The output from Nmap is a list of scanned targets, with supplemental information on each depending on the options used. Key among that information is the "interesting ports table". That table lists the port number and protocol, service name, and state. The state is either open, filtered, closed, or unfiltered. Open means that an application on the target machine is listening for connections/packets on that port. Filtered means that a firewall, filter, or other network obstacle is blocking the port so that Nmap cannot tell whether it is open or closed. Closed ports have no application listening on them, though they could open up at any time. Ports are classified as unfiltered when they are responsive to Nmap's probes, but Nmap cannot determine whether they are open or closed. Nmap reports the state combinations open|filtered and closed|filtered when it cannot determine which of the two states describe a port. The port table may also include software version details when version detection has been requested. When an IP protocol scan is requested (-sO), Nmap provides information on supported IP protocols rather than listening ports.
In addition to the interesting ports table, Nmap can provide further information on targets, including reverse DNS names, operating system guesses, device types, and MAC addresses.
A typical Nmap scan is shown in Example 14.1, "A representative Nmap scan". The only Nmap arguments used in this example are -A, to enable OS and version detection, script scanning, and traceroute; -T4 for faster execution; and then the two target hostnames.
Example 14.1. A representative Nmap scan
# nmap -A -T4 scanme.nmap.org playground Starting nmap ( http://insecure.org/nmap/ ) Interesting ports on scanme.nmap.org (205.217.153.62): (The 1663 ports scanned but not shown below are in state: filtered) PORT STATE SERVICE VERSION 22/tcp open ssh OpenSSH 3.9p1 (protocol 1.99) 53/tcp open domain 70/tcp closed gopher 80/tcp open http Apache httpd 2.0.52 ((Fedora)) 113/tcp closed auth Device type: general purpose Running: Linux 2.4.X|2.5.X|2.6.X OS details: Linux 2.4.7 - 2.6.11, Linux 2.6.0 - 2.6.11 Uptime 33.908 days (since Thu Jul 21 03:38:03 2005) Interesting ports on playground.nmap.org (192.168.0.40): (The 1659 ports scanned but not shown below are in state: closed) PORT STATE SERVICE VERSION 135/tcp open msrpc Microsoft Windows RPC 139/tcp open netbios-ssn 389/tcp open ldap? 445/tcp open microsoft-ds Microsoft Windows XP microsoft-ds 1002/tcp open windows-icfw? 1025/tcp open msrpc Microsoft Windows RPC 1720/tcp open H.323/Q.931 CompTek AquaGateKeeper 5800/tcp open vnc-http RealVNC 4.0 (Resolution 400x250; VNC port: 5900) 5900/tcp open vnc VNC (protocol 3.8) MAC Address: 00:A0:CC:63:85:4B (Lite-on Communications) Device type: general purpose Running: Microsoft Windows NT/2K/XP OS details: Microsoft Windows XP Pro RC1+ through final release Service Info: OSs: Windows, Windows XP Nmap finished: 2 IP addresses (2 hosts up) scanned in 88.392 seconds
The newest version of Nmap can be obtained from http://insecure.org/nmap/. The newest version of the man page is available from http://insecure.org/nmap/man/.
This options summary is printed when Nmap is run with no arguments, and the latest version is always available at http://insecure.org/nmap/data/nmap.usage.txt. It helps people remember the most common options, but is no substitute for the in-depth documentation in the rest of this manual. Some obscure options aren't even included here.
Nmap 4.50 ( http://insecure.org ) Usage: nmap [Scan Type(s)] [Options] {target specification} TARGET SPECIFICATION: Can pass hostnames, IP addresses, networks, etc. Ex: scanme.nmap.org, microsoft.com/24, 192.168.0.1; 10.0.0-255.1-254 -iL <inputfilename>: Input from list of hosts/networks -iR <num hosts>: Choose random targets --exclude <host1[,host2][,host3],...>: Exclude hosts/networks --excludefile <exclude_file>: Exclude list from file HOST DISCOVERY: -sL: List Scan - simply list targets to scan -sP: Ping Scan - go no further than determining if host is online -PN: Treat all hosts as online -- skip host discovery -PS/PA/PU [portlist]: TCP SYN/ACK or UDP discovery to given ports -PE/PP/PM: ICMP echo, timestamp, and netmask request discovery probes -PO [protocol list]: IP Protocol Ping -n/-R: Never do DNS resolution/Always resolve [default: sometimes] --dns-servers <serv1[,serv2],...>: Specify custom DNS servers --system-dns: Use OS's DNS resolver SCAN TECHNIQUES: -sS/sT/sA/sW/sM: TCP SYN/Connect()/ACK/Window/Maimon scans -sU: UDP Scan -sN/sF/sX: TCP Null, FIN, and Xmas scans --scanflags <flags>: Customize TCP scan flags -sI <zombie host[:probeport]>: Idle scan -sO: IP protocol scan -b <FTP relay host>: FTP bounce scan --traceroute: Trace hop path to each host --reason: Display the reason a port is in a particular state PORT SPECIFICATION AND SCAN ORDER: -p <port ranges>: Only scan specified ports Ex: -p22; -p1-65535; -p U:53,111,137,T:21-25,80,139,8080 -F: Fast mode - Scan fewer ports than the default scan -r: Scan ports consecutively - don't randomize --top-ports <number>: Scan <number> most common ports --port-ratio <ratio>: Scan ports more common than <ratio> SERVICE/VERSION DETECTION: -sV: Probe open ports to determine service/version info --version-intensity <level>: Set from 0 (light) to 9 (try all probes) --version-light: Limit to most likely probes (intensity 2) --version-all: Try every single probe (intensity 9) --version-trace: Show detailed version scan activity (for debugging) SCRIPT SCAN: -sC: equivalent to --script=safe,intrusive --script=<Lua scripts>: <Lua scripts> is a comma separated list of directories, script-files or script-categories --script-args=<n1=v1,[n2=v2,...]>: provide arguments to scripts --script-trace: Show all data sent and received --script-updatedb: Update the script database. OS DETECTION: -O: Enable OS detection --osscan-limit: Limit OS detection to promising targets --osscan-guess: Guess OS more aggressively TIMING AND PERFORMANCE: Options which take <time> are in milliseconds, unless you append 's' (seconds), 'm' (minutes), or 'h' (hours) to the value (e.g. 30m). -T[0-5]: Set timing template (higher is faster) --min-hostgroup/max-hostgroup <size>: Parallel host scan group sizes --min-parallelism/max-parallelism <time>: Probe parallelization --min-rtt-timeout/max-rtt-timeout/initial-rtt-timeout <time>: Specifies probe round trip time. --max-retries <tries>: Caps number of port scan probe retransmissions. --host-timeout <time>: Give up on target after this long --scan-delay/--max-scan-delay <time>: Adjust delay between probes FIREWALL/IDS EVASION AND SPOOFING: -f; --mtu <val>: fragment packets (optionally w/given MTU) -D <decoy1,decoy2[,ME],...>: Cloak a scan with decoys -S <IP_Address>: Spoof source address -e <iface>: Use specified interface -g/--source-port <portnum>: Use given port number --data-length <num>: Append random data to sent packets --ip-options <options>: Send packets with specified ip options --ttl <val>: Set IP time-to-live field --spoof-mac <mac address/prefix/vendor name>: Spoof your MAC address --badsum: Send packets with a bogus TCP/UDP checksum OUTPUT: -oN/-oX/-oS/-oG <file>: Output scan in normal, XML, s|<rIpt kIddi3, and Grepable format, respectively, to the given filename. -oA <basename>: Output in the three major formats at once -v: Increase verbosity level (use twice for more effect) -d[level]: Set or increase debugging level (Up to 9 is meaningful) --open: Only show open (or possibly open) ports --packet-trace: Show all packets sent and received --iflist: Print host interfaces and routes (for debugging) --log-errors: Log errors/warnings to the normal-format output file --append-output: Append to rather than clobber specified output files --resume <filename>: Resume an aborted scan --stylesheet <path/URL>: XSL stylesheet to transform XML output to HTML --webxml: Reference stylesheet from Insecure.Org for more portable XML --no-stylesheet: Prevent associating of XSL stylesheet w/XML output MISC: -6: Enable IPv6 scanning -A: Enables OS detection and Version detection, Script scanning and Traceroute --datadir <dirname>: Specify custom Nmap data file location --send-eth/--send-ip: Send using raw ethernet frames or IP packets --privileged: Assume that the user is fully privileged --unprivileged: Assume the user lacks raw socket privileges -V: Print version number -h: Print this help summary page. EXAMPLES: nmap -v -A scanme.nmap.org nmap -v -sP 192.168.0.0/16 10.0.0.0/8 nmap -v -iR 10000 -PN -p 80 SEE THE MAN PAGE FOR MANY MORE OPTIONS, DESCRIPTIONS, AND EXAMPLES
Everything on the Nmap command-line that isn't an option (or option argument) is treated as a target host specification. The simplest case is to specify a target IP address or hostname for scanning.
