TCP Maintenance and Minor F. Gont Extensions (tcpm) UTN/FRH Internet-Draft September 9, 2004 Expires: March 10, 2005 ICMP attacks against TCP draft-gont-tcpm-icmp-attacks-01.txt Status of this Memo This document is an Internet-Draft and is subject to all provisions of section 3 of RFC 3667. By submitting this Internet-Draft, each author represents that any applicable patent or other IPR claims of which he or she is aware have been or will be disclosed, and any of which he or she become aware will be disclosed, in accordance with RFC 3668. This document may not be modified, and derivative works of it may not be created, except to publish it as an RFC and to translate it into languages other than English. Internet-Drafts are working documents of the Internet Engineering Task Force (IETF), its areas, and its working groups. Note that other groups may also distribute working documents as Internet-Drafts. Internet-Drafts are draft documents valid for a maximum of six months and may be updated, replaced, or obsoleted by other documents at any time. It is inappropriate to use Internet-Drafts as reference material or to cite them other than as "work in progress." The list of current Internet-Drafts can be accessed at http://www.ietf.org/ietf/1id-abstracts.txt. The list of Internet-Draft Shadow Directories can be accessed at http://www.ietf.org/shadow.html. This Internet-Draft will expire on March 10, 2005. Copyright Notice Copyright (C) The Internet Society (2004). Abstract This document discusses the use of the Internet Control Message Protocol (ICMP) to perform a variety of attacks against the Transmission Control Protocol (TCP) and other similar protocols. It proposes a work-around to eliminate or minimize the impact of this type of attack. Gont Expires March 10, 2005 [Page 1] Internet-Draft ICMP attacks against TCP September 2004 Table of Contents 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 3 2. Background . . . . . . . . . . . . . . . . . . . . . . . . . . 3 2.1 Internet Control Message Protocol (ICMP) . . . . . . . . . 3 2.2 Handling of ICMP errors . . . . . . . . . . . . . . . . . 4 3. ICMP attacks against TCP . . . . . . . . . . . . . . . . . . . 5 3.1 Blind connection-reset attacks . . . . . . . . . . . . . . 5 3.2 Degrading the performance of a connection . . . . . . . . 6 4. Constraints in the possible solutions . . . . . . . . . . . . 6 5. Solutions to the problem . . . . . . . . . . . . . . . . . . . 6 5.1 TCP sequence number checking . . . . . . . . . . . . . . . 6 5.2 Delaying the connection-reset . . . . . . . . . . . . . . 7 5.3 Changing the handling of ICMP error messages for connections in the synchronized states . . . . . . . . . . 7 5.3.1 ICMP type 2 (Destination Unreachable), code 2 (protocol unreachable) . . . . . . . . . . . . . . . . 7 5.3.2 ICMP type 2 (Destination Unreachable), code 3 (port unreachable) . . . . . . . . . . . . . . . . . . 8 5.3.3 ICMP type 2 (Destination Unreachable), code 4 (fragmentation needed and DF bit set) . . . . . . . . 8 5.4 Ignoring ICMP Source Quench messages . . . . . . . . . . . 8 5.5 Port randomization . . . . . . . . . . . . . . . . . . . . 8 5.6 Authentication . . . . . . . . . . . . . . . . . . . . . . 9 6. Future work . . . . . . . . . . . . . . . . . . . . . . . . . 9 7. Security Considerations . . . . . . . . . . . . . . . . . . . 9 8. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . 9 9. References . . . . . . . . . . . . . . . . . . . . . . . . . . 9 9.1 Normative References . . . . . . . . . . . . . . . . . . . . 9 9.2 Informative References . . . . . . . . . . . . . . . . . . . 10 Author's Address . . . . . . . . . . . . . . . . . . . . . . . 11 A. Changes from draft-gont-tcpm-icmp-attacks-00 . . . . . . . . . 11 Intellectual Property and Copyright Statements . . . . . . . . 12 Gont Expires March 10, 2005 [Page 2] Internet-Draft ICMP attacks against TCP September 2004 1. Introduction Recently, awareness has been raised about several threats against the TCP [1] protocol, which include blind connection-reset attacks [6]. These attacks are based on sending forged TCP segments to any of the TCP endpoints, requiring the attacker to be able to guess the four-tuple that identifies the connection to be attacked. While these attacks were known by the research community, they were considered to be unfeasible. However, increases in bandwidth availability, and the use of larger TCP windows [7] have made these attacks feasible. Several solutions have been proposed to either eliminate or minimize the impact of these attacks [8][9][10]. However, there is still a possibility for performing a number of attacks against the TCP protocol, by means of ICMP [2]. These attacks include, among others, blind connection-reset attacks. This document aims to raise awareness of the use of ICMP to perform a number of attacks against TCP, and proposes several counter-measures that can eliminate or minimize the impact of these attacks. The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in RFC 2119 [3]. 