README revision 146773
1@(#) $Header: /tcpdump/master/tcpdump/README,v 1.65 2004/10/12 02:01:59 guy Exp $ (LBL) 2 3TCPDUMP 3.9 4Now maintained by "The Tcpdump Group" 5See www.tcpdump.org 6 7Please send inquiries/comments/reports to tcpdump-workers@tcpdump.org 8 9Anonymous CVS is available via: 10 cvs -d :pserver:cvs.tcpdump.org:/tcpdump/master login 11 (password "anoncvs") 12 cvs -d :pserver:cvs.tcpdump.org:/tcpdump/master checkout tcpdump 13 14Version 3.9 of TCPDUMP can be retrived with the CVS tag "tcpdump_3_9rel1": 15 cvs -d :pserver:cvs.tcpdump.org:/tcpdump/master checkout -r tcpdump_3_9rel1 tcpdump 16 17Please send patches against the master copy to patches@tcpdump.org. 18 19formerly from Lawrence Berkeley National Laboratory 20 Network Research Group <tcpdump@ee.lbl.gov> 21 ftp://ftp.ee.lbl.gov/tcpdump.tar.Z (3.4) 22 23This directory contains source code for tcpdump, a tool for network 24monitoring and data acquisition. This software was originally 25developed by the Network Research Group at the Lawrence Berkeley 26National Laboratory. The original distribution is available via 27anonymous ftp to ftp.ee.lbl.gov, in tcpdump.tar.Z. More recent 28development is performed at tcpdump.org, http://www.tcpdump.org/ 29 30Tcpdump uses libpcap, a system-independent interface for user-level 31packet capture. Before building tcpdump, you must first retrieve and 32build libpcap, also originally from LBL and now being maintained by 33tcpdump.org; see http://www.tcpdump.org/ . 34 35Once libpcap is built (either install it or make sure it's in 36../libpcap), you can build tcpdump using the procedure in the INSTALL 37file. 38 39The program is loosely based on SMI's "etherfind" although none of the 40etherfind code remains. It was originally written by Van Jacobson as 41part of an ongoing research project to investigate and improve tcp and 42internet gateway performance. The parts of the program originally 43taken from Sun's etherfind were later re-written by Steven McCanne of 44LBL. To insure that there would be no vestige of proprietary code in 45tcpdump, Steve wrote these pieces from the specification given by the 46manual entry, with no access to the source of tcpdump or etherfind. 47 48Over the past few years, tcpdump has been steadily improved by the 49excellent contributions from the Internet community (just browse 50through the CHANGES file). We are grateful for all the input. 51 52Richard Stevens gives an excellent treatment of the Internet protocols 53in his book ``TCP/IP Illustrated, Volume 1''. If you want to learn more 54about tcpdump and how to interpret its output, pick up this book. 55 56Some tools for viewing and analyzing tcpdump trace files are available 57from the Internet Traffic Archive: 58 59 http://www.acm.org/sigcomm/ITA/ 60 61Another tool that tcpdump users might find useful is tcpslice: 62 63 ftp://ftp.ee.lbl.gov/tcpslice.tar.Z 64 65It is a program that can be used to extract portions of tcpdump binary 66trace files. See the above distribution for further details and 67documentation. 68 69Problems, bugs, questions, desirable enhancements, etc. should be sent 70to the address "tcpdump-workers@tcpdump.org". Bugs, support requests, 71and feature requests may also be submitted on the SourceForge site for 72tcpdump at 73 74 http://sourceforge.net/projects/tcpdump/ 75 76Source code contributions, etc. should be sent to the email address 77"patches@tcpdump.org", or submitted as patches on the SourceForge site 78for tcpdump. 79 80Current versions can be found at www.tcpdump.org, or the SourceForge 81site for tcpdump. 82 83 - The TCPdump team 84 85original text by: Steve McCanne, Craig Leres, Van Jacobson 86 87------------------------------------- 88This directory also contains some short awk programs intended as 89examples of ways to reduce tcpdump data when you're tracking 90particular network problems: 91 92send-ack.awk 93 Simplifies the tcpdump trace for an ftp (or other unidirectional 94 tcp transfer). Since we assume that one host only sends and 95 the other only acks, all address information is left off and 96 we just note if the packet is a "send" or an "ack". 97 98 There is one output line per line of the original trace. 99 Field 1 is the packet time in decimal seconds, relative 100 to the start of the conversation. Field 2 is delta-time 101 from last packet. Field 3 is packet type/direction. 102 "Send" means data going from sender to receiver, "ack" 103 means an ack going from the receiver to the sender. A 104 preceding "*" indicates that the data is a retransmission. 105 A preceding "-" indicates a hole in the sequence space 106 (i.e., missing packet(s)), a "#" means an odd-size (not max 107 seg size) packet. Field 4 has the packet flags 108 (same format as raw trace). Field 5 is the sequence 109 number (start seq. num for sender, next expected seq number 110 for acks). The number in parens following an ack is 111 the delta-time from the first send of the packet to the 112 ack. A number in parens following a send is the 113 delta-time from the first send of the packet to the 114 current send (on duplicate packets only). Duplicate 115 sends or acks have a number in square brackets showing 116 the number of duplicates so far. 117 118 Here is a short sample from near the start of an ftp: 119 3.00 0.20 send . 