Home > Articles > Operating Systems, Server > Linux/UNIX/Open Source

Security Tools

  • Print
  • + Share This
Is your toolbox getting pretty full from my last two articles on monitoring tools and troubleshooting tools? Well, make some room. This article covers six security tools that every administrator ought to know.
Pat Eyler is the author of Networking Linux: A Practical Guide to TCP/IP (New Riders, 2001, ISBN 0-7357-1031-7).
Like this article? We recommend

Monitoring Tools

This article presents six tools that should become part of your security toolkit. nmap is a port scanner that has become the de facto standard for system and network administrators. Nessus is a security scanner that has replaced the older SATAN as the tool of choice for catching known security problems. iptables is the interface to the next generation of Linux packet filtering and IP masquerading. Xinetd and tcp wrappers provide two methods of controlling access to a specific host. They can be used separately or in conjunction with one another. OPIE provides a password-obscuring mechanism to allow fairly secure logins without encrypting the session between two hosts.


nmap is a tool for scanning a machine (or machines) for security problems and is exercised from the network. nmap is often run from the command line, but a GTK+-based front end (nmapfe) is also available. nmap is written and maintained by Fyodor. nmapfe was originally written by Zach Smith and is now maintained by Fyodor.

nmap and related tools are something of a mixed blessing, and thus have a murky reputation. While these tools are incredibly useful for a system or network administrator, they can also be used to the advantage of system crackers. Some people would prefer that tools like nmap weren't made publicly available. I tend to side with the other part of the community. If tools like this weren't available to the good guys, it would give the bad guys an incredible advantage because they certainly won't give up their tools.

nmap is an external security scanner, or a port scanner. It works by sending IP packets to the host(s) it is checking and by watching for what kind of response (if any) it receives.

nmap presents a summary of the responses it receives to provide a security overview of the target host(s).This overview can show

  • A list of open ports

  • Owners of remote processes on open ports

  • RPC services matched against the port on which they are provided

  • Information about TCP sequence numbers

  • Remote operating systems

Getting and Installing nmap

Binaries and source of nmap are available from http://www.nmap.org. The binaries are as easy to install as you would expect. Building nmap and nmapfe is easy as well.

After putting the source tarball into /usr/local/source, you just follow the regular three step process:

[root@cherry nmap-2.53]# ./configure
[root@cherry nmap-2.53]# make
[root@cherry nmap-2.53]# make install

nmap at Work

After nmap has been built and installed, you're ready to run with it. nmap provides a number of different security-scanning options.

nmap on the Command Line

The simplest nmap scan is run as shown in Listing 1.

Listing 1 Running an nmap Scan

[root@cherry nmap-2.53]# nmap

Starting nmap V. 2.54BETA1 by fyodor@insecure.org
( http://www.insecure.org/nmap/ )
Interesting ports on (
(The 1519 ports scanned but not shown below are in state: closed)
Port State Service
25/tcp open smtp
111/tcp open sunrpc
113/tcp open auth
515/tcp open printer
939/tcp open unknown
1024/tcp open kdm

nmap run completed -- 1 IP address (1 host up) scanned in 17 seconds
[root@cherry nmap-2.53]#

Several different kinds of scans available through nmap are described in Table 1.

Table 1 Scans Available Through nmap

Scan Type



TCP connect() scan


The most basic form of scanning. This opens a connection to every potentially interesting port on the target machine. Any user can use this kind of scan. It is easily detected, with many connection messages showing up in the logs of the target machine.

TCP SYN scan


The "half-open" scan. This scan sends a TCP SYN packet as though it is trying to open the connection. If it receives a SYN-ACK response, it sends an immediate RST to shut down the connection. Because this scan doesn't open the connection, it is less likely to be logged. Only users with root privilege can send TCP SYN scans.

Stealth FIN


This scan attempts to pass through packet filters by sending a TCP FIN packet.

Christmas Tree


This scan attempts to pass through packet filters by sending a packet with FIN, URG, and PUSH flags set.



This scan attempts to pass through packet filters by sending a packet without any flags turned on.



This limits the scan to only conducting a ping sweep to look for connected systems. It does not do port scans.

UDP scan


This sends 0 byte UDP packets to each port on the target machine(s).

ACK scan


This scan is used to help check packet filters. An ACK packet with random acknowledgment and sequence numbers is sent. If nothing is returned, the port is marked as filtered.

Window scan


This scan is similar to the ACK scan, but depends on anomalies in the TCP window size handling of some OSes.

RPC scan


This scan checks all open ports found by other scan types and sends RPC NULL commands to see if they are RPC ports. If they are RPC ports, this scan attempts to determine what program and version number they serve.

FTP bounce scan

-b <ftp relay host>

This scan relives a historical foible of FTP servers. Older FTP servers were able to serve bounced FTP sessions; that is, they connected to another host to deliver data to you. By providing a relay host in the format username:password@server:port, you can use this FTP bounce (mis)feature to scan ports that might otherwise be protected.

In addition to the types of scans that nmap can run, a number of options modify its behavior. These options include timing, target identification, output, and others. Some of the more useful options are shown in Table 2.

Table 2 nmap Options




Tells nmap not to ping hosts before scanning. (This is used to scan hosts that sit behind packet filters that don't allow ICMP traffic.)


Uses TCP to look for hosts on the target network. ACK packets are sent, and nmap looks for RST packets to be returned. An optional PORT number can be given. It tells nmap which port to attempt its connections against.


Causes nmap to use SYN packets instead of ACK packets when checking for hosts on the target network (root users only).


Uses ICMP only when looking for hosts on the target network.


Uses both ICMP and TCP ACK to sweep the target network for hosts. This is the default behavior.


Causes nmap to attempt remote host identification based on the way the target system handles TCP packets containing certain types of errors.


Causes nmap to use an identd scan. This causes the owner of each server process to be shown as well.


Causes nmap to fragment its scanning packets, making it more difficult to block the scan with packet filters.


Puts nmap into verbose mode, causing it to display much more information about what it's doing.

-oN <logfile>

Writes output into a human readable logfile.

-oM <logfile>

Writes output into a machine-parsable logfile.

--resume <logfile&;

Resumes an incomplete scan from a logfile.

-iL <logfile>

Causes nmap to read input from a logfile instead of really scanning a host.

-g <portnumber>

Allows you to define the port nmap uses as its source port.

-p <port range>

Allows you to define the range of ports nmap will scan. If no range is given, nmap will scan all the ports listed in its own services file. A range of ports can be given in the following format: -p 20-30,79,6000-6010.

nmap allows you to list target IP addresses in one of four styles. For single hosts, you can write the IP address or hostname. For networks, you can write the number in slash notation using the CIDR-style network mask (for example, The most flexible form of target listing allows you to wildcard portions of the address (or list them as ranges). This allows you to look at specific hosts within a group of networks (for instance, if you know that all your routers use the first IP address of their Class C address, you can scan your internal routers with a target such as 192.168.*.1). Finally, you are also able to provide a list of hosts (in any of the previous styles).

  • + Share This
  • 🔖 Save To Your Account