Sometimes you wish to scan a whole network of adjacent hosts. For this, Nmap supports CIDR-style addressing. You can append /numbits to an IP address or hostname and Nmap will scan every IP address for which the first numbits are the same as for the reference IP or hostname given. For example, 192.168.10.0/24 would scan the 256 hosts between 192.168.10.0 (binary: 11000000 10101000 00001010 00000000) and 192.168.10.255 (binary: 11000000 10101000 00001010 11111111), inclusive. 192.168.10.40/24 would do exactly the same thing. Given that the host scanme.nmap.org is at the IP address 205.217.153.62, the specification scanme.nmap.org/16 would scan the 65,536 IP addresses between 205.217.0.0 and 205.217.255.255. The smallest allowed value is /1, which scans half the Internet. The largest value is 32, which scans just the named host or IP address because all address bits are fixed.
CIDR notation is short but not always flexible enough. For example, you might want to scan 192.168.0.0/16 but skip any IPs ending with .0 or .255 because they are commonly broadcast addresses. Nmap supports this through octet range addressing. Rather than specify a normal IP address, you can specify a comma separated list of numbers or ranges for each octet. For example, 192.168.0-255.1-254 will skip all addresses in the range that end in .0 and or .255. Ranges need not be limited to the final octets: the specifier 0-255.0-255.13.37 will perform an Internet-wide scan for all IP addresses ending in 13.37. This sort of broad sampling can be useful for Internet surveys and research.
IPv6 addresses can only be specified by their fully qualified IPv6 address or hostname. CIDR and octet ranges aren't supported for IPv6 because they are rarely useful.
Nmap accepts multiple host specifications on the command line, and they don't need to be the same type. The command nmap scanme.nmap.org 192.168.0.0/16 10.0.0,1,3-7.0-255 does what you would expect.
While targets are usually specified on the command lines, the following options are also available to control target selection:
-iL <inputfilename> (Input from list)
-iR <num hosts> (Choose random targets)
--exclude <host1[,host2][,host3],...> (Exclude hosts/networks)
--excludefile <exclude_file> (Exclude list from file)
One of the very first steps in any network reconnaissance mission is to reduce a (sometimes huge) set of IP ranges into a list of active or interesting hosts. Scanning every port of every single IP address is slow and usually unnecessary. Of course what makes a host interesting depends greatly on the scan purposes. Network administrators may only be interested in hosts running a certain service, while security auditors may care about every single device with an IP address. An administrator may be comfortable using just an ICMP ping to locate hosts on his internal network, while an external penetration tester may use a diverse set of dozens of probes in an attempt to evade firewall restrictions.
Because host discovery needs are so diverse, Nmap offers a wide variety of options for customizing the techniques used. Host discovery is sometimes called ping scan, but it goes well beyond the simple ICMP echo request packets associated with the ubiquitous ping tool. Users can skip the ping step entirely with a list scan (-sL) or by disabling ping (-PN), or engage the network with arbitrary combinations of multi-port TCP SYN/ACK, UDP, and ICMP probes. The goal of these probes is to solicit responses which demonstrate that an IP address is actually active (is being used by a host or network device). On many networks, only a small percentage of IP addresses are active at any given time. This is particularly common with RFC1918-blessed private address space such as 10.0.0.0/8. That network has 16 million IPs, but I have seen it used by companies with less than a thousand machines. Host discovery can find those machines in a sparsely allocated sea of IP addresses.
If no host discovery options are given, Nmap sends a TCP ACK packet destined for port 80 and an ICMP echo request query to each target machine. An exception to this is that an ARP scan is used for any targets which are on a local ethernet network. For unprivileged Unix shell users, a SYN packet is sent instead of the ack using the connect() system call. These defaults are equivalent to the -PA -PE options. This host discovery is often sufficient when scanning local networks, but a more comprehensive set of discovery probes is recommended for security auditing.
The -P* options (which select ping types) can be combined. You can increase your odds of penetrating strict firewalls by sending many probe types using different TCP ports/flags and ICMP codes. Also note that ARP discovery (-PR) is done by default against targets on a local ethernet network even if you specify other -P* options, because it is almost always faster and more effective.
By default, Nmap does host discovery and then performs a port scan against each host it determines is online. This is true even if you specify non-default host discovery types such as UDP probes (-PU). Read about the -sP option to learn how to perform only host discovery, or use -PN to skip host discovery and port scan all target hosts. The following options control host discovery:
-sL (List Scan)
Since the idea is to simply print a list of target hosts, options for higher level functionality such as port scanning, OS detection, or ping scanning cannot be combined with this. If you wish to disable ping scanning while still performing such higher level functionality, read up on the -PN option.
-sP (Ping Scan)
Systems administrators often find this option valuable as well. It can easily be used to count available machines on a network or monitor server availability. This is often called a ping sweep, and is more reliable than pinging the broadcast address because many hosts do not reply to broadcast queries.
The -sP option sends an ICMP echo request and a TCP packet to port 80 by default. When executed by an unprivileged user, only a SYN packet is sent (using a connect() call) to port 80 on the target. When a privileged user tries to scan targets on a local ethernet network, ARP requests (-PR) are used unless --send-ip was specified. The -sP option can be combined with any of the discovery probe types (the -P* options, excluding -PN) for greater flexibility. If any of those probe type and port number options are used, the default probes (ACK and echo request) are overridden. When strict firewalls are in place between the source host running Nmap and the target network, using those advanced techniques is recommended. Otherwise hosts could be missed when the firewall drops probes or their responses.
-PN (No ping)
-PS [portlist] (TCP SYN Ping)
The SYN flag suggests to the remote system that you are attempting to establish a connection. Normally the destination port will be closed, and a RST (reset) packet sent back. If the port happens to be open, the target will take the second step of a TCP 3-way-handshake by responding with a SYN/ACK TCP packet. The machine running Nmap then tears down the nascent connection by responding with a RST rather than sending an ACK packet which would complete the 3-way-handshake and establish a full connection. The RST packet is sent by the kernel of the machine running Nmap in response to the unexpected SYN/ACK, not by Nmap itself.
Nmap does not care whether the port is open or closed. Either the RST or SYN/ACK response discussed previously tell Nmap that the host is available and responsive.
On Unix boxes, only the privileged user root is generally able to send and receive raw TCP packets. For unprivileged users, a workaround is automatically employed whereby the connect() system call is initiated against each target port. This has the effect of sending a SYN packet to the target host, in an attempt to establish a connection. If connect() returns with a quick success or an ECONNREFUSED failure, the underlying TCP stack must have received a SYN/ACK or RST and the host is marked available. If the connection attempt is left hanging until a timeout is reached, the host is marked as down. This workaround is also used for IPv6 connections, as raw IPv6 packet building support is not yet available in Nmap.
