IPAudit Logo



ipaudit - IP traffic summarizer






ipaudit is an IP traffic monitor. It listens to a network interface in promiscuous mode (or reads data previously dumped to a file) and tallies the total traffic for every "connection". A connection is a combination of ip address pairs, protocol and for tcp/udp protocols, the ports numbers.



If hosts telnets to this creates a single connection. ipaudit will tally the total bytes of traffic and total packet count in each direction between the two hosts.

When ipaudit reports its findings, this telnet connection will be listed as 6 1105 23 19934 2993 41 33

The columns are (1) host 1 ip (2) host 2 ip, (3) protocol, (4) host 1 port, (5) host 2 port, (6) host 1 bytes received, (7) host 2 bytes received, (8) host 1 packets received, (9) host 2 packets received.

For protocol other then tcp/udp where there are no ports, the port numbers are set to 0 in the output. Note that the host ip octets are padded on the left with 0 to make 3 digits. This help make the output more readable.

As another example, if the first host now sends 10 ping packets to a non-existent host then it will be reported as an output line 1 0 0 0 680 0 10

This says that address received 680 bytes in 10 packets of protocol 1 - even though there is no such address!

The output is extended using the options. The -t option adds the connection starting and ending times to each output line, for example 6 1105 23 19934 2993 41 33
   14:01:32.0495 14:04:23.9491 2 1

The -e option adds the Ethernet address for the two hosts to each output line, 6 1105 23 19934 2993 41 33
   005000c0fa45 005000d1cc21



Write traffic list in binary format (experimental).

-c npacket
Terminate program after reading npacket packets.

Turn on debugging output.

Write Ethernet addresses for each host. Program stops with an error if you are using a non-Ethernet interface.

Filter incoming packets according to filter string. For example,

   ipaudit -f "host" eth0

will pass the string "host" to the pcap library's filter routine. Thus ipaudit will only see packets with in one of the two ip addresses. The filter commands are extensive and are explained fully in the tcpdump man page.

Note that if you are trying to observe vlan traffic with your filter, you will have to specify the 'vlan' filter before other filters. In fact if all your traffic was vlan tagged, the command above would filter out all traffic, because the 'vlan' specifier was not included. To see vlan traffic the example above would need to change to

   ipaudit -f "vlan and host" eth0

Note that the -f command is different from the -p command. The -f command filters packets before ipaudit reads them. Thus -f affects what is stored in the normal output (directed by the -o command) AND in the packet dump (directed by the -w command). The -p command filter packets after ipaudit reads them, but before it writes them to the dump file. Thus the normal output (-o) is not affected, but the packet dump (-w) is.

-g config_file
Reads configuration from config_file. See CONFIGURATION FILE section below.

-i pidfile
Save the program process id to pidfile. This is used to terminate program execution when sampling time has expired. Use command like

kill -2 `cat pidfile`

to signal ipaudit to stop. Once ipaudit receives this signal, it will print a list of traffic that it has seen.

-l ip-range[:ip-range[:ip-range[..]]]
Normally, the two ip address within a line of output are printed in ip sorted order. But if you are monitoring a link between two networks, you might want the addresses of one (local) network to come first on an output line. ip-range is list of ip ranges that defines the local network. Any ip address in this range is considered local and is printed first on the output line. If neither or both ip address belongs to the group of ip-range then they are printed in the default sorted order.

There can be many ip-ranges separate by colons. No spaces may appear in the argument. Each ip-range can be either a single ip address such as which indicates a range of one, a partial ip address such as 127.0.5 which indicates a range from to, a low and high ip address separate by a hypen (-), and a single ip address with a slash (/) and either an integer between 0 and 32 (a "net address") or a network such as which indicates a network.

If you run ipaudit with the debug option (-d) the program will print the entire list of ip ranges, so you can check their values.

Here is a list of arguments to -l along with the corresponding range.

   COMMAND: ipaudit -l 137.99.11

   COMMAND: ipaudit -l 137.99.11:127.0.5/23

   COMMAND: ipaudit -l

The following example,

ipaudit -c 10000 -l 137.11 eth0

will tally 10,000 packets and list the results on-screen, placing ip addresses beginning with 137.11 first on each output line. Note that ipaudit pads the bytes with leading zeros, so that the output will actually start like

137.011.nnn.nnn ...

You can turn off this padding with the -S option.

