary line is printed showing the number of chunks,
 131         the number of packets it took to send that many chunks
 132         (if there are no lost or duplicated packets, the number
 133         of packets should equal the number of chunks) and the
 134         number of acks.
 135
 136         Following the summary line is one line of information
 137         per chunk.  The line contains eight fields:
 138            1 - the chunk number
 139            2 - the start sequence number for this chunk
 140            3 - time of first send
 141            4 - time of last send
 142            5 - time of first ack
 143            6 - time of last ack
 144            7 - number of times chunk was sent
 145            8 - number of times chunk was acked
 146         (all times are in decimal seconds, relative to the start
 147         of the conversation.)
 148
 149         As an example, here is the first part of the output for
 150         an ftp trace:
 151
 152         # 134 chunks.  536 packets sent.  508 acks.
 153         1       1       0.00    5.80    0.20    0.20    4       1
 154         2       513     0.28    6.20    0.40    0.40    4       1
 155         3       1025    1.16    6.32    1.20    1.20    4       1
 156         4       1561    1.86    15.00   2.00    2.00    6       1
 157         5       2049    2.16    15.44   2.20    2.20    5       1
 158         6       2585    2.64    16.44   2.80    2.80    5       1
 159         7       3073    3.00    16.66   3.20    3.20    4       1
 160         8       3609    3.20    17.24   3.40    5.82    4       11
 161         9       4097    6.02    6.58    6.20    6.80    2       5
 162
 163         This says that 134 chunks were transferred (about 70K
 164         since the average packet size was 512 bytes).  It took
 165         536 packets to transfer the data (i.e., on the average
 166         each chunk was transmitted four times).  Looking at,
 167         say, chunk 4, we see it represents the 512 bytes of
 168         sequence space from 1561 to 2048.  It was first sent
 169         1.86 seconds into the conversation.  It was last
 170         sent 15 seconds into the conversation and was sent
 171         a total of 6 times (i.e., it was retransmitted every
 172         2 seconds on the average).  It was acked once, 140ms
 173         after it first arrived.
 174
 175 stime.awk
 176 atime.awk
 177         Output one line per send or ack, respectively, in the form
 178                 <time> <seq. number>
 179         where <time> is the time in seconds since the start of the
 180         transfer and <seq. number> is the sequence number being sent
 181         or acked.  I typically plot this data looking for suspicious
 182         patterns.
 183
 184
 185 The problem I was looking at was the bulk-data-transfer
 186 throughput of medium delay network paths (1-6 sec.  round trip
 187 time) under typical DARPA Internet conditions.  The trace of the
 188 ftp transfer of a large file was used as the raw data source.
 189 The method was:
 190
 191   - On a local host (but not the Sun running tcpdump), connect to
 192     the remote ftp.
 193
 194   - On the monitor Sun, start the trace going.  E.g.,
 195       tcpdump host local-host and remote-host and port ftp-data >tracefile
 196
 197   - On local, do either a get or put of a large file (~500KB),
 198     preferably to the null device (to minimize effects like
 199     closing the receive window while waiting for a disk write).
 200
 201   - When transfer is finished, stop tcpdump.  Use awk to make up
 202     two files of summary data (maxsize is the maximum packet size,
 203     tracedata is the file of tcpdump tracedata):
 204       awk -f send-ack.awk packetsize=avgsize tracedata >sa
 205       awk -f packetdat.awk packetsize=avgsize tracedata >pd
 206
 207   - While the summary data files are printing, take a look at
 208     how the transfer behaved:
 209       awk -f stime.awk tracedata | xgraph
 210     (90% of what you learn seems to happen in this step).
 211
 212   - Do all of the above steps several times, both directions,
 213     at different times of day, with different protocol
 214     implementations on the other end.
 215
 216   - Using one of the Unix data analysis packages (in my case,
 217     S and Gary Perlman's Unix|Stat), spend a few months staring
 218     at the data.
 219
 220   - Change something in the local protocol implementation and
 221     redo the steps above.
 222
 223   - Once a week, tell your funding agent that you're discovering
 224     wonderful things and you'll write up that research report
 225     "real soon now".
 226 ```