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