Why the first 1056 Ports?
 Internet
ports are numbered from 1 through 65535, but the first 1023 ports are
special. By tradition, and some enforcement, ports 1 through 1023 are
generally reserved for the acceptance of incoming connections by
services running on the receiving system. Internet services "listen" on
various standard low-numbered ports so that clients wishing to have
access to those services know where they may be found. Web servers
traditionally listen on port 80, eMail servers listen on ports 25 and
110, FTP servers listen on port 21 and Telnet servers listen on port
23. And the list goes on. Here's the official Internet Assigned Numbers Authority (IANA) port assignment list.
Although
it is possible to have higher-numbered ports listening for incoming
connections, our scan of the entire "service port range" will detect
all standard services running and listening on the standard service
ports.
Due
to the insecure behavior of Microsoft's Windows operating systems, we
have added an additional 33 ports to these first 1023 ports, bringing
the total to 1056. Windows has a tendency to establish globally
available listening services on the first few ports in the "client
port" range which begins just past 1023. If you are not running a
personal firewall, or you are allowing ShieldsUP! probes into your
network, you may discover one or more additional open ports at, or just
above, 1024.
Strange Results? Personal
firewalls are beginning to exhibit "adaptive behavior". The grid shown
to the left starts off showing ports mostly closed with a few open
(mostly blue with a few red cells). Then at some point it suddenly
switches into "stealth mode". This can occur when a firewall "adapts"
to the scanning IP and raises its defenses against just the attacker.
This complicates the job of accurately checking a system's security.
 Two things you can do:
If you are not certain whether your firewall is adaptive, you can
re-run any test here to compare the results. Differing behavior often
indicates that your firewall has "learned" that it is being probed from
our IP and is treating it differently. For the most accurate scan
results, disable any adaptive behavior during the testing.
 You can use these tests to learn exactly how your firewall deals with probes to specific ports, and to port ranges.
Beyond providing a comprehensive test of your system's first 1056 ports, this service ports scan can be used for additional research:
Service Ports Scan Application Guide ( Cool things you can do with our Service Ports Probe )
 |
Detecting Ports Blocked by Your ISP
 Internet
service providers often block specific traffic entering their network
before it reaches their customers, or after leaving their customers
before it exits their network. This is sometimes done to block the
exploitation of common security vulnerabilities, and sometimes to
prevent their customers from offering proscribed Internet services.
As
a customer, it can be useful and interesting to know which service
ports, if any, an ISP has chosen to preemptively block in order to
restrict their customers' global Internet traffic.
ISP
port blocking can be easily tested, often quite rapidly, by arranging
to allow the ShieldsUP! probe to have access to an unprotected
computer. Since all non-stealth machines will respond to every open
request — either affirmatively or negatively — ports appearing as
STEALTH will be those blocked by your ISP, corporate firewall, or other
external agency.

 | | If your system is unprotected, without any personal firewall or NAT router, any ports showing as stealth
are being blocked somewhere between your computer and the public
Internet. This is probably being done by your ISP. Internet traffic
directed to your computer at the stealth ports will be dropped before
reaching your machine. |

 | | If
your system has a personal firewall that can be instructed to "trust" a
specific remote IP, you can temporarily instruct it to trust the
ShieldsUP! probe IP of [4.79.142.206]. If, after doing so, most of the
service ports change to either open or closed , you have succeeded and any which remain stealth are being blocked by your ISP. |

 | | If
your system is operating behind a residential "NAT" router, the router
will be acting as a natural and excellent hardware firewall. But that's
not what you want for the moment. You can temporarily remove your NAT
router and connect an unprotected computer directly to your cable modem
or DSL line. Or, if you are comfortable reconfiguring your NAT router,
you may be able to point the router's "DMZ" at one of your computers
which has been instructed to "trust" our probe IP of [4.79.142.206]. If, after doing so, most of the service ports change to either open or closed , you have succeeded and any remaining stealth are being blocked by your ISP. |

 | | Finally,
if your Internet security system, NAT router, personal firewall, or
whatever, can produce detailed logs of incoming Internet packets, you
could leave your existing security in place, clear your log, run the
service ports scan, then carefully inspect your log for any
consistently missing port probes. We send out four sets of probing
packets because individual packets are sometimes dropped along the way.
Therefore, it won't be unusual to see occasional missing packets from
your logs. What you're looking for is a complete lack of packets bound
for a specific port. A careful and detailed examination of your log
will reveal any missing ports which are being blocked before they reach
your logging tool. (Note that this technique is not quite as foolproof
as the other approaches since ISPs could be blocking outbound packets
from their customers, which the other approaches would detect but
log-watching would not.) |
After
completing the experiments above, remember to return your system to its
previous tight security and verify that everything is safe again by
re-running any of our tests.

