This Security FAQ is a resource provided by:
Internet Security Systems, Inc.
2000 Miller Court West Tel: (770) 441-2531
Norcross, Georgia 30071 Fax: (770) 441-2431
– Internet Scanner … the most comprehensive “attack simulator”
To get the newest updates of Security files check the following services:
mail firstname.lastname@example.org with “send index” in message
ftp iss.net /pub/
This Sniffer FAQ will hopefully give administrators a clear understanding of
sniffing problems and hopefully possible solutions to follow up with. Sniffers
is one of the main causes of mass break-ins on the Internet today.
This FAQ will be broken down into:
* What a sniffer is and how it works
* Where are sniffers available
* How to detect if a machine is being sniffed
* Stopping sniffing attacks:
o Active hubs
o One-time password technology
o Non-promiscuous interfaces
What a sniffer is and how it works
Unlike telephone circuits, computer networks are shared communication channels.
It is simply too expensive to dedicate local loops to the switch (hub) for each
pair of communicating computers. Sharing means that computers can receive
information that was intended for other machines. To capture the information
going over the network is called sniffing.
Most popular way of connecting computers is through ethernet. Ethernet protocol
works by sending packet information to all the hosts on the same circuit. The
packet header contains the proper address of the destination machine. Only the
machine with the matching address is suppose to accept the packet. A machine
that is accepting all packets, no matter what the packet header says, is said
to be in promiscuous mode.
Because, in a normal networking environment, account and password information
is passed along ethernet in clear-text, it is not hard for an intruder once
they obtain root to put a machine into promiscuous mode and by sniffing,
compromise all the machines on the net.
Where are sniffers available
Sniffing is one of the most popular forms of attacks used by hackers. One
special sniffer, called Esniff.c, is very small, designed to work on Sunos, and
only captures the first 300 bytes of all telnet, ftp, and rlogin sessions. It
was published in Phrack, one of the most widely read freely available
underground hacking magazines. You can find Phrack on many FTP sites. Esniff.c
is also available on many FTP sites such as coombs.anu.edu.au:/pub/net/log.
You may want to run Esniff.c on an authorized network to quickly see how
effective it is in compromising local machines.
Other sniffers that are widely available which are intended to debug network
* Etherfind on SunOs4.1.x
* Snoop on Solaris 2.x and SunOs 4.1 (on ftp playground.sun.com)
* Tcpdump 3.0 uses bpf for a multitude of platforms.
* Packetman, Interman, Etherman, Loadman works on the following platforms:
SunOS, Dec-Mips, SGI, Alpha, and Solaris. It is available on
Packetman was designed to capture packets, while Interman, Etherman, and
Loadman monitor traffic of various kinds.
DOS based sniffers
* Gobbler for IBM DOS Machines
* ethdump v1.03
Available on ftp
* ethload v1.04
Companion utility to a ethernet monitor. Available on ftp
Commercial Sniffers are available at:
* Network General.
Network General produces a number of products. The most
important are the Expert Sniffer, which not only sniffs on the
wire, but also runs the packet through a high-performance expert
system, diagnosing problems for you. There is an extension onto
this called the “Distributed Sniffer System” that allows you to
put the console to the expert sniffer on you Unix workstation
and to distribute the collection agents at remote sites.
* Microsoft’s Net Monitor
” My commercial site runs many protocols on one wire – NetBeui,
IPX/SPX, TCP/IP, 802.3 protocols of various flavors, most
notably SNA. This posed a big problem when trying to find a
sniffer to examine the network problems we were having, since I
found that some sniffers that understood Ethernet II parse out
some 802.3 traffic as bad packets, and vice versa. I found that
the best protocol parser was in Microsoft’s Net Monitor product,
also known as Bloodhound in its earlier incarnations. It is able
to correctly identify such oddities as NetWare control packets,
NT NetBios name service broadcasts, etc, which etherfind on a
Sun simply registered as type 0000 packet broadcasts. It
requires MS Windows 3.1 and runs quite fast on a HP XP60 Pentium
box. Top level monitoring provides network statistics and
information on conversations by mac address (or hostname, if you
bother with an ethers file). Looking at tcpdump style details is
as simple as clicking on a conversation. The filter setup is
also one of the easiest to implement that I’ve seen, just click
in a dialog box on the hosts you want to monitor. The number of
bad packets it reports on my network is a tiny fraction of that
reported by other sniffers I’ve used. One of these other
sniffers in particular was reporting a large number of bad
packets with src mac addresses of aa:aa:aa:aa:aa:aa but I don’t
see them at all using the MS product. – Anonymous
How to detect a sniffer running.
To detect a sniffing device that only collects data and does not respond to any
of the information, requires physically checking all your ethernet connections
by walking around and checking the ethernet connections individually.
It is also impossible to remotely check by sending a packet or ping if a
machine is sniffing.