2. Background 2.1 Internet Control Message Protocol (ICMP) The Internet Control Message Protocol (ICMP) is used by the Internet Architecture to perform the fault-isolation function, that is, the group of actions that hosts and routers take to determine that there is some network failure [11]. When an intermediate router detects a network problem while trying to forward an IP packet, it will send an ICMP error message to the source host, to raise awareness of the network problem. In the same way, there are a number of cases in which an end-system may generate an ICMP error message when it finds a problem while processing a datagram. The internet header plus the first 64 bits of the packet that elicited the ICMP message are included in the payload of the ICMP error message, so that the receiving host can match the error to the instance of the transport protocol that elicited the error message. Thus, it is assumed that all data needed to identify a transport protocol instance is contained in the first 64 bits of the transport Gont Expires March 10, 2005 [Page 3] Internet-Draft ICMP attacks against TCP September 2004 protocol header. When the transport protocol is notified of the error condition, it will perform a fault recovery function. That is, it will try to survive the network failure. In the case of TCP, the fault recovery policy is as follows: o If the network problem being reported is a hard error, abort the corresponding connection. o If the network problem being reported is a soft error, just record this information, and repeatedly retransmit the segment until either it gets acknowledged, or the connection times out. The Host Requirements RFC [4] defines "hard errors" as ICMP error messages of type 3 (Destination Unreachable) codes 2 (protocol unreachable), 3 (port unreachable), and 4 (fragmentation needed and DF bit set). Thus, any of these ICMP messages could elicit a connection abort. [12] provides information about which ICMP error messages are produced by hosts, intermediate routers, or both. 2.2 Handling of ICMP errors The Host Requirements RFC [4] states that a TCP instance should be notified of ICMP error messages received for its corresponding connection. However, neither the Host Requirements RFC nor the original TCP specification [1] recommend any additional security checks on the received ICMP messages. Therefore, as long as the ICMP payload contains the correct four-tuple that identifies the communication instance, it will be processed by the corresponding transport-protocol instance, and the corresponding action will be performed. Thus, in order to perform any of the attacks discussed in this document, an attacker only needs to guess the four-tuple that identifies the communication instance to be attacked. As discussed in [6], there are a number of scenarios in which an attacker may be able to know or guess this four-tuple. Furthermore, it must be noted that most services use the so-called "well-known" ports, so that only the client port would need to be guessed. In the event that an attacker had no knowledge about the range of port numbers used by clients, this would mean that an attacker would need to send, at most, 65535 packets to perform any of Gont Expires March 10, 2005 [Page 4] Internet-Draft ICMP attacks against TCP September 2004 the attacks described in this document. It is clear that additional security checks should be performed on the received ICMP error messages. 3. ICMP attacks against TCP ICMP messages can be used to perform a variety of attacks. These attacks have been discussed by the research community to a large extent. Some TCP/IP implementations have added extra security checks on the received ICMP error messages to minimize the impact of these attacks. However, as there has not been any official proposal about what would be the best way to deal with these attacks, these additional security checks have not been widely implemented. The following subsections discuss some of the possible attacks, and propose work-arounds to eliminate or minimize the impact of these attacks. 