-PA [portlist] (TCP ACK Ping)
The -PA option uses the same default port as the SYN probe (80) and can also take a list of destination ports in the same format. If an unprivileged user tries this, or an IPv6 target is specified, the connect() workaround discussed previously is used. This workaround is imperfect because connect() is actually sending a SYN packet rather than an ACK.
The reason for offering both SYN and ACK ping probes is to maximize the chances of bypassing firewalls. Many administrators configure routers and other simple firewalls to block incoming SYN packets except for those destined for public services like the company web site or mail server. This prevents other incoming connections to the organization, while allowing users to make unobstructed outgoing connections to the Internet. This non-stateful approach takes up few resources on the firewall/router and is widely supported by hardware and software filters. The Linux Netfilter/iptables firewall software offers the --syn convenience option to implement this stateless approach. When stateless firewall rules such as this are in place, SYN ping probes (-PS) are likely to be blocked when sent to closed target ports. In such cases, the ACK probe shines as it cuts right through these rules.
Another common type of firewall uses stateful rules that drop unexpected packets. This feature was initially found mostly on high-end firewalls, though it has become much more common over the years. The Linux Netfilter/iptables system supports this through the --state option, which categorizes packets based on connection state. A SYN probe is more likely to work against such a system, as unexpected ACK packets are generally recognized as bogus and dropped. A solution to this quandary is to send both SYN and ACK probes by specifying -PS and -PA.
-PU [portlist] (UDP Ping)
Upon hitting a closed port on the target machine, the UDP probe should elicit an ICMP port unreachable packet in return. This signifies to Nmap that the machine is up and available. Many other types of ICMP errors, such as host/network unreachables or TTL exceeded are indicative of a down or unreachable host. A lack of response is also interpreted this way. If an open port is reached, most services simply ignore the empty packet and fail to return any response. This is why the default probe port is 31338, which is highly unlikely to be in use. A few services, such as chargen, will respond to an empty UDP packet, and thus disclose to Nmap that the machine is available.
The primary advantage of this scan type is that it bypasses firewalls and filters that only screen TCP. For example, I once owned a Linksys BEFW11S4 wireless broadband router. The external interface of this device filtered all TCP ports by default, but UDP probes would still elicit port unreachable messages and thus give away the device.
-PE; -PP; -PM (ICMP Ping Types)
While echo request is the standard ICMP ping query, Nmap does not stop there. The ICMP standard (RFC 792[2]) also specifies timestamp request, information request, and address mask request packets as codes 13, 15, and 17, respectively. While the ostensible purpose for these queries is to learn information such as address masks and current times, they can easily be used for host discovery. A system that replies is up and available. Nmap does not currently implement information request packets, as they are not widely supported. RFC 1122 insists that "a host SHOULD NOT implement these messages". Timestamp and address mask queries can be sent with the -PP and -PM options, respectively. A timestamp reply (ICMP code 14) or address mask reply (code 18) discloses that the host is available. These two queries can be valuable when administrators specifically block echo request packets while forgetting that other ICMP queries can be used for the same purpose.
-PO [protolist] (IP Protocol Ping)
This host discovery method looks for responses in the same protocol as the probes, or ICMP Protocol Unreachable messages which signify the specified IP protocol isn't supported on the host (which gives away that it's up).
-PR (ARP Ping)
ARP scan puts Nmap and its optimized algorithms in charge of ARP requests. And if it gets a response back, Nmap doesn't even need to worry about the IP-based ping packets since it already knows the host is up. This makes ARP scan much faster and more reliable than IP-based scans. So it is done by default when scanning ethernet hosts that Nmap detects are on a local ethernet network. Even if different ping types (such as -PE or -PS) are specified, Nmap uses ARP instead for any of the targets which are on the same LAN. If you absolutely don't want to do an ARP scan, specify --send-ip.
--traceroute (Trace path to host)
Traceroute works by sending packets with a low TTL (time-to-live) in an attempt to elicit ICMP Time Exceeded messages from intermediate hops between the scanner and the target host. Standard traceroute implementation start with a TTL of 1 and increment the TTL until the destination host is reached. Nmap's traceroute starts with a high TTL and then decrements the TTL until it reaches 0. Doing it backwards lets nmap employ clever caching algorithms to speed up traces over multiple hosts. On average nmap sends 5-10 fewer packets per host, depending on network conditions. If a single subnet is being scanned (i.e. 192.168.0.0/24) nmap may only have to send a single packet to most hosts.
--reason (Host and port state reasons)
-n (No DNS resolution)
-R (DNS resolution for all targets)
--system-dns (Use system DNS resolver)
--dns-servers <server1[,server2],...> (Servers to use for reverse DNS queries)
This option also comes in handy when scanning private networks. Sometimes only a few name servers provide proper rDNS information, and you may not even know where they are. You can scan the network for port 53 (perhaps with version detection), then try Nmap list scans (-sL) specifying each name server one at a time with --dns-servers until you find one which works.
While Nmap has grown in functionality over the years, it began as an efficient port scanner, and that remains its core function. The simple command nmap target scans more than 1660 TCP ports on the host target. While many port scanners have traditionally lumped all ports into the open or closed states, Nmap is much more granular. It divides ports into six states: open, closed, filtered, unfiltered, open|filtered, or closed|filtered.
These states are not intrinsic properties of the port itself, but describe how Nmap sees them. For example, an Nmap scan from the same network as the target may show port 135/tcp as open, while a scan at the same time with the same options from across the Internet might show that port as filtered.
The six port states recognized by Nmap
open
closed
filtered
unfiltered
open|filtered
closed|filtered
As a novice performing automotive repair, I can struggle for hours trying to fit my rudimentary tools (hammer, duct tape, wrench, etc.) to the task at hand. When I fail miserably and tow my jalopy to a real mechanic, he invariably fishes around in a huge tool chest until pulling out the perfect gizmo which makes the job seem effortless. The art of port scanning is similar. Experts understand the dozens of scan techniques and choose the appropriate one (or combination) for a given task. Inexperienced users and script kiddies, on the other hand, try to solve every problem with the default SYN scan. Since Nmap is free, the only barrier to port scanning mastery is knowledge. That certainly beats the automotive world, where it may take great skill to determine that you need a strut spring compressor, then you still have to pay thousands of dollars for it.
Most of the scan types are only available to privileged users. This is because they send and receive raw packets, which requires root access on Unix systems. Using an administrator account on Windows is recommended, though Nmap sometimes works for unprivileged users on that platform when WinPcap has already been loaded into the OS. Requiring root privileges was a serious limitation when Nmap was released in 1997, as many users only had access to shared shell accounts. Now, the world is different. Computers are cheaper, far more people have always-on direct Internet access, and desktop Unix systems (including Linux and Mac OS X) are prevalent. A Windows version of Nmap is now available, allowing it to run on even more desktops. For all these reasons, users have less need to run Nmap from limited shared shell accounts. This is fortunate, as the privileged options make Nmap far more powerful and flexible.
While Nmap attempts to produce accurate results, keep in mind that all of its insights are based on packets returned by the target machines (or firewalls in front of them). Such hosts may be untrustworthy and send responses intended to confuse or mislead Nmap. Much more common are non-RFC-compliant hosts that do not respond as they should to Nmap probes. FIN, null, and Xmas scans are particularly susceptible to this problem. Such issues are specific to certain scan types and so are discussed in the individual scan type entries.
