RFC 9109




Internet Engineering Task Force (IETF)                           F. Gont
Request for Comments: 9109                                       G. Gont
Updates: 5905                                               SI6 Networks
Category: Standards Track                                     M. Lichvar
ISSN: 2070-1721                                                  Red Hat
                                                             August 2021


          Network Time Protocol Version 4: Port Randomization

Abstract



   The Network Time Protocol (NTP) can operate in several modes.  Some
   of these modes are based on the receipt of unsolicited packets and
   therefore require the use of a well-known port as the local port.
   However, in the case of NTP modes where the use of a well-known port
   is not required, employing such a well-known port unnecessarily
   facilitates the ability of attackers to perform blind/off-path
   attacks.  This document formally updates RFC 5905, recommending the
   use of transport-protocol ephemeral port randomization for those
   modes where use of the NTP well-known port is not required.

Status of This Memo



   This is an Internet Standards Track document.

   This document is a product of the Internet Engineering Task Force
   (IETF).  It represents the consensus of the IETF community.  It has
   received public review and has been approved for publication by the
   Internet Engineering Steering Group (IESG).  Further information on
   Internet Standards is available in Section 2 of RFC 7841.

   Information about the current status of this document, any errata,
   and how to provide feedback on it may be obtained at
   https://www.rfc-editor.org/info/rfc9109.

Copyright Notice



   Copyright (c) 2021 IETF Trust and the persons identified as the
   document authors.  All rights reserved.

   This document is subject to BCP 78 and the IETF Trust's Legal
   Provisions Relating to IETF Documents
   (https://trustee.ietf.org/license-info) in effect on the date of
   publication of this document.  Please review these documents
   carefully, as they describe your rights and restrictions with respect
   to this document.  Code Components extracted from this document must
   include Simplified BSD License text as described in Section 4.e of
   the Trust Legal Provisions and are provided without warranty as
   described in the Simplified BSD License.

Table of Contents



   1.  Introduction
   2.  Terminology
   3.  Considerations about Port Randomization in NTP
     3.1.  Mitigation against Off-Path Attacks
     3.2.  Effects on Path Selection
     3.3.  Filtering of NTP Traffic
     3.4.  Effect on NAPT Devices
   4.  Update to RFC 5905
   5.  IANA Considerations
   6.  Security Considerations
   7.  References
     7.1.  Normative References
     7.2.  Informative References
   Acknowledgments

   Authors' Addresses



1.  Introduction



   The Network Time Protocol (NTP) is one of the oldest Internet
   protocols and is currently specified in [RFC5905].  Since its
   original implementation, standardization, and deployment, a number of
   vulnerabilities have been found both in the NTP specification and in
   some of its implementations [NTP-VULN].  Some of these
   vulnerabilities allow for blind/off-path attacks, where an attacker
   can send forged packets to one or both NTP peers to achieve Denial of
   Service (DoS), time shifts, or other undesirable outcomes.  Many of
   these attacks require the attacker to guess or know at least a target
   NTP association, typically identified by the tuple {srcaddr, srcport,
   dstaddr, dstport, keyid} (see Section 9.1 of [RFC5905]).  Some of
   these parameters may be known or easily guessed.

   NTP can operate in several modes.  Some of these modes rely on the
   ability of nodes to receive unsolicited packets and therefore require
   the use of the NTP well-known port (123).  However, for modes where
   the use of a well-known port is not required, employing the NTP well-
   known port unnecessarily facilitates the ability of attackers to
   perform blind/off-path attacks (since knowledge of the port numbers
   is typically required for such attacks).  A recent study [NIST-NTP]
   that analyzes the port numbers employed by NTP clients suggests that
   numerous NTP clients employ the NTP well-known port as their local
   port, or select predictable ephemeral port numbers, thus
   unnecessarily facilitating the ability of attackers to perform blind/
   off-path attacks against NTP.