Does not turn on promiscuous mode on interface. By default promiscuous mode is enabled. Note that interface may be in promiscuous mode for other reasons.

-o outfile
Writes traffic list to outfile upon completion. By default, it writes traffic to stdout. The file name outfile can contain time format strings (see 'man strftime' for format options) that represent time at the start of ipaudit.

-p prot[,port..][:prot[,port..]..
Only dump packets with specific protocols and ports. For example, -p1:6:17 dumps only packets with protocols 1 (icmp), 6 (tcp) and 17 (udp). You can also break down udp and tcp packets by port numbers - for example -p1:6,21,23 will only dump icmp packets, ftp packets (protocol 6, port 21) and telnet packets (protocol 6, port 23).

formats output as SQL statements which can be directly piped into any SQL type database. It should work just fine for MySQL, Postgress and Oracle. Tested against MySQL currently. All that happens is that instead of a column display, output is formatted as INSERT statements that can be fed into a database as follows:

   ipaudit -q eth0 | mysql -ppassword ipaudit

or to save traffic info every 60 seconds

   ipaudit -q -D60 eth0 | mysql -ppassword ipaudit

This assumes that the table name is ipaudit and the database is also called ipaudit. The INSERT statements look as follows:

   INSERT INTO ipaudit SET ip1='',ip2='',
   INSERT INTO ipaudit SET ip1='',ip2='',

The table structure for the database is as follows:

   CREATE TABLE ipaudit (
     ip1 varchar(15) DEFAULT '' NOT NULL,
     ip2 varchar(15) DEFAULT '' NOT NULL,
     protocol tinyint(3) unsigned DEFAULT '0' NOT NULL,
     ip1port smallint(5) unsigned DEFAULT '0' NOT NULL,
     ip2port smallint(5) unsigned DEFAULT '0' NOT NULL,
     ip1bytes int(10) unsigned DEFAULT '0' NOT NULL,
     ip2bytes int(10) unsigned DEFAULT '0' NOT NULL,
     ip1pkts int(10) unsigned DEFAULT '0' NOT NULL,
     ip2pkts int(10) unsigned DEFAULT '0' NOT NULL,
     eth1 varchar(12) DEFAULT '' NOT NULL,
     eth2 varchar(12) DEFAULT '' NOT NULL,
     constart time DEFAULT '00:00:00' NOT NULL,
     constartmsec smallint(5) unsigned DEFAULT '0' NOT NULL,
     constop time DEFAULT '00:00:00' NOT NULL,
     constopmsec smallint(5) unsigned DEFAULT '0' NOT NULL,
     probename varchar(80) DEFAULT '' NOT NULL

The above structure describes the full table required if all the options are enabled. In most cases, the start, stop, ethernet addresses and probename are not required as these have to be explicitly enabled via command line options. They can be dropped from the table for faster database inputs and queries if they are not required.

Based on the info in the table all manner of computations can be done and graphs can be drawn.

-r dumpfile
Reads network info from dumpfile instead of reading live from network. Such a dumpfile could have been produced by the programs ipaudit, tcpdump or ethereal among others. You can read from standard input using '-' as the file name.

-s nlen
Save no more than first nlen packet bytes. Default is 96, minimum is 68.

Adds packets connection times to each line of traffic output. The time are in the format HH:MM:SS.SSSS where HH, MM and SS are hours, minutes and seconds (to a precision of 1/10,000). The first (second) time is the time the first (last) packet in the connection was detected. This is followed by two integers between 1 and 2. The first (second) integer is the source machine of the first packet (second) packet. A 1 (2) means the first (second) ip address on the line was the source.

Print version and exit.

-w dumpfile
Writes first nlen bytes of every packet to dumpfile (see -s option about nlen). Can later be read by ipaudit, tcpdump, ethereal, etc. The -w option does not affect traffic analysis - it will still take place. By default this writes ALL of the packets to the file. Use the -p option to write traffic for specific protocols and ports only. However the -p option does not affect normal output, all packets are still logged and stored. Contrast this with the -f option. The option -w- will write packets to standard output where they can then be piped to another program which reads pcap packets. See EXAMPLES section below for an example of sending output from ipaudit to tcpdump and also a second instance of ipaudit. Use the -W option to limit the number of packets written to dumpfile, to avoid creating large files.