This section documents the dozen or so port scan techniques supported by Nmap. Only one method may be used at a time, except that UDP scan (-sU) may be combined with any one of the TCP scan types. As a memory aid, port scan type options are of the form -sC, where C is a prominent character in the scan name, usually the first. The one exception to this is the deprecated FTP bounce scan (-b). By default, Nmap performs a SYN Scan, though it substitutes a connect scan if the user does not have proper privileges to send raw packets (requires root access on Unix) or if IPv6 targets were specified. Of the scans listed in this section, unprivileged users can only execute connect and FTP bounce scans.
-sS (TCP SYN scan)
This technique is often referred to as half-open scanning, because you don't open a full TCP connection. You send a SYN packet, as if you are going to open a real connection and then wait for a response. A SYN/ACK indicates the port is listening (open), while a RST (reset) is indicative of a non-listener. If no response is received after several retransmissions, the port is marked as filtered. The port is also marked filtered if an ICMP unreachable error (type 3, code 1,2, 3, 9, 10, or 13) is received.
-sT (TCP connect scan)
When SYN scan is available, it is usually a better choice. Nmap has less control over the high level connect() call than with raw packets, making it less efficient. The system call completes connections to open target ports rather than performing the half-open reset that SYN scan does. Not only does this take longer and require more packets to obtain the same information, but target machines are more likely to log the connection. A decent IDS will catch either, but most machines have no such alarm system. Many services on your average Unix system will add a note to syslog, and sometimes a cryptic error message, when Nmap connects and then closes the connection without sending data. Truly pathetic services crash when this happens, though that is uncommon. An administrator who sees a bunch of connection attempts in her logs from a single system should know that she has been connect scanned.
-sU (UDP scans)
UDP scan is activated with the -sU option. It can be combined with a TCP scan type such as SYN scan (-sS) to check both protocols during the same run.
UDP scan works by sending an empty (no data) UDP header to every targeted port. If an ICMP port unreachable error (type 3, code 3) is returned, the port is closed. Other ICMP unreachable errors (type 3, codes 1, 2, 9, 10, or 13) mark the port as filtered. Occasionally, a service will respond with a UDP packet, proving that it is open. If no response is received after retransmissions, the port is classified as open|filtered. This means that the port could be open, or perhaps packet filters are blocking the communication. Versions scan (-sV) can be used to help differentiate the truly open ports from the filtered ones.
A big challenge with UDP scanning is doing it quickly. Open and filtered ports rarely send any response, leaving Nmap to time out and then conduct retransmissions just in case the probe or response were lost. Closed ports are often an even bigger problem. They usually send back an ICMP port unreachable error. But unlike the RST packets sent by closed TCP ports in response to a SYN or connect scan, many hosts rate limit ICMP port unreachable messages by default. Linux and Solaris are particularly strict about this. For example, the Linux 2.4.20 kernel limits destination unreachable messages to one per second (in net/ipv4/icmp.c).
Nmap detects rate limiting and slows down accordingly to avoid flooding the network with useless packets that the target machine will drop. Unfortunately, a Linux-style limit of one packet per second makes a 65,536-port scan take more than 18 hours. Ideas for speeding your UDP scans up include scanning more hosts in parallel, doing a quick scan of just the popular ports first, scanning from behind the firewall, and using --host-timeout to skip slow hosts.
-sN; -sF; -sX (TCP Null, FIN, and Xmas scans)
When scanning systems compliant with this RFC text, any packet not containing SYN, RST, or ACK bits will result in a returned RST if the port is closed and no response at all if the port is open. As long as none of those three bits are included, any combination of the other three (FIN, PSH, and URG) are OK. Nmap exploits this with three scan types:
Null scan (-sN)
FIN scan (-sF)
Xmas scan (-sX)
These three scan types are exactly the same in behavior except for the TCP flags set in probe packets. If a RST packet is received, the port is considered closed, while no response means it is open|filtered. The port is marked filtered if an ICMP unreachable error (type 3, code 1, 2, 3, 9, 10, or 13) is received.
The key advantage to these scan types is that they can sneak through certain non-stateful firewalls and packet filtering routers. Another advantage is that these scan types are a little more stealthy than even a SYN scan. Don't count on this though---most modern IDS products can be configured to detect them. The big downside is that not all systems follow RFC 793 to the letter. A number of systems send RST responses to the probes regardless of whether the port is open or not. This causes all of the ports to be labeled closed. Major operating systems that do this are Microsoft Windows, many Cisco devices, BSDI, and IBM OS/400. This scan does work against most Unix-based systems though. Another downside of these scans is that they can't distinguish open ports from certain filtered ones, leaving you with the response open|filtered.
-sA (TCP ACK scan)
The ACK scan probe packet has only the ACK flag set (unless you use --scanflags). When scanning unfiltered systems, open and closed ports will both return a RST packet. Nmap then labels them as unfiltered, meaning that they are reachable by the ACK packet, but whether they are open or closed is undetermined. Ports that don't respond, or send certain ICMP error messages back (type 3, code 1, 2, 3, 9, 10, or 13), are labeled filtered.
-sW (TCP Window scan)
This scan relies on an implementation detail of a minority of systems out on the Internet, so you can't always trust it. Systems that don't support it will usually return all ports closed. Of course, it is possible that the machine really has no open ports. If most scanned ports are closed but a few common port numbers (such as 22, 25, 53) are filtered, the system is most likely susceptible. Occasionally, systems will even show the exact opposite behavior. If your scan shows 1000 open ports and 3 closed or filtered ports, then those three may very well be the truly open ones.
-sM (TCP Maimon scan)
--scanflags (Custom TCP scan)
The --scanflags argument can be a numerical flag value such as 9 (PSH and FIN), but using symbolic names is easier. Just mash together any combination of URG, ACK, PSH, RST, SYN, and FIN. For example, --scanflags URGACKPSHRSTSYNFIN sets everything, though it's not very useful for scanning. The order these are specified in is irrelevant.
In addition to specifying the desired flags, you can specify a TCP scan type (such as -sA or -sF). That base type tells Nmap how to interpret responses. For example, a SYN scan considers no-response to indicate a filtered port, while a FIN scan treats the same as open|filtered. Nmap will behave the same way it does for the base scan type, except that it will use the TCP flags you specify instead. If you don't specify a base type, SYN scan is used.
-sI <zombie host[:probeport]> (idle scan)
Besides being extraordinarily stealthy (due to its blind nature), this scan type permits mapping out IP-based trust relationships between machines. The port listing shows open ports from the perspective of the zombie host. So you can try scanning a target using various zombies that you think might be trusted (via router/packet filter rules).
You can add a colon followed by a port number to the zombie host if you wish to probe a particular port on the zombie for IP ID changes. Otherwise Nmap will use the port it uses by default for TCP pings (80).
-sO (IP protocol scan)
Besides being useful in its own right, protocol scan demonstrates the power of open source software. While the fundamental idea is pretty simple, I had not thought to add it nor received any requests for such functionality. Then in the summer of 2000, Gerhard Rieger conceived the idea, wrote an excellent patch implementing it, and sent it to the nmap-hackers mailing list. I incorporated that patch into the Nmap tree and released a new version the next day. Few pieces of commercial software have users enthusiastic enough to design and contribute their own improvements!
