A resurgence of Denial of Service Attacks [1] aimed at various
targets in the Internet have produced new challenges within the
Internet Service Provider (ISP) and network security communities to
find new and innovative methods to mitigate these types of attacks.
The difficulties in reaching this goal are numerous; some simple
tools already exist to limit the effectiveness and scope of these
attacks, but they have not been widely implemented.
This method of attack has been known for some time. Defending against
it, however, has been an ongoing concern. Bill Cheswick is quoted in
[2] as saying that he pulled a chapter from his book, "Firewalls and
Internet Security" [3], at the last minute because there was no way
for an administrator of the system under attack to effectively defend
the system. By mentioning the method, he was concerned about
encouraging it's use.
While the filtering method discussed in this document does
absolutely nothing to protect against flooding attacks which
originate from valid prefixes (IP addresses), it will prohibit an
attacker within the originating network from launching an attack of
this nature using forged source addresses that do not conform to
ingress filtering rules. All providers of Internet connectivity are
urged to implement filtering described in this document to prohibit
attackers from using forged source addresses which do not reside
within a range of legitimately advertised prefixes. In other words,
if an ISP is aggregating routing announcements for multiple
downstream networks, strict traffic filtering should be used to
prohibit traffic which claims to have originated from outside of
these aggregated announcements.
An additional benefit of implementing this type of filtering is that
it enables the originator to be easily traced to it's true source,
since the attacker would have to use a valid, and legitimately
reachable, source address.
A simplified diagram of the TCP SYN flooding problem is depicted
below:
9.0.0.0/8
host <----- router <--- Internet <----- router <-- attacker
TCP/SYN
<---------------------------------------------
Source: 192.168.0.4/32
Ferguson & Senie Informational [Page 2]
RFC 2267 Network Ingress Filtering January 1998
SYN/ACK
no route
TCP/SYN
<---------------------------------------------
Source: 10.0.0.13/32
SYN/ACK
no route
TCP/SYN
<---------------------------------------------
Source: 172.16.0.2/32
SYN/ACK
no route
[etc.]
Assume:
o The "host" is the targeted machine.
o The attacker resides within the "valid" prefix, 9.0.0.0/8.
o The attacker launches the attack using randomly changing source
addresses; in this example, the source addresses are depicted as
from within [4], which are not generally present in the global
Internet routing tables, and therefore, unreachable. However, any
unreachable prefix could be used to perpetrate this attack
method.
Also worthy of mention is a case wherein the source address is forged
to appear to have originated from within another legitimate network
which appears in the global routing table(s). For example, an
attacker using a valid network address could wreak havoc by making
the attack appear to come from an organization which did not, in
fact, originate the attack and was completely innocent. In such
cases, the administrator of a system under attack may be inclined to
filter all traffic coming from the apparent attack source. Adding
such a filter would then result in a denial of service to
legitimate, non-hostile end-systems. In this case, the administrator
of the system under attack unwittingly becomes an accomplice of the
attacker.
Further complicating matters, TCP SYN flood attacks will result in
SYN-ACK packets being sent to one or many hosts which have no
involvement in the attack, but which become secondary victims. This
allows the attacker to abuse two or more systems at once.
Ferguson & Senie Informational [Page 3]
RFC 2267 Network Ingress Filtering January 1998
Similar attacks have been attempted using UDP and ICMP flooding.
The former attack (UDP flooding) uses forged packets to try and
connect the chargen UDP service to the echo UDP service at another
site. Systems administrators should NEVER allow UDP packets destined
for system diagnostic ports from outside of their administrative
domain to reach their systems. The latter attack (ICMP flooding),
uses an insidious feature in IP subnet broadcast replication
mechanics. This attack relies on a router serving a large multi-
access broadcast network to frame an IP broadcast address (such as
one destined for 10.255.255.255) into a Layer 2 broadcast frame (for
ethernet, FF:FF:FF:FF:FF:FF). Ethernet NIC hardware (MAC-layer
hardware, specifically) will only listen to a select number of
addresses in normal operation. The one MAC address that all devices
share in common in normal operation is the media broadcast, or
FF:FF:FF:FF:FF:FF. In this case, a device will take the packet and
send an interrupt for processing. Thus, a flood of these broadcast
frames will consume all available resources on an end-system [9]. It
is perhaps prudent that system administrators should consider
ensuring that their border routers do not allow directed broadcast
packets to be forwarded through their routers as a default.
When an TCP SYN attack is launched using unreachable source address,
the target host attempts to reserve resources waiting for a
response. The attacker repeatedly changes the bogus source address
on each new packet sent, thus exhausting additional host resources.