3.1 Blind connection-reset attacks The Host Requirements RFC [4] states that a host SHOULD abort the corresponding connection when receiving an ICMP error message that indicates a hard error. For ICMP messages of type 2 (Destination Unreachable) code 2 (protocol unreachable), specifically, the Host Requirements RFC states that even those transport protocols that have their own mechanisms to indicate that a port is unreachable MUST accept these ICMP error messages for the same purpose. That is, they MUST abort the corresponding connection when an ICMP port unreachable message is received. Thus, an attacker could use ICMP to perform a blind connection-reset attack. That is, even being off-path, an attacker could reset any TCP connection taking place. In order to perform such an attack, an attacker would send any ICMP error message that indicates a "hard error", to either of the two TCP endpoints of the connection. Because of TCP's fault recovery policy, the connection would be immediately aborted. All an attacker needs to know to perform such an attack is the socket pair that identifies the TCP connection to be attacked. In some scenarios, the IP addresses and port numbers in use may be easily guessed or known to the attacker [6]. Gont Expires March 10, 2005 [Page 5] Internet-Draft ICMP attacks against TCP September 2004 There are some points to be considered about this type of attack: o The source address of the ICMP error message need not be forged. Thus, simple egress-filtering based on the source address of IP packets would not serve as a counter-measure against this type of attack. o Even if TCP itself were protected against the blind connection-reset attack described in [6] and [8], the type of attack described in this document could still succeed. 3.2 Degrading the performance of a connection The Host requirements RFC states hosts MUST react to ICMP Source Quench messages by slowing transmission on the connection. Thus, an attacker could send ICMP Source Quench [2] messages to a TCP endpoint to make it reduce the rate at which it sends data to the other party. While this would not reset the connection, it would certainly degrade the performance of the data transfer taking place over it. 4. Constraints in the possible solutions The original ICMP specification [2] requires nodes generating ICMP errors to include the IP header of the packet that elicited the ICMP error message, plus the first 64 bits of its payload, in the payload of the ICMP error message. For TCP, that means that the only fields that will be included are: the source port number, the destination port number, and the 32-bit sequence number. This imposes a constraint on the possible solutions, as there is not much information avalable on which to perform additional security checks. While there exists a proposal to recommend hosts to include more data from the original datagram in the payload of ICMP error messages [13], we cannot yet propose any work-around based on any data past the first 64 bits of the payload of the original IP datagram that elicited the ICMP error message. 5. Solutions to the problem There are a number of counter-measures against this type of attack. Rather than being alternative measures, they could be implemented together to increase the protection against this type of attack. 5.1 TCP sequence number checking TCP SHOULD check that the sequence number in the TCP header contained in the payload of the ICMP error message is within the range SND.UNA < SEG.SEQ < SND.NXT. This means that the sequence number should be Gont Expires March 10, 2005 [Page 6] Internet-Draft ICMP attacks against TCP September 2004 within the range of the data already sent but not yet acknowledged. If an ICMP error message doesn't pass this check, it SHOULD be discarded. Even if an attacker were able to guess the four-tuple that identifies the TCP connection, this additional check would reduce the possibility of success of the attacker to Flight_Size/2^^32 (where Flight_Size is the number of data bytes already sent to the remote peer, but not yet acknowledged [14]). For connections in the SYN-SENT or SYN-RECEIVED states, this would reduce the probability of success of a blind-connection reset attack during the connection-establishment phase to 1/2^^32. For a TCP endpoint with no data "in flight", this would completely eliminate the possibility of success of this attack. 