NetBSD/usr.sbin/tcpdump/README

Fri May 22 17:19:41 PDT 1992
Version 2.2

This directory contains tcpdump version 2.2, the Berkely Packet Filter (BPF),
and tcpslice, a tool for manipulating raw packet traces.  Note that there
was no 2.1 release.  Version 2.1beta was released with the BSD Networking 2
tape, but we never got around to a general 2.1 release.

Tcpdump runs on following platforms:

	machine		os				packet filter
	-------		--				-------------
	hp300		4.3BSD Tahoe/Reno		bpf
	sparc		SunOS 4.x			bpf, nit
	sun3		SunOS 3.5, SunOS 4.x		bpf, nit
	Decstation	Ultrix 4.0 (and higher)		packetfilter
	IBM RT		4.3BSD				enet
	386/486		4.3BSD netII			bpf

BPF can be installed in SunOS kernels, provided you have source, 4.3BSD
kernels, and is now standard in BSD systems.  (The Networking II release
from CSRG has BPF support.  Tcpdump can be temporarily found in 
/usr/src/contrib.)  See bpf/README for further details and an installation
procedure.

In addition to bug fixes, changes from version 2.0 include:

- Easy access to icmp packets, via the 'icmp' keyword.  For example,

	% tcpdump 'icmp[0] != 8 and icmp[0] != 0'

  matches non-echo/reply ICMP packets.

- An improved filter code optimizer.

- A multicast keyword.  Also, the broadcast keyword can now be qualified 
  with a protocol layer.  For instance, "ip broadcast" and "ether multicast"
  are valid filters.

- Support for monitoring the loopback interface (i.e. 'tcpdump -i lo').
  Jeffrey Honig (jch@MITCHELL.CIT.CORNELL.EDU) contributed the kernel
  patches to netinet/if_loop.c.

- Support for the Ungermann-Bass Ethernet on IBM/PC-RTs running AOS.
  Contact Jeffrey Honig for the diffs.

- Decoding of EGP and OSPF packets, thanks to Jeffrey Honig.

- The tcpslice program, thanks to Vern Paxson (vern@ee.lbl.gov).  Check
  out the man page (tcpslice.1) for more information.

The BPF kernel interface has changed and is not backward compatible
with the interface from the 2.0 release.  tcpdump-2.1 won't work with 
old versions of BPF, but will work with the CSRG Networking II release.
If you've made the bpf changes to network drivers, you'll need to update
them (see bpf/README).  Also, if you've written any BPF applications,
they may need some minor changes (with respect to ioctls).

The BPF man page is improved and contains an explanation of the
filtering language.

Tcpdump's makefile has continued to evolve.  Multiple platforms are
supported using subdirectories.  See INSTALL for more details.

Problems, bugs, questions, desirable enhancements, etc., should be
sent to the email address "tcpdump@ee.lbl.gov".  

 - Steve McCanne (mccanne@ee.lbl.gov)
   Craig Leres (leres@ee.lbl.gov)
   Van Jacobson (van@ee.lbl.gov)

------------------------------

Old news:

 - A packet dumper has been added (thanks to Jeff Mogul of DECWRL). 
   With this option, you can create an architecture independent binary 
   trace file in real time, without the overhead of the packet printer.  
   At a later time, the packets can be filtered (again) and printed.

 - BSD is supported.  You must have BPF in your kernel.
   Since the filtering is now done in the kernel, fewer packets are
   dropped.  In fact, with BPF and the packet dumper option, a measly
   Sun 3/50 can keep up with a busy network.

 - Compressed SLIP packets can now be dumped, provided you use our
   SLIP software and BPF.  These packets are dumped as any other IP
   packet; the compressed headers are dumped with the '-e' option.

 - Machines with little-endian byte ordering are supported (thanks to
   Jeff Mogul).

 - Ultrix 4.0 is supported (also thanks to Jeff Mogul).

 - IBM RT and Stanford Enetfilter support has been added by
   Rayan Zachariassen <rayan@canet.ca>.  Tcpdump has been tested under
   both the vanilla enetfilter interface, and the extended interface 
   (#ifdef'd by IBMRTPC) present in the MERIT version of the enetfilter.

 - TFTP packets are now printed (requests only).

 - BOOTP packets are now printed.

 - SNMP packets are now printed. (thanks to John LoVerso of Xylogics).

 - Sparc architectures, including the Sparcstation-1, are now
   supported thanks to Steve McCanne and Craig Leres.

