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

   1 # tcpdump
   2
   3 [![Build
   4 Status](https://travis-ci.org/the-tcpdump-group/tcpdump.png)](https://travis-ci.org/the-tcpdump-group/tcpdump)
   5
   6 TCPDUMP 4.x.y
   7 Now maintained by "The Tcpdump Group"
   8 See             www.tcpdump.org
   9
  10 Please send inquiries/comments/reports to:
  11
  12 * tcpdump-workers@lists.tcpdump.org
  13
  14 Anonymous Git is available via:
  15
  16         git clone git://bpf.tcpdump.org/tcpdump
  17
  18 Version 4.x.y of TCPDUMP can be retrieved with the CVS tag `tcpdump_4_xrely`:
  19
  20         cvs -d :pserver:cvs.tcpdump.org:/tcpdump/master checkout -r tcpdump_4_xrely tcpdump
  21
  22 Please submit patches by forking the branch on GitHub at:
  23
  24 *       http://github.com/the-tcpdump-group/tcpdump/tree/master
  25
  26 and issuing a pull request.
  27
  28 formerly from   Lawrence Berkeley National Laboratory
  29                 Network Research Group <tcpdump@ee.lbl.gov>
  30                 ftp://ftp.ee.lbl.gov/tcpdump.tar.Z (3.4)
  31
  32 This directory contains source code for tcpdump, a tool for network
  33 monitoring and data acquisition.  This software was originally
  34 developed by the Network Research Group at the Lawrence Berkeley
  35 National Laboratory.  The original distribution is available via
  36 anonymous ftp to `ftp.ee.lbl.gov`, in `tcpdump.tar.Z`.  More recent
  37 development is performed at tcpdump.org, http://www.tcpdump.org/
  38
  39 Tcpdump uses libpcap, a system-independent interface for user-level
  40 packet capture.  Before building tcpdump, you must first retrieve and
  41 build libpcap, also originally from LBL and now being maintained by
  42 tcpdump.org; see http://www.tcpdump.org/ .
  43
  44 Once libpcap is built (either install it or make sure it's in
  45 `../libpcap`), you can build tcpdump using the procedure in the `INSTALL.txt`
  46 file.
  47
  48 The program is loosely based on SMI's "etherfind" although none of the
  49 etherfind code remains.  It was originally written by Van Jacobson as
  50 part of an ongoing research project to investigate and improve tcp and
  51 internet gateway performance.  The parts of the program originally
  52 taken from Sun's etherfind were later re-written by Steven McCanne of
  53 LBL.  To insure that there would be no vestige of proprietary code in
  54 tcpdump, Steve wrote these pieces from the specification given by the
  55 manual entry, with no access to the source of tcpdump or etherfind.
  56
  57 Over the past few years, tcpdump has been steadily improved by the
  58 excellent contributions from the Internet community (just browse
  59 through the `CHANGES` file).  We are grateful for all the input.
  60
  61 Richard Stevens gives an excellent treatment of the Internet protocols
  62 in his book *"TCP/IP Illustrated, Volume 1"*. If you want to learn more
  63 about tcpdump and how to interpret its output, pick up this book.
  64
  65 Some tools for viewing and analyzing tcpdump trace files are available
  66 from the Internet Traffic Archive:
  67
  68 * http://www.acm.org/sigcomm/ITA/
  69
  70 Another tool that tcpdump users might find useful is tcpslice:
  71
  72 * ftp://ftp.ee.lbl.gov/tcpslice.tar.Z
  73
  74 It is a program that can be used to extract portions of tcpdump binary
  75 trace files. See the above distribution for further details and
  76 documentation.
  77
  78 Problems, bugs, questions, desirable enhancements, etc. should be sent
  79 to the address "tcpdump-workers@lists.tcpdump.org".  Bugs, support
  80 requests, and feature requests may also be submitted on the GitHub issue
  81 tracker for tcpdump at:
  82
  83 * https://github.com/the-tcpdump-group/tcpdump/issues
  84
  85 Source code contributions, etc. should be sent to the email address
  86 above or submitted by forking the branch on GitHub at:
  87
  88 * http://github.com/the-tcpdump-group/tcpdump/tree/master
  89
  90 and issuing a pull request.
  91
  92 Current versions can be found at www.tcpdump.org.
  93
  94  - The TCPdump team
  95
  96 original text by: Steve McCanne, Craig Leres, Van Jacobson
  97
  98 -------------------------------------
  99 ```
 100 This directory also contains some short awk programs intended as
 101 examples of ways to reduce tcpdump data when you're tracking
 102 particular network problems:
 103
 104 send-ack.awk
 105         Simplifies the tcpdump trace for an ftp (or other unidirectional
 106         tcp transfer).  Since we assume that one host only sends and
 107         the other only acks, all address information is left off and
 108         we just note if the packet is a "send" or an "ack".
 109
 110         There is one output line per line of the original trace.
 111         Field 1 is the packet time in decimal seconds, relative
 112         to the start of the conversation.  Field 2 is delta-time
 113         from last packet.  Field 3 is packet type/direction.
 114         "Send" means data going from sender to receiver, "ack"
 115         means an ack going from the receiver to the sender.  A
 116         preceding "*" indicates that the data is a retransmission.
 117         A preceding "-" indicates a hole in the sequence space
 118         (i.e., missing packet(s)), a "#" means an odd-size (not max
 119         seg size) packet.  Field 4 has the packet flags
 120         (same format as raw trace).  Field 5 is the sequence
 121         number (start seq. num for sender, next expected seq number
 122         for acks).  The number in parens following an ack is
 123         the delta-time from the first send of the packet to the
 124         ack.  A number in parens following a send is the
 125         delta-time from the first send of the packet to the
 126         current send (on duplicate packets only).  Duplicate
 127         sends or acks have a number in square brackets showing
 128         the number of duplicates so far.