Protocol scan works in a similar fashion to UDP scan. Instead of iterating through the port number field of a UDP packet, it sends IP packet headers and iterates through the 8-bit IP protocol field. The headers are usually empty, containing no data and not even the proper header for the claimed protocol. The three exceptions are TCP, UDP, and ICMP. A proper protocol header for those is included since some systems won't send them otherwise and because Nmap already has functions to create them. Instead of watching for ICMP port unreachable messages, protocol scan is on the lookout for ICMP protocol unreachable messages. If Nmap receives any response in any protocol from the target host, Nmap marks that protocol as open. An ICMP protocol unreachable error (type 3, code 2) causes the protocol to be marked as closed Other ICMP unreachable errors (type 3, code 1, 3, 9, 10, or 13) cause the protocol to be marked filtered (though they prove that ICMP is open at the same time). If no response is received after retransmissions, the protocol is marked open|filtered
-b <FTP relay host> (FTP bounce scan)
This vulnerability was widespread in 1997 when Nmap was released, but has largely been fixed. Vulnerable servers are still around, so it is worth trying when all else fails. If bypassing a firewall is your goal, scan the target network for open port 21 (or even for any FTP services if you scan all ports with version detection), then try a bounce scan using each. Nmap will tell you whether the host is vulnerable or not. If you are just trying to cover your tracks, you don't need to (and, in fact, shouldn't) limit yourself to hosts on the target network. Before you go scanning random Internet addresses for vulnerable FTP servers, consider that sysadmins may not appreciate you abusing their servers in this way.
In addition to all of the scan methods discussed previously, Nmap offers options for specifying which ports are scanned and whether the scan order is randomized or sequential. By default, Nmap scans all ports up to and including 1024 as well as higher numbered ports listed in the nmap-services file for the protocol(s) being scanned.
-p <port ranges> (Only scan specified ports)
When scanning both TCP and UDP ports, you can specify a particular protocol by preceding the port numbers by T: or U:. The qualifier lasts until you specify another qualifier. For example, the argument -p U:53,111,137,T:21-25,80,139,8080 would scan UDP ports 53,111,and 137, as well as the listed TCP ports. Note that to scan both UDP & TCP, you have to specify -sU and at least one TCP scan type (such as -sS, -sF, or -sT). If no protocol qualifier is given, the port numbers are added to all protocol lists.
Ports can also be specified by name according to what the port is referred to in the nmap-services. You can even use the wildcards * and ? with the names. For example, to scan FTP and all ports whose names begin with http, use -p ftp,http*. Be careful about shell expansions and quote the argument to -p if unsure.
Ranges of ports can be surrounded by square brackets to indicate ports inside that range that appear in nmap-services. For example, the following will scan all ports in nmap-services equal to or below 1024: -p [-1024]. Be careful with shell expansions and quote the argument to -p if unsure.
-F (Fast (limited port) scan)
-r (Don't randomize ports)
Point Nmap at a remote machine and it might tell you that ports 25/tcp, 80/tcp, and 53/udp are open. Using its nmap-services database of about 2,200 well-known services, Nmap would report that those ports probably correspond to a mail server (SMTP), web server (HTTP), and name server (DNS) respectively. This lookup is usually accurate---the vast majority of daemons listening on TCP port 25 are, in fact, mail servers. However, you should not bet your security on this! People can and do run services on strange ports.
Even if Nmap is right, and the hypothetical server above is running SMTP, HTTP, and DNS servers, that is not a lot of information. When doing vulnerability assessments (or even simple network inventories) of your companies or clients, you really want to know which mail and DNS servers and versions are running. Having an accurate version number helps dramatically in determining which exploits a server is vulnerable to. Version detection helps you obtain this information.
After TCP and/or UDP ports are discovered using one of the other scan methods, version detection interrogates those ports to determine more about what is actually running. The nmap-service-probes database contains probes for querying various services and match expressions to recognize and parse responses. Nmap tries to determine the service protocol (e.g. FTP, SSH, telnet, http), the application name (e.g. ISC BIND, Apache httpd, Solaris telnetd), the version number, hostname, device type (e.g. printer, router), the OS family (e.g. Windows, Linux) and sometimes miscellaneous details like whether an X server is open to connections, the SSH protocol version, or the KaZaA user name). Of course, most services don't provide all of this information. If Nmap was compiled with OpenSSL support, it will connect to SSL servers to deduce the service listening behind that encryption layer. When RPC services are discovered, the Nmap RPC grinder (-sR) is automatically used to determine the RPC program and version numbers. Some UDP ports are left in the open|filtered state after a UDP port scan is unable to determine whether the port is open or filtered. Version detection will try to elicit a response from these ports (just as it does with open ports), and change the state to open if it succeeds. open|filtered TCP ports are treated the same way. Note that the Nmap -A option enables version detection among other things. A paper documenting the workings, usage, and customization of version detection is available at http://insecure.org/nmap/vscan/.
When Nmap receives responses from a service but cannot match them to its database, it prints out a special fingerprint and a URL for you to submit if to if you know for sure what is running on the port. Please take a couple minutes to make the submission so that your find can benefit everyone. Thanks to these submissions, Nmap has about 3,000 pattern matches for more than 350 protocols such as SMTP, FTP, HTTP, etc.
Version detection is enabled and controlled with the following options:
-sV (Version detection)
--allports (Don't exclude any ports from version detection)
--version-intensity <intensity> (Set version scan intensity)
--version-light (Enable light mode)
--version-all (Try every single probe)
--version-trace (Trace version scan activity)
-sR (RPC scan)
One of Nmap's best-known features is remote OS detection using TCP/IP stack fingerprinting. Nmap sends a series of TCP and UDP packets to the remote host and examines practically every bit in the responses. After performing dozens of tests such as TCP ISN sampling, TCP options support and ordering, IP ID sampling, and the initial window size check, Nmap compares the results to its nmap-os-db database of more than 800 known OS fingerprints and prints out the OS details if there is a match. Each fingerprint includes a freeform textual description of the OS, and a classification which provides the vendor name (e.g. Sun), underlying OS (e.g. Solaris), OS generation (e.g. 10), and device type (general purpose, router, switch, game console, etc).
If Nmap is unable to guess the OS of a machine, and conditions are good (e.g. at least one open port and one closed port were found), Nmap will provide a URL you can use to submit the fingerprint if you know (for sure) the OS running on the machine. By doing this you contribute to the pool of operating systems known to Nmap and thus it will be more accurate for everyone.
OS detection enables several other tests which make use of information that is gathered during the process anyway. One of these is uptime measurement, which uses the TCP timestamp option (RFC 1323) to guess when a machine was last rebooted. This is only reported for machines which provide this information. Another is TCP Sequence Predictability Classification. This measures approximately how hard it is to establish a forged TCP connection against the remote host. It is useful for exploiting source-IP based trust relationships (rlogin, firewall filters, etc) or for hiding the source of an attack. This sort of spoofing is rarely performed any more, but many machines are still vulnerable to it. The actual difficulty number is based on statistical sampling and may fluctuate. It is generally better to use the English classification such as "worthy challenge" or "trivial joke". This is only reported in normal output in verbose (-v) mode. When verbose mode is enabled along with -O, IP ID sequence generation is also reported. Most machines are in the "incremental" class, which means that they increment the ID field in the IP header for each packet they send. This makes them vulnerable to several advanced information gathering and spoofing attacks.
A paper documenting the workings, usage, and customization of OS detection is available at http://insecure.org/nmap/osdetect/.
OS detection is enabled and controlled with the following options:
-O (Enable OS detection)
--osscan-limit (Limit OS detection to promising targets)
--osscan-guess; --fuzzy (Guess OS detection results)
--max-os-tries (Set the maximum number of OS detection tries against a target)
The Nmap Scripting Engine (NSE) combines the efficiency of Nmap's network handling with the versatility of the lightweight scripting language Lua[6], thus providing innumerable opportunities. A more extensive documentation of the NSE (including its API) can be found at: http://www.insecure.org/nmap/nse. The target of the NSE is to provide Nmap with a flexible infrastructure for extending its capabilities and offering its users a simple way of creating customized tests. Uses for the NSE include (but definitely are not limited to):
Enhanced version detection (category version)---While Nmap already offers its Service and Version detection system, which is unmatched in terms of efficiency and scope, this power has its downside when it comes to services requiring more complex probes. The Skype-Protocol version 2 for instance can be identified by sending 2 independent probes to it, which the builtin system is not laid out for: a simple NSE-script can do the job and update the port's service information.