 - SunOS 4 is now supported thanks to Micky Liu of Columbia
   University (micky@cunixc.cc.columbia.edu).

 - IP options are now printed.

 - RIP packets are now printed.

 - There's a -v flag that prints out more information than the
   default (e.g., it will enable printing of IP ttl, tos and id)
   and -q flag that prints out less (e.g., it will disable
   interpretation of AppleTalk-in-UDP).

 - The grammar has undergone substantial changes (if you have an
   earlier version of tcpdump, you should re-read the manual
   entry).

   The most useful change is the addition of an expression
   syntax that lets you filter on arbitrary fields or values in the
   packet.  E.g., "ip[0] > 0x45" would print only packets with IP
   options, "tcp[13] & 3 != 0" would print only TCP SYN and FIN
   packets.

   The most painful change is that concatenation no longer means
   "and" -- e.g., you have to say "host foo and port bar" instead
   of "host foo port bar".  The up side to this down is that
   repeated qualifiers can be omitted, making most filter
   expressions shorter.  E.g., you can now say "ip host foo and
   (bar or baz)" to look at ip traffic between hosts foo and bar or
   between hosts foo and baz.  [The old way of saying this was "ip
   host foo and (ip host bar or ip host baz)".]

------------------------------

The program is loosely based on SMI's "etherfind" although none
of the etherfind code remains.  It was originally written by Van
Jacobson, Lawrence Berkeley Laboratory, as part of an ongoing
research project to investigate and improve tcp and internet
gateway performance.  The parts of the program originally taken
from Sun's etherfind were later re-written by Steve McCanne of
LBL.  To insure that would be no vestige of proprietary code in
tcpdump, Steve wrote these pieces from the specification given
by the manual entry, with no access to the source of tcpdump or
etherfind.

The current versions of these files are available via anonymous
ftp from host ftp.ee.lbl.gov (currently at address 128.3.254.68)
file tcpdump.tar.Z (a compressed Unix tar file).

This program is subject to the 'standard' Berkeley network software
copyright:
  
  Copyright (c) 1988-1990 The Regents of the University of California.
  All rights reserved.

  Redistribution and use in source and binary forms, with or without
  modification, are permitted provided that: (1) source code distributions
  retain the above copyright notice and this paragraph in its entirety, (2)
  distributions including binary code include the above copyright notice and
  this paragraph in its entirety in the documentation or other materials
  provided with the distribution, and (3) all advertising materials mentioning
  features or use of this software display the following acknowledgement:
  ``This product includes software developed by the University of California,
  Lawrence Berkeley Laboratory and its contributors.'' Neither the name of
  the University nor the names of its contributors may be used to endorse
  or promote products derived from this software without specific prior
  written permission.
  THIS SOFTWARE IS PROVIDED ``AS IS'' AND WITHOUT ANY EXPRESS OR IMPLIED
  WARRANTIES, INCLUDING, WITHOUT LIMITATION, THE IMPLIED WARRANTIES OF
  MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE.

Enjoy.

  - Van Jacobson, Steve McCanne, Craig Leres
----------------------------

This directory also contains some short awk programs intended as
examples of ways to reduce tcpdump data when you're tracking
particular network problems:

send-ack.awk
	Simplifies the tcpdump trace for an ftp (or other unidirectional
	tcp transfer).  Since we assume that one host only sends and
	the other only acks, all address information is left off and
	we just note if the packet is a "send" or an "ack".

	There is one output line per line of the original trace. 
	Field 1 is the packet time in decimal seconds, relative
	to the start of the conversation.  Field 2 is delta-time
	from last packet.  Field 3 is packet type/direction. 
	"Send" means data going from sender to receiver, "ack"
	means an ack going from the receiver to the sender.  A
	preceding "*" indicates that the data is a retransmission.
	A preceding "-" indicates a hole in the sequence space
	(i.e., missing packet(s)), a "#" means an odd-size (not max
	seg size) packet.  Field 4 has the packet flags
	(same format as raw trace).  Field 5 is the sequence
	number (start seq. num for sender, next expected seq number
	for acks).  The number in parens following an ack is
	the delta-time from the first send of the packet to the
	ack.  A number in parens following a send is the
	delta-time from the first send of the packet to the
	current send (on duplicate packets only).  Duplicate
	sends or acks have a number in square brackets showing
	the number of duplicates so far. 