Alternatively, if the attacker uses someone else's valid host
address as the source address, the system under attack will send a
large number of SYN/ACK packets to what it believes is the originator
of the connection establishment sequence. In this fashion, the
attacker does damage to two systems: the destination target system,
as well as the system which is actually using the spoofed address in
the global routing system.
The result of both attack methods is extremely degraded performance,
or worse, a system crash.
In response to this threat, most operating system vendors have
modified their software to allow the targeted servers to sustain
attacks with very high connection attempt rates. This is a welcome
and necessary part of the solution to the problem. Ingress filtering
will take time to be implemented pervasively and be fully effective,
but the extensions to the operating systems can be implemented
quickly. This combination should prove effective against source
address spoofing. See [1] for vendor and platform software upgrade
information.
Ferguson & Senie Informational [Page 4]
RFC 2267 Network Ingress Filtering January 1998
The problems encountered with this type of attack are numerous, and
involve shortcomings in host software implementations, routing
methodologies, and the TCP/IP protocols themselves. However, by
restricting transit traffic which originates from a downstream
network to known, and intentionally advertised, prefix(es), the
problem of source address spoofing can be virtually eliminated in
this attack scenario.
11.0.0.0/8
/
router 1
/
/
/ 9.0.0.0/8
ISP <----- ISP <---- ISP <--- ISP <-- router <-- attacker
A B C D 2
/
/
/
router 3
/
12.0.0.0/8
In the example above, the attacker resides within 9.0.0.0/8, which is
provided Internet connectivity by ISP D. An input traffic filter on
the ingress (input) link of "router 2", which provides connectivity
to the attacker's network, restricts traffic to allow only traffic
originating from source addresses within the 9.0.0.0/8 prefix, and
prohibits an attacker from using "invalid" source addresses which
reside outside of this prefix range.
In other words, the ingress filter on "router 2" above would check:
IF packet's source address from within 9.0.0.0/8
THEN forward as appropriate
IF packet's source address is anything else
THEN deny packet
Network administrators should log information on packets which are
dropped. This then provides a basis for monitoring any suspicious
activity.
Ferguson & Senie Informational [Page 5]
RFC 2267 Network Ingress Filtering January 1998
Additional functions should be considered for future platform
implementations. The following one is worth noting:
o Implementation of automatic filtering on remote access servers.
In most cases, a user dialing into an access server is an
individual user on a single PC. The ONLY valid source IP address
for packets originating from that PC is the one assigned by the
ISP (whether statically or dynamically assigned). The remote
access server could check every packet on ingress to ensure the
user is not spoofing the source address on the packets which he
is originating. Obviously, provisions also need to be made for
cases where the customer legitimately is attaching a net or
subnet via a remote router, but this could certainly be
implemented as an optional parameter. We have received reports
that some vendors and some ISPs are already starting to
implement this capability.
We considered suggesting routers also validate the source IP address
of the sender as suggested in [8], but that methodology will not
operate well in the real networks out there today. The method
suggested is to look up source addresses to see that the return path
to that address would flow out the same interface as the packet
arrived upon. With the number of asymmetric routes in the Internet,
this would clearly be problematic.
Filtering of this nature has the potential to break some types of
"special" services. It is in the best interest of the ISP offering
these types of special services, however, to consider alternate
methods of implementing these services to avoid being affected by
ingress traffic filtering.
Mobile IP, as defined in [6], is specifically affected by ingress
traffic filtering. As specified, traffic to the mobile node is
tunneled, but traffic from the mobile node is not tunneled. This
results in packets from the mobile node(s) which have source
addresses that do not match with the network where the station is
attached. The Mobile IP Working Group is addressing this problem by
specifying "reverse tunnels" in [7]. This work in progress provides
a method for the data transmitted from the mobile node to be tunneled
to the home agent before transmission to the Internet. There are
additional benefits to the reverse tunneling scheme, including better
handling of multicast traffic. Those implementing mobile IP systems
are encouraged to implement this method of reverse tunneling.
Ferguson & Senie Informational [Page 6]
RFC 2267 Network Ingress Filtering January 1998
As mentioned previously, while ingress traffic filtering drastically
reduces the success of source address spoofing, it does not preclude
an attacker using a forged source address of another host within the
permitted prefix filter range. It does, however, ensure that when an
attack of this nature does indeed occur, a network administrator can
be sure that the attack is actually originating from within the known
prefixes that are being advertised. This simplifies tracking down the
culprit, and at worst, the administrator can block a range of source
addresses until the problem is resolved.