Malware-detection (categories malware and backdoor)- Both attackers and worms often leave backdoors---be it in form of SMTP-servers listening on uncommon ports mostly used by spammers for mail relay, or in form of an FTP-server giving crackers access to critical data. A few lines of Lua code can help to identify those loopholes easily.
Vulnerability Detection (category vulnerability)- NSE's capacity in detecting risks ranges from checking for default passwords on Apache distributions to testing whether a SMTP-server supports relaying mail from arbitrary domains.
Network Discovery and Information Gathering (categories safe, intrusive and discovery)---By providing you with a scripting language and a really efficient asynchronous network API on the one hand and the information gathered during earlier stages of a scan on the other hand the NSE is suited to write client programs for the services listening on a target machine. These clients may collect information like: listings of available NFS/SMB/RPC shares, the number of channels of an irc-network or currently logged on users.
To reflect those different uses and to simplify the choice of which scripts to run, each script contains a field associating it with one or more of the above mentioned categories. To maintain the matching from scripts to categories a file called script.db is installed along with the distributed scripts. Therefore, if you, for example, want to see if a machine is infected by any worm Nmap provides a script for you can simply run nmap --script=malware target-ip and check the output afterwards. The version scripts are always run implicitly when a script-scan is requested. The script.db is a Lua-script itself and can be updated through the --script-updatedb option.
A NSE-script basically is a chunk of Lua-code which has (among some informational fields, like name, id and categories) 2 functions: a test whether the particular script should be run against a certain host or port (called a hostrule or portrule respectively) and an action to be carried out if the test returns true. Scripts have access to most information gathered by Nmap during earlier stages. For each host this includes the IP address, hostname and (if available) operating system. If a script is targeted at a port it has access to the portnumber, the protocol (tcp, udp or ssl), the service running behind that port, and optionally information from a version-scan. NSE-scripts have by convention a .nse-extension. Although you are not required to follow this for the moment, this may change in the future. Nmap will issue a warning if a file has any other extension. More extensive documentation on the NSE, including a description of its API can be found at http://insecure.org/nmap/nse/.
-sC
--script <script-categories|directory|filename|all>
If a directory is specified and found, Nmap loads all NSE scripts (any filenames ending with .nse) from that directory. They must have the filename extension nse. Nmap does not recurse into subdirectories to find scripts. When individual file names are specified, the file extension does not have to be nse.
Nmap scripts are stored in a scripts subdirectory of the Nmap data directory by default. Scripts are indexed in a database stored in scripts/script.db. The database lists all of the scripts in each category. A single script may be in several categories.
--script-args=<name1=value1,name2={name3=value3},name4=value4>
--script-trace
--script-updatedb
One of my highest Nmap development priorities has always been performance. A default scan (nmap hostname) of a host on my local network takes a fifth of a second. That is barely enough time to blink, but adds up when you are scanning tens or hundreds of thousands of hosts. Moreover, certain scan options such as UDP scanning and version detection can increase scan times substantially. So can certain firewall configurations, particularly response rate limiting. While Nmap utilizes parallelism and many advanced algorithms to accelerate these scans, the user has ultimate control over how Nmap runs. Expert users carefully craft Nmap commands to obtain only the information they care about while meeting their time constraints.
Techniques for improving scan times include omitting non-critical tests, and upgrading to the latest version of Nmap (performance enhancements are made frequently). Optimizing timing parameters can also make a substantial difference. Those options are listed below.
Some options accept a time parameter. This is specified in milliseconds by default, though you can append 's', 'm', or 'h' to the value to specify seconds, minutes, or hours. So the --host-timeout arguments 900000, 900s, and 15m all do the same thing.
--min-hostgroup <numhosts>; --max-hostgroup <numhosts> (Adjust parallel scan group sizes)
By default, Nmap takes a compromise approach to this conflict. It starts out with a group size as low as five so the first results come quickly and then increases the groupsize to as high as 1024. The exact default numbers depend on the options given. For efficiency reasons, Nmap uses larger group sizes for UDP or few-port TCP scans.
When a maximum group size is specified with --max-hostgroup, Nmap will never exceed that size. Specify a minimum size with --min-hostgroup and Nmap will try to keep group sizes above that level. Nmap may have to use smaller groups than you specify if there are not enough target hosts left on a given interface to fulfill the specified minimum. Both may be set to keep the group size within a specific range, though this is rarely desired.
These options do not have an effect during the host discovery phase of a scan. This includes plain ping scans (-sP). Host discovery always works in large groups of hosts to improve speed and accuracy.
The primary use of these options is to specify a large minimum group size so that the full scan runs more quickly. A common choice is 256 to scan a network in Class C sized chunks. For a scan with many ports, exceeding that number is unlikely to help much. For scans of just a few port numbers, host group sizes of 2048 or more may be helpful.
--min-parallelism <numprobes>; --max-parallelism <numprobes> (Adjust probe parallelization)
The most common usage is to set --min-parallelism to a number higher than one to speed up scans of poorly performing hosts or networks. This is a risky option to play with, as setting it too high may affect accuracy. Setting this also reduces Nmap's ability to control parallelism dynamically based on network conditions. A value of ten might be reasonable, though I only adjust this value as a last resort.
The --max-parallelism option is sometimes set to one to prevent Nmap from sending more than one probe at a time to hosts. This can be useful in combination with --scan-delay (discussed later), although the latter usually serves the purpose well enough by itself.
--min-rtt-timeout <time>, --max-rtt-timeout <time>, --initial-rtt-timeout <time> (Adjust probe timeouts)
Specifying a lower --max-rtt-timeout and --initial-rtt-timeout than the defaults can cut scan times significantly. This is particularly true for pingless (-PN) scans, and those against heavily filtered networks. Don't get too aggressive though. The scan can end up taking longer if you specify such a low value that many probes are timing out and retransmitting while the response is in transit.
If all the hosts are on a local network, 100 milliseconds is a reasonable aggressive --max-rtt-timeout value. If routing is involved, ping a host on the network first with the ICMP ping utility, or with a custom packet crafter such as hping2 that is more likely to get through a firewall. Look at the maximum round trip time out of ten packets or so. You might want to double that for the --initial-rtt-timeout and triple or quadruple it for the --max-rtt-timeout. I generally do not set the maximum RTT below 100ms, no matter what the ping times are. Nor do I exceed 1000ms.
--min-rtt-timeout is a rarely used option that could be useful when a network is so unreliable that even Nmap's default is too aggressive. Since Nmap only reduces the timeout down to the minimum when the network seems to be reliable, this need is unusual and should be reported as a bug to the nmap-dev mailing list.
--max-retries <numtries> (Specify the maximum number of port scan probe retransmissions)
The default (with no -T template) is to allow ten retransmissions. If a network seems reliable and the target hosts aren't rate limiting, Nmap usually only does one retransmission. So most target scans aren't even affected by dropping --max-retries to a low value such as three. Such values can substantially speed scans of slow (rate limited) hosts. You usually lose some information when Nmap gives up on ports early, though that may be preferable to letting the --host-timeout expire and losing all information about the target.
--host-timeout <time> (Give up on slow target hosts)
--scan-delay <time>; --max-scan-delay <time> (Adjust delay between probes)
When Nmap adjusts the scan delay upward to cope with rate limiting, the scan slows down dramatically. The --max-scan-delay option specifies the largest delay that Nmap will allow. Setting this value too low can lead to wasteful packet retransmissions and possible missed ports when the target implements strict rate limiting.
Another use of --scan-delay is to evade threshold based intrusion detection and prevention systems (IDS/IPS).