5.2 Delaying the connection-reset For connections in any of the synchronized states, an additional counter-measure against the blind connection-reset attack could be taken. Rather than immediately aborting a connection, a TCP could abort a connection only after an ICMP error message indicating a hard error has been received a specified number of times, and the corresponding data have already been retransmitted more than some specified number of times. For example, hosts could abort connections only after a fourth ICMP error message indicating a hard error is received, and the corresponding data have already been retransmitted more than six times. The rationale behind this proposed fix is that if a host can make forward progress on a connection, it can completely disregard the "hard errors" being indicated by the received ICMP error messages. 5.3 Changing the handling of ICMP error messages for connections in the synchronized states An analysis of the circumstances in which ICMP messages that indicate hard errors may be received can shed some light to minimize (or even eliminate) the impact of blind connection-reset attacks. 5.3.1 ICMP type 2 (Destination Unreachable), code 2 (protocol unreachable) This ICMP error message indicates that the host sending the ICMP error message received a packet meant for a transport protocol it does not support. For connection-oriented protocols such as TCP, one could expect to receive such an error as the result of a connection Gont Expires March 10, 2005 [Page 7] Internet-Draft ICMP attacks against TCP September 2004 request. However, it would be strange to get such an error during the life of a connection, as this would indicate that support for that transport protocol has been removed from the host sending the error message during the life of the corresponding connection. Thus, it would be fair to ignore ICMP protocol unreachable error messages meant for connections that are in synchronized states. For TCP, this would mean one would ignore ICMP port unreachable error messages meant for connections that are in the ESTABLISHED, FIN-WAIT-1, FIN-WAIT-2, CLOSE-WAIT, CLOSING, LAST-ACK or TIME-WAIT states. 5.3.2 ICMP type 2 (Destination Unreachable), code 3 (port unreachable) This error message indicates that the host sending the ICMP error message received a packet meant for a socket (IP address, port number) on which there is no process listening. Those transport protocols which have their own mechanisms for notifying this condition should not be receiving these error messages. However, the Host Requirements RFC [4] states that even those transport protocols that have their own mechanism for notifying the sender that a port is unreachable MUST nevertheless accept an ICMP Port Unreachable for the same purpose. For security reasons, it would be fair to ignore ICMP port unreachable messages that are meant for protocols that have their own mechanisms for reporting this condition. 5.3.3 ICMP type 2 (Destination Unreachable), code 4 (fragmentation needed and DF bit set) This error message indicates that an intermediate node needed to fragment a datagram, but the DF (Don't Fragment) bit in the IP header was set. Those TCP/IP stacks implementing the Path-MTU Discovery (PMTUD) mechanism [5] should not abort the corresponding connection when such an error message is received. 5.4 Ignoring ICMP Source Quench messages The Host Requirements RFC [4] states that hosts MUST react to ICMP Source Quench messages by slowing transmission on the connection. However, as discussed in the Requirements for IP Version 4 Routers RFC [15], research seems to suggest ICMP Source Quench is an ineffective (and unfair) antidote for congestion. Thus, we recommend hosts to completely ignore ICMP Source Quench messages. 5.5 Port randomization As discussed in the previous sections, in order to perform any of the attacks described in this document, an attacker needs to guess (or know) the four-tuple that identifies the connection to be attacked. Randomizing the ephemeral ports used by the clients would reduce the Gont Expires March 10, 2005 [Page 8] Internet-Draft ICMP attacks against TCP September 2004 chances of success by an attacker. A proposal exists to enable TCP to reassign a well-known port number to a random value [16]. 5.6 Authentication Hosts could require ICMP error messages to be authenticated [12], in order to act upon them. However, while this requirement could make sense for those ICMP error messages sent by hosts, it would not be feasible for those ICMP error messages generated by intermediate routers. [12] contains a discussion on the authentication of ICMP messages. 6. Future work The same considerations discussed in this document should be applied to other similar protocols. 7. Security Considerations This document describes the use of ICMP error messages to perform a number of attacks against the TCP protocol, and proposes a number of counter-measures that either eliminate or reduce the impact of these attacks. 