If ingress filtering is used in an environment where DHCP or BOOTP is
used, the network administrator would be well advised to ensure that
packets with a source address of 0.0.0.0 and a destination of
255.255.255.255 are allowed to reach the relay agent in routers when
appropriate. The scope of directed broadcast replication should be
controlled, however, and not arbitrarily forwarded.
Ingress traffic filtering at the periphery of Internet connected
networks will reduce the effectiveness of source address spoofing
denial of service attacks. Network service providers and
administrators have already begun implementing this type of filtering
on periphery routers, and it is recommended that all service
providers do so as soon as possible. In addition to aiding the
Internet community as a whole to defeat this attack method, it can
also assist service providers in locating the source of the attack if
service providers can categorically demonstrate that their network
already has ingress filtering in place on customer links.
Corporate network administrators should implement filtering to ensure
their corporate networks are not the source of such problems. Indeed,
filtering could be used within an organization to ensure users do not
cause problems by improperly attaching systems to the wrong networks.
The filtering could also, in practice, block a disgruntled employee
from anonymous attacks.
It is the responsibility of all network administrators to ensure they
do not become the unwitting source of an attack of this nature.
The primary intent of this document is to inherently increase
security practices and awareness for the Internet community as a
whole; as more Internet Providers and corporate network
administrators implement ingress filtering, the opportunity for an
attacker to use forged source addresses as an attack methodology will
significantly lessen. Tracking the source of an attack is simplified
Ferguson & Senie Informational [Page 7]
RFC 2267 Network Ingress Filtering January 1998
when the source is more likely to be "valid." By reducing the number
and frequency of attacks in the Internet as a whole, there will be
more resources for tracking the attacks which ultimately do occur.
The North American Network Operators Group (NANOG) [5] group as a
whole deserves special credit for openly discussing these issues and
actively seeking possible solutions. Also, thanks to Justin Newton
[Priori Networks] and Steve Bielagus [OpenROUTE Networks, Inc.] for
their comments and contributions.
[1] CERT Advisory CA-96.21; TCP SYN Flooding and IP Spoofing
Attacks; September 24, 1996.
[2] B. Ziegler, "Hacker Tangles Panix Web Site", Wall Street
Journal, 12 September 1996.
[3] "Firewalls and Internet Security: Repelling the Wily Hacker";
William R. Cheswick and Steven M. Bellovin, Addison-Wesley
Publishing Company, 1994; ISBN 0-201-63357-4.
[4] Rekhter, Y., Moskowitz, R., Karrenberg, D., de Groot, G., and
E. Lear, "Address Allocation for Private Internets", RFC 1918,
February 1996.
[5] The North American Network Operators Group;
http://www.nanog.org.
[6] Perkins, C., "IP Mobility Support", RFC 2002, October 1996.
[7] Montenegro, G., "Reverse Tunneling Mobile IP",
Work in Progress.
[8] Baker, F., "Requirements for IP Version 4 Routers", RFC 1812,
June 1995.
[9] Thanks to: Craig Huegen;
See: http://www.quadrunner.com/~chuegen/smurf.txt.
Ferguson & Senie Informational [Page 8]
RFC 2267 Network Ingress Filtering January 1998
Paul Ferguson
cisco Systems, Inc.
400 Herndon Parkway
Herndon, VA USA 20170
EMail: ferguson@cisco.com
Daniel Senie
BlazeNet, Inc.
4 Mechanic Street
Natick, MA USA 01760
EMail: dts@senie.com
Ferguson & Senie Informational [Page 9]
RFC 2267 Network Ingress Filtering January 1998
Copyright (C) The Internet Society (1998). All Rights Reserved.
This document and translations of it may be copied and furnished to
others, and derivative works that comment on or otherwise explain it
or assist in its implementation may be prepared, copied, published
and distributed, in whole or in part, without restriction of any
kind, provided that the above copyright notice and this paragraph are
included on all such copies and derivative works. However, this
document itself may not be modified in any way, such as by removing
the copyright notice or references to the Internet Society or other
Internet organizations, except as needed for the purpose of
developing Internet standards in which case the procedures for
copyrights defined in the Internet Standards process must be
followed, or as required to translate it into languages other than
English.
The limited permissions granted above are perpetual and will not be
revoked by the Internet Society or its successors or assigns.
This document and the information contained herein is provided on an
"AS IS" basis and THE INTERNET SOCIETY AND THE INTERNET ENGINEERING
TASK FORCE DISCLAIMS ALL WARRANTIES, EXPRESS OR IMPLIED, INCLUDING
BUT NOT LIMITED TO ANY WARRANTY THAT THE USE OF THE INFORMATION
HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED WARRANTIES OF
MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE.
Ferguson & Senie Informational [Page 10]