A sniffer running on a machine puts the interface into promiscuous mode, which
accepts all the packets. On some Unix boxes, it is possible to detect a
promiscuous interface. It is possible to run a sniffer in non-promiscuous mode,
but it will only capture sessions from the machine it is running on. It is also
possible for the intruder to do similiar capture of sessions by trojaning many
programs such as sh, telnet, rlogin, in.telnetd, and so on to write a log file
of what the user did. They can easily watch the tty and kmem devices as well.
These attacks will only compromise sessions coming from that one machine, while
promiscuous sniffing compromises all sessions on the ethernet.
For SunOs, NetBSD, and other possible BSD derived Unix systems, there is a
that will tell you information about all the interfaces and if they are in
promiscuous mode. DEC OSF/1 and IRIX and possible other OSes require the device
to be specified. One way to find out what interface is on the system, you can
# netstat -r
Destination Gateway Flags Refs Use Interface
default iss.net UG 1 24949 le0
localhost localhost UH 2 83 lo0
Then you can test for each interface by doing the following command:
inet 127.0.0.1 netmask 0xffffff00 broadcast 255.0.0.1
Intruders often replace commands such as ifconfig to avoid detection. Make sure
you verify its checksum.
There is a program called cpm available on ftp.cert.org:/pub/tools/cpm that
only works on Sunos and is suppose to check the interface for promiscuous flag.
Ultrix can possibly detect someone running a sniffer by using the commands
pfstat and pfconfig.
pfconfig allows you to set who can run a sniffer
pfstat shows you if the interface is in promiscuous mode.
These commands only work if sniffing is enabled by linking it into the kernel.
by default, the sniffer is not linked into the kernel. Most other Unix systems,
such as Irix, Solaris, SCO, etc, do not have any flags indication whether they
are in promiscuous mode or not, therefore an intruder could be sniffing your
whole network and there is no way to detect it.
Often a sniffer log becomes so large that the file space is all used up. On a
high volume network, a sniffer will create a large load on the machine. These
sometimes trigger enough alarms that the administrator will discover a sniffer.
I highly suggest using lsof (LiSt Open Files) available from
coast.cs.purdue.edu:/pub/Purdue/lsof for finding log files and finding programs
that are accessing the packet device such as /dev/nit on SunOs.
There is no commands I know of to detect a promiscuous IBM PC compatible
machine, but they atleast usually do not allow command execution unless from
the console, therefore remote intruders can not turn a PC machine into a
sniffer without inside assistance.
Stopping sniffing attacks
Active hubs send to each system only packets intended for it rendering
promiscuous sniffing useless. This is only effective for 10-Base T.
The following vendors have available active hubs:
There are several packages out there that allow encryption between connections
therefore an intruder could capture the data, but could not decypher it to make
any use of it.
Some packages available are:
* deslogin is one package available at ftp
* swIPe is another package available at
* Netlock encrypts all (tcp, udp, and raw ip based) communications
transparently. It has automatic (authenticated Diffie-Helman) distibuted
key management mechanism for each host and runs on the SUN 4.1 and HP 9.x
systems. The product comes with a Certification Authority Management
application which generates host certificates (X.509) used for
authentication between the hosts. and provides centralized control of each
Hosts communications rules.
The product is built by Hughes Aircraft and they can be reached at
800-825-LOCK or email at email@example.com.
Kerberos is another package that encrypts account information going over the
network. Some of its draw backs are that all the account information is held on
one host and if that machine is compromised, the whole network is vulnerable.
It is has been reported a major difficulty to set up. Kerberos comes with a
stream-encrypting rlogind, and stream-encrypting telnetd is available. This
prevents intruders from capturing what you did after you logged in.
There is a Kerberos FAQ at ftp at rtfm.mit.edu in
One time password technology
S/key and other one time password technology makes sniffing account information
almost useless. S/key concept is having your remote host already know a
password that is not going to go over insecure channels and when you connect,
you get a challenge. You take the challenge information and password and plug
it into an algorithm which generates the response that should get the same
answer if the password is the same on the both sides. Therefore the password
never goes over the network, nor is the same challenge used twice. Unlike
SecureID or SNK, with S/key you do not share a secret with the host. S/key is
available on ftp:thumper.bellcore.com:/pub/nmh/skey
Other one time password technology is card systems where each user gets a card
that generates numbers that allow access to their account. Without the card, it
is improbable to guess the numbers.
The following are companies that offer solutions that are provide better
password authenication (ie, handheld password devices):
Secure Net Key (SNK)
Digital Pathways, Inc.
201 Ravendale Dr. Mountainview, Ca.
Phone: 415-964-0707 Fax: (415) 961-7487
One Alewife Center
Cambridge, MA 02140-2312
USA Phone: 617-547-7820
Fax: (617) 354-8836
Secure ID uses time slots as authenication rather than challenge/response.
ArKey and OneTime Pass
PO Box 1480
Hudson, OH 44236
Tel:US+216-686-0090 Fax: US+216-686-0092
OneTime Pass (OTP):
This program provides unrestricted one-time pass codes on a user by user basis
without any need for cryptographic protocols or hardware devices. The user
takes a list of usable pass codes and scratches out each one as it is used. The
system tracks usage, removing each passcode from the available list when it is
used. Comes with a very small and fast password tester and password and pass
phrase generation systems.