	Here is a short sample from near the start of an ftp:
		3.00    0.20   send . 512
		3.20    0.20    ack . 1024  (0.20)
		3.20    0.00   send P 1024
		3.40    0.20    ack . 1536  (0.20)
		3.80    0.40 * send . 0  (3.80) [2]
		3.82    0.02 *  ack . 1536  (0.62) [2]
	Three seconds into the conversation, bytes 512 through 1023
	were sent.  200ms later they were acked.  Shortly thereafter
	bytes 1024-1535 were sent and again acked after 200ms. 
	Then, for no apparent reason, 0-511 is retransmitted, 3.8
	seconds after its initial send (the round trip time for this
	ftp was 1sec, +-500ms).  Since the receiver is expecting
	1536, 1536 is re-acked when 0 arrives. 

packetdat.awk
	Computes chunk summary data for an ftp (or similar
	unidirectional tcp transfer). [A "chunk" refers to
	a chunk of the sequence space -- essentially the packet
	sequence number divided by the max segment size.]

	A summary line is printed showing the number of chunks,
	the number of packets it took to send that many chunks
	(if there are no lost or duplicated packets, the number
	of packets should equal the number of chunks) and the
	number of acks.

	Following the summary line is one line of information
	per chunk.  The line contains eight fields:
	   1 - the chunk number
	   2 - the start sequence number for this chunk
	   3 - time of first send
	   4 - time of last send
	   5 - time of first ack
	   6 - time of last ack
	   7 - number of times chunk was sent 
	   8 - number of times chunk was acked
	(all times are in decimal seconds, relative to the start
	of the conversation.)

	As an example, here is the first part of the output for
	an ftp trace:

	# 134 chunks.  536 packets sent.  508 acks.
	1       1       0.00    5.80    0.20    0.20    4       1
	2       513     0.28    6.20    0.40    0.40    4       1
	3       1025    1.16    6.32    1.20    1.20    4       1
	4       1561    1.86    15.00   2.00    2.00    6       1
	5       2049    2.16    15.44   2.20    2.20    5       1
	6       2585    2.64    16.44   2.80    2.80    5       1
	7       3073    3.00    16.66   3.20    3.20    4       1
	8       3609    3.20    17.24   3.40    5.82    4       11
	9       4097    6.02    6.58    6.20    6.80    2       5

	This says that 134 chunks were transfered (about 70K
	since the average packet size was 512 bytes).  It took
	536 packets to transfer the data (i.e., on the average
	each chunk was transmitted four times).  Looking at,
	say, chunk 4, we see it represents the 512 bytes of
	sequence space from 1561 to 2048.  It was first sent
	1.86 seconds into the conversation.  It was last
	sent 15 seconds into the conversation and was sent
	a total of 6 times (i.e., it was retransmitted every
	2 seconds on the average).  It was acked once, 140ms
	after it first arrived.

stime.awk
atime.awk
	Output one line per send or ack, respectively, in the form
		<time> <seq. number>
	where <time> is the time in seconds since the start of the
	transfer and <seq. number> is the sequence number being sent
	or acked.  I typically plot this data looking for suspicious
	patterns.


The problem I was looking at was the bulk-data-transfer
throughput of medium delay network paths (1-6 sec.  round trip
time) under typical DARPA Internet conditions.  The trace of the
ftp transfer of a large file was used as the raw data source. 
The method was:

  - On a local host (but not the Sun running tcpdump), connect to
    the remote ftp.

  - On the monitor Sun, start the trace going.  E.g.,
      tcpdump host local-host and remote-host and port ftp-data >tracefile

  - On local, do either a get or put of a large file (~500KB),
    preferably to the null device (to minimize effects like
    closing the receive window while waiting for a disk write).

  - When tranfer is finished, stop tcpdump.  Use awk to make up
    two files of summary data (maxsize is the maximum packet size,
    tracedata is the file of tcpdump tracedata):
      awk -f send-ack.awk packetsize=avgsize tracedata >sa
      awk -f packetdat.awk packetsize=avgsize tracedata >pd

  - While the summary data files are printing, take a look at
    how the transfer behaved:
      awk -f stime.awk tracedata | xgraph
    (90% of what you learn seems to happen in this step).

  - Do all of the above steps several times, both directions,
    at different times of day, with different protocol
    implementations on the other end.

  - Using one of the Unix data analysis packages (in my case,
    S and Gary Perlman's Unix|Stat), spend a few months staring
    at the data.

  - Change something in the local protocol implementation and
    redo the steps above.

  - Once a week, tell your funding agent that you're discovering
    wonderful things and you'll write up that research report
    "real soon now".