 129
 130         Here is a short sample from near the start of an ftp:
 131                 3.00    0.20   send . 512
 132                 3.20    0.20    ack . 1024  (0.20)
 133                 3.20    0.00   send P 1024
 134                 3.40    0.20    ack . 1536  (0.20)
 135                 3.80    0.40 * send . 0  (3.80) [2]
 136                 3.82    0.02 *  ack . 1536  (0.62) [2]
 137         Three seconds into the conversation, bytes 512 through 1023
 138         were sent.  200ms later they were acked.  Shortly thereafter
 139         bytes 1024-1535 were sent and again acked after 200ms.
 140         Then, for no apparent reason, 0-511 is retransmitted, 3.8
 141         seconds after its initial send (the round trip time for this
 142         ftp was 1sec, +-500ms).  Since the receiver is expecting
 143         1536, 1536 is re-acked when 0 arrives.
 144
 145 packetdat.awk
 146         Computes chunk summary data for an ftp (or similar
 147         unidirectional tcp transfer). [A "chunk" refers to
 148         a chunk of the sequence space -- essentially the packet
 149         sequence number divided by the max segment size.]
 150
 151         A summary line is printed showing the number of chunks,
 152         the number of packets it took to send that many chunks
 153         (if there are no lost or duplicated packets, the number
 154         of packets should equal the number of chunks) and the
 155         number of acks.
 156
 157         Following the summary line is one line of information
 158         per chunk.  The line contains eight fields:
 159            1 - the chunk number
 160            2 - the start sequence number for this chunk
 161            3 - time of first send
 162            4 - time of last send
 163            5 - time of first ack
 164            6 - time of last ack
 165            7 - number of times chunk was sent
 166            8 - number of times chunk was acked
 167         (all times are in decimal seconds, relative to the start
 168         of the conversation.)
 169
 170         As an example, here is the first part of the output for
 171         an ftp trace:
 172
 173         # 134 chunks.  536 packets sent.  508 acks.
 174         1       1       0.00    5.80    0.20    0.20    4       1
 175         2       513     0.28    6.20    0.40    0.40    4       1
 176         3       1025    1.16    6.32    1.20    1.20    4       1
 177         4       1561    1.86    15.00   2.00    2.00    6       1
 178         5       2049    2.16    15.44   2.20    2.20    5       1
 179         6       2585    2.64    16.44   2.80    2.80    5       1
 180         7       3073    3.00    16.66   3.20    3.20    4       1
 181         8       3609    3.20    17.24   3.40    5.82    4       11
 182         9       4097    6.02    6.58    6.20    6.80    2       5
 183
 184         This says that 134 chunks were transferred (about 70K
 185         since the average packet size was 512 bytes).  It took
 186         536 packets to transfer the data (i.e., on the average
 187         each chunk was transmitted four times).  Looking at,
 188         say, chunk 4, we see it represents the 512 bytes of
 189         sequence space from 1561 to 2048.  It was first sent
 190         1.86 seconds into the conversation.  It was last
 191         sent 15 seconds into the conversation and was sent
 192         a total of 6 times (i.e., it was retransmitted every
 193         2 seconds on the average).  It was acked once, 140ms
 194         after it first arrived.
 195
 196 stime.awk
 197 atime.awk
 198         Output one line per send or ack, respectively, in the form
 199                 <time> <seq. number>
 200         where <time> is the time in seconds since the start of the
 201         transfer and <seq. number> is the sequence number being sent
 202         or acked.  I typically plot this data looking for suspicious
 203         patterns.
 204
 205
 206 The problem I was looking at was the bulk-data-transfer
 207 throughput of medium delay network paths (1-6 sec.  round trip
 208 time) under typical DARPA Internet conditions.  The trace of the
 209 ftp transfer of a large file was used as the raw data source.
 210 The method was:
 211
 212   - On a local host (but not the Sun running tcpdump), connect to
 213     the remote ftp.
 214
 215   - On the monitor Sun, start the trace going.  E.g.,
 216       tcpdump host local-host and remote-host and port ftp-data >tracefile
 217
 218   - On local, do either a get or put of a large file (~500KB),
 219     preferably to the null device (to minimize effects like
 220     closing the receive window while waiting for a disk write).
 221
 222   - When transfer is finished, stop tcpdump.  Use awk to make up
 223     two files of summary data (maxsize is the maximum packet size,
 224     tracedata is the file of tcpdump tracedata):
 225       awk -f send-ack.awk packetsize=avgsize tracedata >sa
 226       awk -f packetdat.awk packetsize=avgsize tracedata >pd
 227
 228   - While the summary data files are printing, take a look at
 229     how the transfer behaved:
 230       awk -f stime.awk tracedata | xgraph
 231     (90% of what you learn seems to happen in this step).
 232
 233   - Do all of the above steps several times, both directions,
 234     at different times of day, with different protocol
 235     implementations on the other end.
 236
 237   - Using one of the Unix data analysis packages (in my case,
 238     S and Gary Perlman's Unix|Stat), spend a few months staring
 239     at the data.
 240
 241   - Change something in the local protocol implementation and
 242     redo the steps above.
 243
 244   - Once a week, tell your funding agent that you're discovering
 245     wonderful things and you'll write up that research report
 246     "real soon now".
 247 ```