--defeat-rst-ratelimit
Using this option can reduce accuracy, as some ports will appear non-responsive because Nmap didn't wait long enough for a rate-limited RST response. With a SYN scan, the non-response results in the port being labeled filtered rather than the closed state we see when RST packets are received. This optional is useful when you only care about open ports, and distinguishing between closed and filtered ports isn't worth the extra time.
-T <paranoid|sneaky|polite|normal|aggressive|insane> (Set a timing template)
These templates allow the user to specify how aggressive they wish to be, while leaving Nmap to pick the exact timing values. The templates also make some minor speed adjustments for which fine-grained control options do not currently exist. For example, -T4 prohibits the dynamic scan delay from exceeding 10ms for TCP ports and -T5 caps that value at 5 milliseconds. Templates can be used in combination with fine-grained controls, and the fine-grained controls will you specify will take precedence over the timing template default for that parameter. I recommend using -T4 when scanning reasonably modern and reliable networks. Keep that option even when you add fine-grained controls so that you benefit from those extra minor optimizations that it enables.
If you are on a decent broadband or ethernet connection, I would recommend always using -T4. Some people love -T5 though it is too aggressive for my taste. People sometimes specify -T2 because they think it is less likely to crash hosts or because they consider themselves to be polite in general. They often don't realize just how slow -T polite really is. Their scan may take ten times longer than a default scan. Machine crashes and bandwidth problems are rare with the default timing options (-T3) and so I normally recommend that for cautious scanners. Omitting version detection is far more effective than playing with timing values at reducing these problems.
While -T0 and -T1 may be useful for avoiding IDS alerts, they will take an extraordinarily long time to scan thousands of machines or ports. For such a long scan, you may prefer to set the exact timing values you need rather than rely on the canned -T0 and -T1 values.
The main effects of T0 are serializing the scan so only one port is scanned at a time, and waiting five minutes between sending each probe. T1 and T2 are similar but they only wait 15 seconds and 0.4 seconds, respectively, between probes. T3 is Nmap's default behavior, which includes parallelization. T4 does the equivalent of --max-rtt-timeout 1250 --initial-rtt-timeout 500 --max-retries 6 and sets the maximum TCP scan delay to 10 milliseconds. T5 does the equivalent of --max-rtt-timeout 300 --min-rtt-timeout 50 --initial-rtt-timeout 250 --max-retries 2 --host-timeout 15m as well as setting the maximum TCP scan delay to 5ms.
Many Internet pioneers envisioned a global open network with a universal IP address space allowing virtual connections between any two nodes. This allows hosts to act as true peers, serving and retrieving information from each other. People could access all of their home systems from work, changing the climate control settings or unlocking the doors for early guests. This vision of universal connectivity has been stifled by address space shortages and security concerns. In the early 1990s, organizations began deploying firewalls for the express purpose of reducing connectivity. Huge networks were cordoned off from the unfiltered Internet by application proxies, network address translation, and packet filters. The unrestricted flow of information gave way to tight regulation of approved communication channels and the content that passes over them.
Network obstructions such as firewalls can make mapping a network exceedingly difficult. It will not get any easier, as stifling casual reconnaissance is often a key goal of implementing the devices. Nevertheless, Nmap offers many features to help understand these complex networks, and to verify that filters are working as intended. It even supports mechanisms for bypassing poorly implemented defenses. One of the best methods of understanding your network security posture is to try to defeat it. Place yourself in the mind-set of an attacker, and deploy techniques from this section against your networks. Launch an FTP bounce scan, idle scan, fragmentation attack, or try to tunnel through one of your own proxies.
In addition to restricting network activity, companies are increasingly monitoring traffic with intrusion detection systems (IDS). All of the major IDSs ship with rules designed to detect Nmap scans because scans are sometimes a precursor to attacks. Many of these products have recently morphed into intrusion prevention systems (IPS) that actively block traffic deemed malicious. Unfortunately for network administrators and IDS vendors, reliably detecting bad intentions by analyzing packet data is a tough problem. Attackers with patience, skill, and the help of certain Nmap options can usually pass by IDSs undetected. Meanwhile, administrators must cope with large numbers of false positive results where innocent activity is misdiagnosed and alerted on or blocked.
Occasionally people suggest that Nmap should not offer features for evading firewall rules or sneaking past IDSs. They argue that these features are just as likely to be misused by attackers as used by administrators to enhance security. The problem with this logic is that these methods would still be used by attackers, who would just find other tools or patch the functionality into Nmap. Meanwhile, administrators would find it that much harder to do their jobs. Deploying only modern, patched FTP servers is a far more powerful defense than trying to prevent the distribution of tools implementing the FTP bounce attack.
There is no magic bullet (or Nmap option) for detecting and subverting firewalls and IDS systems. It takes skill and experience. A tutorial is beyond the scope of this reference guide, which only lists the relevant options and describes what they do.
-f (fragment packets); --mtu (using the specified MTU)
-D <decoy1 [,decoy2][,ME],...> (Cloak a scan with decoys)
Separate each decoy host with commas, and you can optionally use ME as one of the decoys to represent the position for your real IP address. If you put ME in the 6th position or later, some common port scan detectors (such as Solar Designer's excellent Scanlogd) are unlikely to show your IP address at all. If you don't use ME, nmap will put you in a random position. You can also use RND to generate a random, non-reserved IP address, or RND:<number> to generate <number> addresses.
Note that the hosts you use as decoys should be up or you might accidentally SYN flood your targets. Also it will be pretty easy to determine which host is scanning if only one is actually up on the network. You might want to use IP addresses instead of names (so the decoy networks don't see you in their nameserver logs).
Decoys are used both in the initial ping scan (using ICMP, SYN, ACK, or whatever) and during the actual port scanning phase. Decoys are also used during remote OS detection (-O). Decoys do not work with version detection or TCP connect scan.
It is worth noting that using too many decoys may slow your scan and potentially even make it less accurate. Also, some ISPs will filter out your spoofed packets, but many do not restrict spoofed IP packets at all.
-S <IP_Address> (Spoof source address)
Another possible use of this flag is to spoof the scan to make the targets think that someone else is scanning them. Imagine a company being repeatedly port scanned by a competitor! The -e option and -PN are generally required for this sort of usage. Note that you usually won't receive reply packets back (they will be addressed to the IP you are spoofing), so Nmap won't produce useful reports.
-e <interface> (Use specified interface)
--source-port <portnumber>; -g <portnumber> (Spoof source port number)
Secure solutions to these problems exist, often in the form of application-level proxies or protocol-parsing firewall modules. Unfortunately there are also easier, insecure solutions. Noting that DNS replies come from port 53 and active FTP from port 20, many administrators have fallen into the trap of simply allowing incoming traffic from those ports. They often assume that no attacker would notice and exploit such firewall holes. In other cases, administrators consider this a short-term stop-gap measure until they can implement a more secure solution. Then they forget the security upgrade.
Overworked network administrators are not the only ones to fall into this trap. Numerous products have shipped with these insecure rules. Even Microsoft has been guilty. The IPsec filters that shipped with Windows 2000 and Windows XP contain an implicit rule that allows all TCP or UDP traffic from port 88 (Kerberos). In another well-known case, versions of the Zone Alarm personal firewall up to 2.1.25 allowed any incoming UDP packets with the source port 53 (DNS) or 67 (DHCP).
Nmap offers the -g and --source-port options (they are equivalent) to exploit these weaknesses. Simply provide a port number and Nmap will send packets from that port where possible. Nmap must use different port numbers for certain OS detection tests to work properly, and DNS requests ignore the --source-port flag because Nmap relies on system libraries to handle those. Most TCP scans, including SYN scan, support the option completely, as does UDP scan.