The file name dumpfile can contain time format strings (see 'man strftime' for format options) that represent time at the start of ipaudit. So for example the command

   ipaudit -w %Y-%m-%d-%H:%M.raw eth0

will save raw packets in the file 2003-02-10-22:44.raw if ipaudit starts at Feb 10, 2003 at 10:44 pm.

-x program
After ipaudit ends, either from receiving a signal or reaching its packet limit or time limit (-c or -E options), call program. This option is intend for calling reports scripts which read ipaudit output after completion. You can send at most one argument to this program, for example

ipaudit -c 10000 -x 'report traffic.out' -o traffic.out eth0

runs ipaudit for 10,000 steps, places the output in the file traffic.out, and then calls the program 'report' with the argument 'traffic.out' - telling 'report' where to find its input file.

Note that to send an argument 'traffic.out' to 'report' we needed to enclose both options together in a single pair of quotes. Without the quotes -x would think that we were running 'report' without an option (and, the option traffic.out would be interpreted as the network interface - not what you wanted at all).

The program strings program can contain time format strings (see 'man strftime' for format options) that represent time at the start of ipaudit.

-z 'config file option'
You can use configuration file options (see CONFIGURATION FILE below) from the command line with the -z option. For example, instead of the '-t' option you can use the equivalanet configuration file option from the command line as follows

ipaudit -z'writetime yes' eth0

-A file[,lim]
Dump all packets to pcap format file "file". Limit number of packets to "lim" (optional).

Print ICMP type and code information in port field of the sending IP address. By default the port fields are set to zero for all ICMP (and all non-TCP and non-UDP traffic). Note that ICMP type/code data is not like TCP/UDP port data, because each IP address in an IP connection has its own port address. In the ICMP case, the type/code info applies to the entire 'connection', not to just the sender or receiver. However, in ipaudit output the type/code information is placed in the port field of the packet sender when the -C option is used.

The type/code fields are each a byte, the port value is assigned is the corresponding two byte value. The following table shows the standard values (taken from TCP/IP Illustrated Vol 1, W.R. Stevens)

    0  Echo reply
  768  Network unreachable
  769  Host unreachable
  770  Protocol unreachable
  771  Port unreachable
  772  Fragmentation needed but don't-fragment bit set
  773  Source route failed
  774  Destination network Unknown
  775  Destination host Unknown
  776  Source host isolated (obsolete)
  777  Destination network administratively prohibited
  777  Destination host administratively prohibited
  778  Network Unreachable for TOS
  779  Communication administratively prohibited by filtering
  780  Host precedence Violation
  781  Precedence cutoff in effect
 1024  Source quench
 1280  Redirect for network
 1281  Redirect for host
 1282  Redirect for type-of-service and network
 1283  Redirect for type-of-service and host
 2048  Echo request
 2304  Router advertisement
 2560  Router solicitation
 2816  Time-to-live equals 0 during transit
 2817  Time-to-live equals 0 during reassembly
 3072  IP header bad
 3073  Required option missing
 3328  Time stamp request
 3584  Time stamp reply
 3840  Information request (obsolete)
 4096  Information reply (obsolete)
 4352  Address mask request
 4608  Address mask reply

-D dumpperiod
Run in daemon mode. Instructs ipaudit to perform network summary every dumpperiod seconds. Ipaudit will spawn a new child process every dumpperiod seconds which will run for dumpperiod seconds before writing out data (as determined by the -w, -o and -x options). If the time formats are used as part of the file names, ipaudit will round the times to the nearest dumpperiod to the starting time. This way if you wish to dump data every 10 minutes, but start at 1:06, if you run the command

   ipaudit -o %H:%M.txt eth0

the first collection period will run from 1:06 to 1:20, and the firs file name will be "01:10.txt", and afterwards the collection periods will run every 10 minutes on the 10 minute interval. This simplifies the task of synchronizing the output file with a regular interval.
   Ipaudit currently has no option to append output with the -o or -w options. If two different collection periods write to the same file, the first data will be overwritten.

-E n
Quit after reading interface(s) for n seconds.

Ignore config file. See CONFIGURATION FILE section below.

Store host IPs only. On output the protocol and port fields will be set to zero.

-I ipaddr
Dump all packets to or from "ipaddr" if using -w option.