   BCP 156 [RFC6056] already recommends the randomization of transport-
   protocol ephemeral ports.  This document aligns NTP with the
   recommendation in BCP 156 [RFC6056] by formally updating [RFC5905]
   such that port randomization is employed for those NTP modes for
   which the use of the NTP well-known port is not needed.

2.  Terminology



   The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
   "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and
   "OPTIONAL" in this document are to be interpreted as described in BCP
   14 [RFC2119] [RFC8174] when, and only when, they appear in all
   capitals, as shown here.

3.  Considerations about Port Randomization in NTP



   The following subsections analyze a number of considerations about
   transport-protocol ephemeral port randomization when applied to NTP.

3.1.  Mitigation against Off-Path Attacks



   There has been a fair share of work in the area of blind/off-path
   attacks against transport protocols and upper-layer protocols, such
   as [RFC4953] and [RFC5927].  Whether the target of the attack is a
   transport-protocol instance (e.g., TCP connection) or an upper-layer
   protocol instance (e.g., an application-protocol instance), the
   attacker is required to know or guess the five-tuple {Protocol, IP
   Source Address, IP Destination Address, Source Port, Destination
   Port} that identifies the target transport-protocol instance or the
   transport-protocol instance employed by the target upper-layer
   protocol instance.  Therefore, increasing the difficulty of guessing
   this five-tuple helps mitigate blind/off-path attacks.

   As a result of these considerations, transport-protocol ephemeral
   port randomization is a best current practice (BCP 156) that helps
   mitigate off-path attacks at the transport layer.  This document
   aligns the NTP specification [RFC5905] with the existing best current
   practice on transport-protocol ephemeral port selection, irrespective
   of other techniques that may (and should) be implemented for
   mitigating off-path attacks.

   We note that transport-protocol ephemeral port randomization is a
   transport-layer mitigation against blind/off-path attacks and does
   not preclude (nor is it precluded by) other possible mitigations for
   off-path attacks that might be implemented at other layers (e.g.,
   [NTP-DATA-MINIMIZATION]).  For instance, some of the aforementioned
   mitigations may be ineffective against some off-path attacks
   [NTP-FRAG] or may benefit from the additional entropy provided by
   port randomization [NTP-security].

3.2.  Effects on Path Selection



   Intermediate systems implementing the Equal-Cost Multipath (ECMP)
   algorithm may select the outgoing link by computing a hash over a
   number of values, including the transport-protocol source port.
   Thus, as discussed in [NTP-CHLNG], the selected client port may have
   an influence on the measured offset and delay.

   If the source port is changed with each request, packets in different
   exchanges will be more likely to take different paths, which could
   cause the measurements to be less stable and have a negative impact
   on the stability of the clock.

   Network paths to/from a given server are less likely to change
   between requests if port randomization is applied on a per-
   association basis.  This approach minimizes the impact on the
   stability of NTP measurements, but it may cause different clients in
   the same network synchronized to the same NTP server to have a
   significant stable offset between their clocks.  This is due to their
   NTP exchanges consistently taking different paths with different
   asymmetry in the network delay.

   Section 4 recommends that NTP implementations randomize the ephemeral
   port number of client/server associations.  The choice of whether to
   randomize the port number on a per-association or a per-request basis
   is left to the implementation.

3.3.  Filtering of NTP Traffic



   In a number of scenarios (such as when mitigating DDoS attacks), a
   network operator may want to differentiate between NTP requests sent
   by clients and NTP responses sent by NTP servers.  If an
   implementation employs the NTP well-known port for the client port,
   requests/responses cannot be readily differentiated by inspecting the
   source and destination port numbers.  Implementation of port
   randomization for nonsymmetrical modes allows for simple
   differentiation of NTP requests and responses and for the enforcement
   of security policies that may be valuable for the mitigation of DDoS
   attacks, when all NTP clients in a given network employ port
   randomization.