 Checking a NAT Router's WAN Security
 Residential
broadband "NAT" routers which allow many computers to share a single
Internet connection are becoming quite popular. We love them for the
security they provide to the machines placed behind them since any NAT
router functions as a natural and excellent hardware firewall.
However,
the Internet or "WAN" (Wide Area Network) side connection of many NAT
routers and DSL gateways is not as secure as it should be. Many routers
ship with web, ftp, or Telnet management ports wide open! And many are
still configured with their well-known default administrative
passwords. Although the router may be protecting the machines behind
it, it might not be protecting itself without your deliberate closing
of remote "WAN" administration ports.
ShieldsUP!
automatically tests your NAT router's WAN-side security because the
router's WAN IP is the single public IP that connects your internal
private network to the public Internet. When a test is initiated by any
system behind a NAT router, we are testing the public-side security of
the router itself and not the security of the individual machines which
are located behind and protected by the router.

 Adaptive IDENT Stealthing Experimentation
 The IDENT protocol's port 113
is quite problematical and tricky to stealth. If the user's port 113 is
completely stealthed, connections to some remote Internet servers such
as eMail, Internet Relay Chat (IRC), and others, may be delayed or
denied altogether. For this reason, many NAT routers and personal
firewalls do not attempt to stealth port 113, they settle for leaving
it closed. One of the first things that caught my eye about the
ZoneAlarm personal firewall was that it was clever about handling port
113: It "adaptively stealthed" the port.
To
understand the following discussion, you should familiarize yourself
with the details of the IDENT protocol and port 113. Please read port 113's Port Authority database page before proceeding.
Even
after many years, the (free) ZoneAlarm personal firewall from Zone Labs
is the only personal firewall to "adaptively" stealth port 113. Unlike
any other firewall or NAT router (any of which could also do the same)
this allows port 113 to be stealthed to any passing Internet scanners
or probes, but "unstealthed" for any valid IDENT connection attempts
originating from remote servers with which the user's computer is
attempting to connect. (Since this could easily be done by any personal
firewall or even NAT routers, I am hopeful that this feature might yet
appear in other products.)
"Adaptive
Stealthing" means that when a TCP SYN packet arrives to request a
connection to your machine's port 113, ZoneAlarm checks, on the fly, to
see whether your machine currently has any sort of "relationship" with
the remote machine (such as a pending outgoing connection attempt). If
so, the remote machine is considered to be "friendly" and its IDENT
request packet is allowed to pass through ZoneAlarm's firewall. But if
the IDENT originating machine is not known to ZoneAlarm as a "friendly"
machine, the connection requesting packet is dropped and discarded,
rendering port 113 stealth to all unknown port scanners. It's very
slick.
IDENT, ZoneAlarm, and ShieldsUP!
 Even
though your computer's web browser already has a relationship with the
web server at GRC, our tests originate from a different "foreign" IP
address. ZoneAlarm therefore drops incoming packets to port 113 from
this different probing IP address and ZoneAlarm users see that port 113
is stealthed to passing Internet scans.
To
demonstrate how ZoneAlarm (and perhaps someday other firewalls or NAT
routers) selectively "unstealth" port 113 — but only for known
"friendly" machines — we simply initiate a connection from your web
browser to the ShieldsUP! scanning IP. Even though the connection
attempt will ultimately fail (since there's no web server at the
probing address), ZoneAlarm will note the outgoing attempt and will
unstealth port 113 for subsequent probes.
Step One:
Verify that our scan currently show port 113 stealthed. (You may wish
to use one of the other remote port tests which will be faster than an
entire 1056-port grid scan.)
Step Two:
Open a secondary web browser window to initiate a connection to the
probing IP. (Users of Microsoft Internet Explorer can press Ctrl-N to
"clone" their current browser window.)
Step Three: In the secondary web browser window, click this URL or enter this address:

http://4.79.142.206
This second connection attempt will ultimately fail, but ZoneAlarm will notice the effort, which is all that's necessary.
Step Four:
Finally, refresh the port probe window or repeat the scan to check your
system's current port status. You should find that port 113 is no
longer "stealth" to the probing IP address because you are attempting
to connect to it and it has been determined to be "friendly".
Step Five:
If you're curious, stop and close the secondary web browser window and
periodically refresh your port probe window to see how long the
"friendly" status persists before Zone Alarm returns the probing IP to
unknown status and port 113 to full stealth. |