--data-length <number> (Append random data to sent packets)
--ip-options <S|R [route]|L [route]|T|U ... >; --ip-options <hex string> (Send packets with specified ip options)
The most powerful way to specify IP options is to simply pass in values as the argument to --ip-options. Precede each hex number with \x then the two digits. You may repeat certain characters by following them with an asterisk and then the number of times you wish them to repeat. For example, \x01\x07\x04\x00*36\x01 is a hex string containing 36 NUL bytes.
Nmap also offers a shortcut mechanism for specifying options. Simply pass the letter R, T, or U to request record-route, record-timestamp, or both options together, respectively. Loose or strict source routing may be specified with an L or S followed by a space and then a space-separated list of IP addresses.
If you wish to see the options in packets sent and received, specify --packet-trace. For more information and examples of using IP options with Nmap, see http://seclists.org/nmap-dev/2006/q3/0052.html.
--ttl <value> (Set IP time-to-live field)
--randomize-hosts (Randomize target host order)
--spoof-mac <MAC address, prefix, or vendor name> (Spoof MAC address)
--badsum (Send packets with bogus TCP/UDP checksums)
Any security tools is only as useful as the output it generates. Complex tests and algorithms are of little value if they aren't presented in an organized and comprehensible fashion. Given the number of ways Nmap is used by people and other software, no single format can please everyone. So Nmap offers several formats, including the interactive mode for humans to read directly and XML for easy parsing by software.
In addition to offering different output formats, Nmap provides options for controlling the verbosity of output as well as debugging messages. Output types may be sent to standard output or to named files, which Nmap can append to or clobber. Output files may also be used to resume aborted scans.
Nmap makes output available in five different formats. The default is called interactive output, and it is sent to standard output (stdout). There is also normal output, which is similar to interactive except that it displays less runtime information and warnings since it is expected to be analyzed after the scan completes rather than interactively.
XML output is one of the most important output types, as it can be converted to HTML, easily parsed by programs such as Nmap graphical user interfaces, or imported into databases.
The two remaining output types are the simple grepable output which includes most information for a target host on a single line, and sCRiPt KiDDi3 0utPUt for users who consider themselves |<-r4d.
While interactive output is the default and has no associated command-line options, the other four format options use the same syntax. They take one argument, which is the filename that results should be stored in. Multiple formats may be specified, but each format may only be specified once. For example, you may wish to save normal output for your own review while saving XML of the same scan for programmatic analysis. You might do this with the options -oX myscan.xml -oN myscan.nmap. While this chapter uses the simple names like myscan.xml for brevity, more descriptive names are generally recommended. The names chosen are a matter of personal preference, though I use long ones that incorporate the scan date and a word or two describing the scan, placed in a directory named after the company I'm scanning.
While these options save results to files, Nmap still prints interactive output to stdout as usual. For example, the command nmap -oX myscan.xml target prints XML to myscan.xml and fills standard output with the same interactive results it would have printed if -oX wasn't specified at all. You can change this by passing a hyphen character as the argument to one of the format types. This causes Nmap to deactivate interactive output, and instead print results in the format you specified to the standard output stream. So the command nmap -oX - target will send only XML output to stdout. Serious errors may still be printed to the normal error stream, stderr.
Unlike some Nmap arguments, the space between the logfile option flag (such as -oX) and the filename or hyphen is mandatory. If you omit the flags and give arguments such as -oG- or -oXscan.xml, a backwards compatibility feature of Nmap will cause the creation of normal format output files named G- and Xscan.xml respectively.
Nmap also offers options to control scan verbosity and to append to output files rather than clobbering them. All of these options are described below.
Nmap Output Formats
-oN <filespec> (normal output)
-oX <filespec> (XML output)
XML offers a stable format that is easily parsed by software. Free XML parsers are available for all major computer languages, including C/C++, Perl, Python, and Java. People have even written bindings for most of these languages to handle Nmap output and execution specifically. Examples are Nmap::Scanner[8] and Nmap::Parser[9] in Perl CPAN. In almost all cases that a non-trivial application interfaces with Nmap, XML is the preferred format.
The XML output references an XSL stylesheet which can be used to format the results as HTML. The easiest way to use this is simply to load the XML output in a web browser such as Firefox or IE. By default, this will only work on the machine you ran Nmap on (or a similarly configured one) due to the hard-coded nmap.xsl filesystem path. Use the --webxml or --stylesheet options to create portable XML files that render as HTML on any web-connected machine.
-oS <filespec> (ScRipT KIdd|3 oUTpuT)
-oG <filespec> (grepable output)
Nevertheless, grepable output is still quite popular. It is a simple format that lists each host on one line and can be trivially searched and parsed with standard Unix tools such as grep, awk, cut, sed, diff, and Perl. Even I usually use it for one-off tests done at the command line. Finding all the hosts with the SSH port open or that are running Solaris takes only a simple grep to identify the hosts, piped to an awk or cut command to print the desired fields.
Grepable output consists of comments (lines starting with a pound (#)) and target lines. A target line includes a combination of 6 labeled fields, separated by tabs and followed with a colon. The fields are Host, Ports, Protocols, Ignored State, OS, Seq Index, IP ID, and Status.
The most important of these fields is generally Ports, which gives details on each interesting port. It is a comma separated list of port entries. Each port entry represents one interesting port, and takes the form of seven slash (/) separated subfields. Those subfields are: Port number, State, Protocol, Owner, Service, SunRPC info, and Version info.
As with XML output, this man page does not allow for documenting the entire format. A more detailed look at the Nmap grepable output format is available from http://www.unspecific.com/nmap-oG-output.
-oA <basename> (Output to all formats)
Verbosity and debugging options
-v (Increase verbosity level)
Most changes only affect interactive output, and some also affect normal and script kiddie output. The other output types are meant to be processed by machines, so Nmap can give substantial detail by default in those formats without fatiguing a human user. However, there are a few changes in other modes where output size can be reduced substantially by omitting some detail. For example, a comment line in the grepable output that provides a list of all ports scanned is only printed in verbose mode because it can be quite long.
-d [level] (Increase or set debugging level)
Debugging output is useful when a bug is suspected in Nmap, or if you are simply confused as to what Nmap is doing and why. As this feature is mostly intended for developers, debug lines aren't always self-explanatory. You may get something like: Timeout vals: srtt: -1 rttvar: -1 to: 1000000 delta 14987 ==> srtt: 14987 rttvar: 14987 to: 100000. If you don't understand a line, your only recourses are to ignore it, look it up in the source code, or request help from the development list (nmap-dev). Some lines are self explanatory, but the messages become more obscure as the debug level is increased.
--packet-trace (Trace packets and data sent and received)
--open (Show only open (or possibly open) ports)
--iflist (List interfaces and routes)
--log-errors (Log errors/warnings to normal mode output file)
Miscellaneous output options
--append-output (Append to rather than clobber output files)
--resume <filename> (Resume aborted scan)
--stylesheet <path or URL> (Set XSL stylesheet to transform XML output)
--webxml (Load stylesheet from Insecure.Org)
--no_stylesheet (Omit XSL stylesheet declaration from XML)
This section describes some important (and not-so-important) options that don't really fit anywhere else.
-6 (Enable IPv6 scanning)
While IPv6 hasn't exactly taken the world by storm, it gets significant use in some (usually Asian) countries and most modern operating systems support it. To use Nmap with IPv6, both the source and target of your scan must be configured for IPv6. If your ISP (like most of them) does not allocate IPv6 addresses to you, free tunnel brokers are widely available and work fine with Nmap. One of the better ones is run by BT Exact at https://tb.ipv6.btexact.com/. I have also used one that Hurricane Electric provides at http://ipv6tb.he.net/. 6to4 tunnels are another popular, free approach.
-A (Aggressive scan options)
--datadir <directoryname> (Specify custom Nmap data file location)