8. Acknowledgements This document was inspired by Mikka Liljeberg, while discussing some issues related to [17] by private e-mail. The author would like to thank James Carlson, Juan Fraschini, Markus Friedl, Guillermo Gont, Michael Kerrisk, Kacheong Poon, and Andrew Powell, for contributing many valuable comments. 9. References 9.1 Normative References [1] Postel, J., "Transmission Control Protocol", STD 7, RFC 793, September 1981. [2] Postel, J., "Internet Control Message Protocol", STD 5, RFC 792, September 1981. [3] Bradner, S., "Key words for use in RFCs to Indicate Requirement Levels", BCP 14, RFC 2119, March 1997. Gont Expires March 10, 2005 [Page 9] Internet-Draft ICMP attacks against TCP September 2004 [4] Braden, R., "Requirements for Internet Hosts - Communication Layers", STD 3, RFC 1122, October 1989. [5] Mogul, J. and S. Deering, "Path MTU discovery", RFC 1191, November 1990. 9.2 Informative References [6] Watson, P., "Slipping in the Window: TCP Reset Attacks", 2004 CanSecWest Conference , 2004. [7] Jacobson, V., Braden, B. and D. Borman, "TCP Extensions for High Performance", RFC 1323, May 1992. [8] Stewart, R., "Transmission Control Protocol security considerations", draft-ietf-tcpm-tcpsecure-01 (work in progress), June 2004. [9] Touch, J., "ANONsec: Anonymous IPsec to Defend Against Spoofing Attacks", draft-touch-anonsec-00 (work in progress), May 2004. [10] Poon, K., "Use of TCP timestamp option to defend against blind spoofing attack", draft-poon-tcp-tstamp-mod-00 (work in progress), June 2004. [11] Clark, D., "Fault isolation and recovery", RFC 816, July 1982. [12] Kent, S. and R. Atkinson, "Security Architecture for the Internet Protocol", RFC 2401, November 1998. [13] Gont, F., "Increasing the payload of ICMP error messages", (work in progress) draft-gont-icmp-payload-00.txt, 2004. [14] Allman, M., Paxson, V. and W. Stevens, "TCP Congestion Control", RFC 2581, April 1999. [15] Baker, F., "Requirements for IP Version 4 Routers", RFC 1812, June 1995. [16] Shepard, T., "Reassign Port Number option for TCP", draft-shepard-tcp-reassign-port-number-00 (work in progress), July 2004. [17] Gont, F., "TCP's Reaction to Soft Errors", draft-gont-tcpm-tcp-soft-errors-00 (work in progress), June 2004. Gont Expires March 10, 2005 [Page 10] Internet-Draft ICMP attacks against TCP September 2004 Author's Address Fernando Gont Universidad Tecnologica Nacional Evaristo Carriego 2644 Haedo, Provincia de Buenos Aires 1706 Argentina Phone: +54 11 4650 8472 EMail: fernando@gont.com.ar Appendix A. Changes from draft-gont-tcpm-icmp-attacks-00 o Added Section 5.3 o Added a summary of the relevant RFCs in several sections o Miscellaneous editorial changes Gont Expires March 10, 2005 [Page 11] Internet-Draft ICMP attacks against TCP September 2004 Intellectual Property Statement The IETF takes no position regarding the validity or scope of any Intellectual Property Rights or other rights that might be claimed to pertain to the implementation or use of the technology described in this document or the extent to which any license under such rights might or might not be available; nor does it represent that it has made any independent effort to identify any such rights. Information on the procedures with respect to rights in RFC documents can be found in BCP 78 and BCP 79. Copies of IPR disclosures made to the IETF Secretariat and any assurances of licenses to be made available, or the result of an attempt made to obtain a general license or permission for the use of such proprietary rights by implementers or users of this specification can be obtained from the IETF on-line IPR repository at http://www.ietf.org/ipr. The IETF invites any interested party to bring to its attention any copyrights, patents or patent applications, or other proprietary rights that may cover technology that may be required to implement this standard. Please address the information to the IETF at ietf-ipr@ietf.org. Disclaimer of Validity This document and the information contained herein are provided on an "AS IS" basis and THE CONTRIBUTOR, THE ORGANIZATION HE/SHE REPRESENTS OR IS SPONSORED BY (IF ANY), THE INTERNET SOCIETY AND THE INTERNET ENGINEERING TASK FORCE DISCLAIM 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. Copyright Statement Copyright (C) The Internet Society (2004). This document is subject to the rights, licenses and restrictions contained in BCP 78, and except as set forth therein, the authors retain all their rights. Acknowledgment Funding for the RFC Editor function is currently provided by the Internet Society. Gont Expires March 10, 2005 [Page 12]