-L hostportlimit, hostlimit
Normally information is stored for every connection , which is a combination of host ip addresses, protocol and ports. Sometimes the traffic that is monitored will have a large number of connections, for example when scans are taking place where the ports numbers change repeatedly. hostportlimit sets a limit on the number of connections stored with unique port addresses. Otherwise ipaudit's hash table can overflow memory. When this limit is exceeded, ipaudit will set all ports to zero and store only host ips and protocols. Under some conditions even this precaution is not enough, for example when a Denial of Service attack where every packet has a unique forged source ip address. hostlimit sets a limit on the number of unique host pairs which are stored. When this limit is exceeded then every host ip is stored as Only the protocol number will remain unchanged.

By default there is no hostportlimit or hostlimit. You might want to set these limits. If so set them larger than your typical connection count. In our case the connection count never exceeds 200000 unless there is a DoS attack. The option -L500000,100000 works for us.

See discussion below READING MULTIPLE NETWORK INTERFACES. The -M option turns off ipaudit removal of duplicate packets.

-N n_hash_slots
ipaudit stores connections in a hash table. The number of slots in this table is set when the program starts. Each slot can hold multiple connections, and storage and retrieval become inefficient when there are too few slots. You can control the number of slots with this option. The default number of slots (typically 1,000,000) is printed when the -v option is given.

-O locip,remip
When using the -L option, ipaudit sets overflow packet's ip addresses to If you configured ipaudit (with -l) to classify ip addresses as local or remote, you can assign local and remote ip addresses to different fixed values. For example, with the option


your overflow packets can have local ip addresses (as defined by the -l option) set to, while remote addresses are labled as This way you can still determine how much traffic travels in and out of your network even when traffic volume is too large to record individual ip addresses. Note there can be no spaces within the argument.

This option adds the hostname as the first column in the output. This is useful to keep track of which host generated output if you have multiple collectors feeding into a common SQL database.

-R packet_interval
If saving selected packets (which involves the options -w and -p), then also save every packet_interval'th packet. This helps in later analysis of heavy traffic that was not pre-selected.

By default ipaudit pads the ip addresses with leading zeroes, for example

This option tells ipaudit to print ip addresses without leading zeroes,

Write out connection start and stop dates and times (in constrast to the -t option which only write out time but not dates).

-V vlan
Read only packets belonging to a specific vlan (802.1q tagging). If packets are not vlan tagged then this option is not used.

-W dumplimit
Limits the number of packets written to the dumpfile (see -w option). This is useful to prevent the dumpfile from becoming too large when traffic is high. A value of 0 (the default) means no limit.



In addition to command line options you can use a configuration file. When IPAUDIT starts it first looks for a file named "ipaudit.conf" in the current directory, then in the home directory. You can also give the '-g' (see COMMAND LINE OPTIONS above) to specify a configuration file.

The configuration can contain comments beginning with #.

The options in the config file are specified by keyword/value pairs. For example, to configure ipaudit to write the time information for each connection the option is

   writetime on

In keyword is 'writetime' and the option is 'on'. Many options are either ON/OFF. The words TRUE, YES and OK are synonymous with ON. Any other word means OFF.

Other options require one or more values, for example

   interface  eth0 eth1

tells ipaudit to read interfaces eth0 and eth1.

You can also use config file options from the command line with the '-z' command line option. While most commands support both option formats, command line options and config file options, some newer commands only have config file options. To execute these newer command from the command line you must invoke the -z option.

Below is a list of config file options:

Like -M option, turns off duplicate packet detection used when reading multiple interfaces.

count packet_limit
Like -c option, maximum number of packets to read.

daemon {period}
Like -D option, ipaudit goes into daemon mode, writes statistics every 'period' seconds.

ethernet {on/off}
Like -e option, prints ethernet addresses on each connection written. Default is off.

filter bpf_filter
Like -f option, specifies bpf filter.

hashslots nslots
Like the -N option, sets the number of slots in the hash table used by ipaudit.

hostip ip
Like -I option, tells ipaudit to write *all* packets (the first length bytes as set by -s or packetlen options) for the host ip, regardless of the setting for -p or saveport options.

hostonly {on/off}
Like -H option, Save and write information only for host pairs, not for "connections", host pairs, protocols and ports.

hostportlimit hostportlimit hostlimit
Like -L option, sets limit for number of unique host-pair/protocol/port connections, and a separate limit for the number of unique host-pair connections.

icmptype {on/off}
Like -C option, writes ICMP type and code information in port field of the sending IP address.

interface i1 [i2] [i3]
List interface(s) to read packets from.

localrange ip_range
Like -l option, determines which range(s) of ip addresses are considered local when writing ipaudit output.

mysql host user password database [table]
Ipaudit can write directly to a MySQL database if such support has been compiled. See the section MYSQL SUPPORT below for detailed information on the table format used. If support has not been compiled in then the program will print an error when started and stop. A simple test is to run the following

    ipaudit -z mysql

If you get the following ERROR then MySQL support was not compiled in.