512 120 3.20 0.20 ack . 1024 (0.20) 121 3.20 0.00 send P 1024 122 3.40 0.20 ack . 1536 (0.20) 123 3.80 0.40 * send . 0 (3.80) [2] 124 3.82 0.02 * ack . 1536 (0.62) [2] 125 Three seconds into the conversation, bytes 512 through 1023 126 were sent. 200ms later they were acked. Shortly thereafter 127 bytes 1024-1535 were sent and again acked after 200ms. 128 Then, for no apparent reason, 0-511 is retransmitted, 3.8 129 seconds after its initial send (the round trip time for this 130 ftp was 1sec, +-500ms). Since the receiver is expecting 131 1536, 1536 is re-acked when 0 arrives. 132 133packetdat.awk 134 Computes chunk summary data for an ftp (or similar 135 unidirectional tcp transfer). [A "chunk" refers to 136 a chunk of the sequence space -- essentially the packet 137 sequence number divided by the max segment size.] 138 139 A summary line is printed showing the number of chunks, 140 the number of packets it took to send that many chunks 141 (if there are no lost or duplicated packets, the number 142 of packets should equal the number of chunks) and the 143 number of acks. 144 145 Following the summary line is one line of information 146 per chunk. The line contains eight fields: 147 1 - the chunk number 148 2 - the start sequence number for this chunk 149 3 - time of first send 150 4 - time of last send 151 5 - time of first ack 152 6 - time of last ack 153 7 - number of times chunk was sent 154 8 - number of times chunk was acked 155 (all times are in decimal seconds, relative to the start 156 of the conversation.) 157 158 As an example, here is the first part of the output for 159 an ftp trace: 160 161 # 134 chunks. 536 packets sent. 508 acks. 162 1 1 0.00 5.80 0.20 0.20 4 1 163 2 513 0.28 6.20 0.40 0.40 4 1 164 3 1025 1.16 6.32 1.20 1.20 4 1 165 4 1561 1.86 15.00 2.00 2.00 6 1 166 5 2049 2.16 15.44 2.20 2.20 5 1 167 6 2585 2.64 16.44 2.80 2.80 5 1 168 7 3073 3.00 16.66 3.20 3.20 4 1 169 8 3609 3.20 17.24 3.40 5.82 4 11 170 9 4097 6.02 6.58 6.20 6.80 2 5 171 172 This says that 134 chunks were transferred (about 70K 173 since the average packet size was 512 bytes). It took 174 536 packets to transfer the data (i.e., on the average 175 each chunk was transmitted four times). Looking at, 176 say, chunk 4, we see it represents the 512 bytes of 177 sequence space from 1561 to 2048. It was first sent 178 1.86 seconds into the conversation. It was last 179 sent 15 seconds into the conversation and was sent 180 a total of 6 times (i.e., it was retransmitted every 181 2 seconds on the average). It was acked once, 140ms 182 after it first arrived. 183 184stime.awk 185atime.awk 186 Output one line per send or ack, respectively, in the form 187 <time> <seq. number> 188 where <time> is the time in seconds since the start of the 189 transfer and <seq. number> is the sequence number being sent 190 or acked. I typically plot this data looking for suspicious 191 patterns. 192 193 194The problem I was looking at was the bulk-data-transfer 195throughput of medium delay network paths (1-6 sec. round trip 196time) under typical DARPA Internet conditions. The trace of the 197ftp transfer of a large file was used as the raw data source. 198The method was: 199 200 - On a local host (but not the Sun running tcpdump), connect to 201 the remote ftp. 202 203 - On the monitor Sun, start the trace going. E.g., 204 tcpdump host local-host and remote-host and port ftp-data >tracefile 205 206 - On local, do either a get or put of a large file (~500KB), 207 preferably to the null device (to minimize effects like 208 closing the receive window while waiting for a disk write). 209 210 - When transfer is finished, stop tcpdump. Use awk to make up 211 two files of summary data (maxsize is the maximum packet size, 212 tracedata is the file of tcpdump tracedata): 213 awk -f send-ack.awk packetsize=avgsize tracedata >sa 214 awk -f packetdat.awk packetsize=avgsize tracedata >pd 215 216 - While the summary data files are printing, take a look at 217 how the transfer behaved: 218 awk -f stime.awk tracedata | xgraph 219 (90% of what you learn seems to happen in this step). 220 221 - Do all of the above steps several times, both directions, 222 at different times of day, with different protocol 223 implementations on the other end. 224 225 - Using one of the Unix data analysis packages (in my case, 226 S and Gary Perlman's Unix|Stat), spend a few months staring 227 at the data. 228 229 - Change something in the local protocol implementation and 230 redo the steps above. 231 232 - Once a week, tell your funding agent that you're discovering 233 wonderful things and you'll write up that research report 234 "real soon now". 235