This is the original Argued Key system that mutually authenticates users and
systems to each other based on their common knowledge. No hardware necessary.
Comes with a very small and fast password tester and password and pass phrase
WatchWord and WatchWord II
480 Spring Park Place
Herndon, VA 22070
1-800-521-6261 ext 217
Arnold Consulting, Inc.
2530 Targhee Street, Madison, Wisconsin
Phone : 608-278-7700 Fax: 608-278-7701
CRYPTOCard is a modern, SecureID-sized, SNK-compatible device.
Enigma Logic, Inc.
2151 Salvio #301
Concord, CA 94520
510-827-5707 Fax: (510)827-2593
For information about Enigma ftp to: ftp.netcom.com in directory
Secure Computing Corporation:
2675 Long Lake Road
Roseville, MN 55113
Tel: (612) 628-2700
Fax: (612) 628-2701
You can try to make sure that most IBM DOS compatible machines have interfaces
that will not allow sniffing. Here is a list of cards that do not support
Test the interface for promiscuous mode by using the Gobbler. If you find a
interface that does do promiscuous mode and it is listed here, please e-mail
firstname.lastname@example.org so I can remove it ASAP.
IBM Token-Ring Network PC Adapter
IBM Token-Ring Network PC Adapter II (short card)
IBM Token-Ring Network PC Adapter II (long card)
IBM Token-Ring Network 16/4 Adapter
IBM Token-Ring Network PC Adapter/A
IBM Token-Ring Network 16/4 Adapter/A
IBM Token-Ring Network 16/4 Busmaster Server Adapter/A
The following cards are rumoured to be unable to go into promiscuous mode, but
that the veracity of those rumours is doubtful.
Microdyne (Excelan) EXOS 205
Microdyne (Excelan) EXOS 205T
Microdyne (Excelan) EXOS 205T/16
Hewlett-Packard 27250A EtherTwist PC LAN Adapter Card/8
Hewlett-Packard 27245A EtherTwist PC LAN Adapter Card/8
Hewlett-Packard 27247A EtherTwist PC LAN Adapter Card/16
Hewlett-Packard 27248A EtherTwist EISA PC LAN Adapter Card/32
HP 27247B EtherTwist Adapter Card/16 TP Plus
HP 27252A EtherTwist Adapter Card/16 TP Plus
HP J2405A EtherTwist PC LAN Adapter NC/16 TP
Adapters based upon the TROPIC chipset generally do not support promiscuous
mode. The TROPIC chipset is used in IBM’s Token Ring adapters such as the 16/4
adapter. Other vendors (notably 3Com) also supply TROPIC based adapters.
TROPIC-based adapters do accept special EPROMs, however, that will allow them
to go into promiscuous mode. However, when in promiscuous mode, these adapters
will spit out a “Trace Tool Present” frame.
I would like to thank the following people for the contribution to this FAQ
that has helped to update and shape it:
* Padgett Peterson (email@example.com)
* Steven Bellovin (firstname.lastname@example.org)
* Wietse Venema (email@example.com)
* Robert D. Graham (robg@NGC.COM)
* Kevin Martinez (firstname.lastname@example.org)
* Frederick B. Cohen (email@example.com)
* James Bonfield (firstname.lastname@example.org)
* Marc Horowitz (marc@MIT.EDU)
* Steve Edwards (email@example.com)
* Andy Poling (Andy.Poling@jhu.edu)
* Jeff Collyer (firstname.lastname@example.org)
* Sara Gordon (email@example.com)
This paper is Copyright (c) 1994, 1995
by Christopher Klaus of Internet Security Systems, Inc.
Permission is hereby granted to give away free copies electronically. You may
distribute, transfer, or spread this paper electronically. You may not pretend
that you wrote it. This copyright notice must be maintained in any copy made.
If you wish to reprint the whole or any part of this paper in any other medium
(ie magazines, books, etc) excluding electronic medium, please ask the author
The information within this paper may change without notice. Use of this
information constitutes acceptance for use in an AS IS condition. There are NO
warranties with regard to this information. In no event shall the author be
liable for any damages whatsoever arising out of or in connection with the use
or spread of this information. Any use of this information is at the user’s own
Address of Author
Please send suggestions, updates, and comments to:
Christopher Klaus <firstname.lastname@example.org> of Internet Security Systems, Inc.
Internet Security Systems, Inc.
Internet Security Systems, Inc, located in Atlanta, Ga., specializes in the
developement of security scanning software tools. Its flagship product,
Internet Scanner, is software that learns an organization’s network and probes
every device on that network for security holes. It is the most comprehensive
“attack simulator” available, checking for over 100 security vulnerabilities.
Christopher William Klaus Voice: (404)441-2531. Fax: (404)441-2431
Internet Security Systems, Inc. Computer Security Consulting
2000 Miller Court West, Norcross, GA 30071