    ERROR:  Cannot output to MySQL database as requested because
    MySQL support was not compiled into this instance of ipaudit.

The host option determines which machine the database resides on. Use 'localhost' if it is on the same machine. The user and password specified must have write permission in database. The optional table is the name of the table written to in the database. The table will be created if it does not already exist. The default table name is 'connections'.

The mysql option does not have a command line equivalent. Use the '-z' option if you want to invoke it from the command line, like this

    ipaudit -z 'mysql localhost ipaudit password testdb' eth0

outfile file_name
Like -o option, name of output file for ipaudit output

Like -O option, assigns label ip addresses to use for local and remote overflow traffic.

packetlen length
Like -s option, maximum number of bytes saved for each packet.

packetsample sample_interval
Like -R option, write out every sample_interval'th packet.

pidfile file_name
Like -i option, write out file with process id.

Like -p option, this option adds the hostname as the first column in the output.

probelabel label
This option is like probename above, but rather than print the hostname in the first column, it prints the specified label.

progfile program_name
Like -x option, gives program name to run after ipaudit completes.

promisc {on/off}
Like -m option, puts interface(s) in promiscuous mode if on. Default is on.

readfile file_name
Like -r option, gives name of pcap packet capture file to be read.

savefile file_name
Like -w option, synonymous with the writefile config file option above.

saveport protocol>[,port[,port]..][:<protocol[,port[,port]..]]..
Like -p option, limits which protocols/ports packets must have to be saved with writefile(savefile) option.

Like -S option, prints ip addresses without leading 0s.

Like -q option, outputs traffic data in SQL format.

vlan vlan#
Like -V option, read only vlan packets with id vlan#.

writefile file_name
Like -w option, gives name of pcap packet capture file to write.

Like -W option, limits number of packets written to the dumpfile.

writetime {on/off}
Like -t option, writes times with connection data.



ipaudit has the ability to read network data from more than one network interface at a time. You can specify multiple interfaces on the command line such as

   ipaudit  -t  eth0:eth1

with the interface names separated by colons (:). Also, in the config file you use the interface option such as

   interface eth0 eth1

There are some details to consider when monitoring multiple interfaces. In particular there is the issue of double counting packets that travel between interfaces. Consider the following example. Suppose you have a router with four interfaces A, B, C and D. Interface A connects to the Internet. Interfaces B, C and D connect to three separate internal networks called Net-B, Net-C and Net-D. Interfaces B, C, D are also mirrored (tapped) by three network interface on your computer running ipaudit. If Net-B sends a packet to Net-C ipaudit will see it twice. Once as it travels into the router through interface B, and a second time when it travels out of the router through interface C. If ipaudit simply recorded every packet on every interface, it would have two copies of every packet that travels between B and C. Similarly they would also be two copies of every packet sent between B and D and between C and D. This is a problem which ipaudit attempts to solve.

The solution is to record the interface for the first packet of each connection. Then, subsequent packets for this connection are recorded only if they are read from the same interface. In this way duplicate packets are ignored.

This technique should work fine under normal circumstances, but it fails when the hostportlimit is reached (see discussion about the -L option). When the hostportlimit is reached ipaudit stops recording ip addresses. In order to save memory it sets all ip address to The correct ip address is needed to successfully remove duplicate packet information. When this information is lost many different connections are stored together, making it impossible to remove duplicate packets by noting which interface they were read from. If you wish, you can run ipaudit without removing duplicate packets with the -M option.



Ipaudit can write its output directly into a MySQL table if such support has been compiled in (see the 'mysql' option under CONFIGURATION FILE above). When MySQL output is selected, the options -t and -p are automatically selected. This includes the packet first and last times and first and last talkers in the database table, as well as the hostname (or probelabel if that has been specified).

By default, the name of the MySQL table is 'connections', but this can be changed in the 'mysql' statement described above. The format of the MySQL table is as follows, where the values in the first column below are the names of the MySQL table columns, and the values second column is the MySQL data type used.

  NAME         DATA TYPE

  probe        varchar(12)
  local        int unsigned
  remote       int unsigned
  prot         tinyint unsigned
  lport        smallint unsigned
  rport        smallint unsigned
  incb         bigint unsigned
  outb         bigint unsigned
  incp         int unsigned
  outp         int unsigned
  sec1         int unsigned
  msec1        int unsigned
  sec2         int unsigned
  msec2        int unsigned
  talk1        tinyint unsigned
  talk2        tinyint unsigned

These columns hold values identical to ipaudit's standard output except fro the columns local, remote, sec1, msec1, sec2 and msec2.

The columns local and remote hold ipaddress as 4 byte integers. You can use the MySQL function INET_NTOA(LOCAL) to convert it to dotted quad notation, for example
   INET_NTOA(3221888773) -> "".

The column sec1 holds the first packet time in units of seconds since Jan 1, 1970 GMT, column msec1 holds the number of microseconds since the start of the last second interval. Columns sec2 and msec2 hold the corresponding data for the last packet time.

A convenient way to convert the columns sec1 and sec2 to a more standard date/time notation is to use the GNU version of the Unix 'data' utility as so,
   date -d "Jan 1, 1970 GMT  <sec> sec" where <sec> is replaced by the integer value of sec1 or sec2. For example,
   date -d "Jan 1, 1970 GMT 1054315274 sec" yields
   Fri May 30 13:21:14 EDT 2003



ipaudit -w dump.fil -p1:2:6,21,23 eth0

This runs tells ipaudit to write all packet header for icmp, telnet, and ftp to dump.fil and to read traffic from the network interface eth0. ipaudit will continue to run until it is interrupted with control-C or the kill command, at which time it will list on the terminal a summary of traffic that it saw.

ipaudit -i ipaudit.pid -o traffic.out le0

Read traffic from interface le0 and place in the file traffic.out when program completes. The process id number is written to ipaudit.pid at the program start. Thus you can terminate ipaudit with the command kill -2 `cat ipaudit.pid`

ipaudit -c 50000 eth0 > traffic.out

Read 50,000 packets of traffic from eth0, pipe default stdout output into file traffic.out.

ipaudit -w- -oipaudit.out eth0:eth1 | tcpdump -r- -wtest.cap port 23 or port 111 or net 10.1.4

Reads data from interfaces eth0 and eth0 while writing output to ipaudit.out. In addition raw packets are sent to standard out (-w- option) and are piped to tcpdump. The trick here is that tcpdump is used to write raw packets instead of ipaudit. The advantage is that tcpdump offers finer control than ipaudit's -p option over what packets are written to the capture file. In this example it is necessary to write the ipaudit output directly to file instead of to standard out (the default) so as not to interfere with the packets written to standard out. You could also use the BPF filters called by ipaudit to accomplish the same thing,

ipaudit -w- -oipaudit.out eth0:eth1 | ipaudit -r- -wtest.cap -f 'port 23 or port 111 or net 10.1.4'



without -t option 6 2076 139  0 58 0 1 17 520 520  0 1092 0 2 17 138 138  0 243 0 1 17 138 138  0 248 0 1 17 123 123  180 180 2 2

Columns are ip address for first (second) machine, protocol number (1 is icmp, 6 is tcp, 17 is udp), udp/tcp ports on first (second) machine (set to 0 if protocol not udp/tcp), bytes received by first (second) machine, number of packets received by first (second) machine.

with -t option 6 2076 139 0 58 0 1 13:38:59.8229 13:38:59.8229 1 1 17 520 520 0 1092 0 2 13:38:59.1435 13:38:59.9248 1 1 17 138 138 0 243 0 1 13:38:56.5373 13:38:56.5373 1 1 17 138 138 0 248 0 1 13:38:54.7191 13:38:54.7191 1 1 17 123 123 180 180 2 2 13:38:54.6649 13:38:57.5350 2 1

As above but with the addition of time the first (last) packet was detected, and the machine which sent the first (last) packet.



Sometimes under Linux when reading lo and eth0 interfaces simultaneously packets on the lo interface are dropped.

Hash table size is fixed at run-time. Would be nice to have dynamically adjusted hash table size.

Report any bugs to jon.rifkin@uconn.edu. If possible, run with -d option and email output. Thanks.



j rifkin jon.rifkin@uconn.edu



0.98 Feb 12, 2003



ipstrings(1) total(1) tcpdump(1) pcap(3)

*** host user password database

SourceForge.net Logo