3.4.  Effect on NAPT Devices



   Some NAPT devices will reportedly not translate the source port of a
   packet when a system port number (i.e., a port number in the range
   0-1023) [RFC6335] is employed.  In networks where such NAPT devices
   are employed, use of the NTP well-known port for the client port may
   limit the number of hosts that may successfully employ NTP client
   implementations at any given time.

      |  NOTES:
      |
      |     NAPT devices are defined in Section 4.1.2 of [RFC2663].
      |  
      |     The reported behavior is similar to the special treatment of
      |     UDP port 500, which has been documented in Section 2.3 of
      |     [RFC3715].

   In the case of NAPT devices that will translate the source port even
   when a system port is employed, packets reaching the external realm
   of the NAPT will not employ the NTP well-known port as the source
   port, as a result of the port translation function being performed by
   the NAPT device.

4.  Update to RFC 5905



   The following text from Section 9.1 (Peer Process Variables) of
   [RFC5905]:

   |  dstport:  UDP port number of the client, ordinarily the NTP port
   |     number PORT (123) assigned by the IANA.  This becomes the
   |     source port number in packets sent from this association.

   is replaced with:

   |  dstport:  UDP port number of the client.  In the case of broadcast
   |     server mode (5) and symmetric modes (1 and 2), it SHOULD
   |     contain the NTP port number PORT (123) assigned by IANA.  In
   |     the client mode (3), it SHOULD contain a randomized port
   |     number, as specified in [RFC6056].  The value in this variable
   |     becomes the source port number of packets sent from this
   |     association.  The randomized port number SHOULD NOT be shared
   |     with other associations, to avoid revealing the randomized port
   |     to other associations.
   |  
   |     If a client implementation performs transport-protocol
   |     ephemeral port randomization on a per-request basis, it SHOULD
   |     close the corresponding socket/port after each request/response
   |     exchange.  In order to prevent duplicate or delayed server
   |     packets from eliciting ICMP port unreachable error messages
   |     [RFC0792] [RFC4443] at the client, the client MAY wait for more
   |     responses from the server for a specific period of time (e.g.,
   |     3 seconds) before closing the UDP socket/port.
   |  
   |  
   |        NOTES:
   |  
   |        Randomizing the ephemeral port number on a per-request basis
   |        will better mitigate blind/off-path attacks, particularly if
   |        the socket/port is closed after each request/response
   |        exchange, as recommended above.  The choice of whether to
   |        randomize the ephemeral port number on a per-request or a
   |        per-association basis is left to the implementation, and it
   |        should consider the possible effects on path selection along
   |        with its possible impact on time measurement.
   |  
   |        On most current operating systems, which implement ephemeral
   |        port randomization [RFC6056], an NTP client may normally
   |        rely on the operating system to perform ephemeral port
   |        randomization.  For example, NTP implementations using POSIX
   |        sockets may achieve ephemeral port randomization by _not_
   |        binding the socket with the bind() function or binding it to
   |        port 0, which has a special meaning of "any port".  Using
   |        the connect() function for the socket will make the port
   |        inaccessible by other systems (that is, only packets from
   |        the specified remote socket will be received by the
   |        application).

5.  IANA Considerations



   This document has no IANA actions.

6.  Security Considerations



   The security implications of predictable numeric identifiers
   [PEARG-NUMERIC-IDS] (and of predictable transport-protocol port
   numbers [RFC6056] in particular) have been known for a long time now.
   However, the NTP specification has traditionally followed a pattern
   of employing common settings even when not strictly necessary, which
   at times has resulted in negative security and privacy implications
   (see, e.g., [NTP-DATA-MINIMIZATION]).  The use of the NTP well-known
   port (123) for the srcport and dstport variables is not required for
   all operating modes.  Such unnecessary usage comes at the expense of
   reducing the amount of work required for an attacker to successfully
   perform blind/off-path attacks against NTP.  Therefore, this document
   formally updates [RFC5905], recommending the use of transport-
   protocol port randomization when use of the NTP well-known port is
   not required.

   This issue has been assigned CVE-2019-11331 [VULN-REPORT] in the U.S.
   National Vulnerability Database (NVD).

7.  References



7.1.  Normative References



   [RFC2119]  Bradner, S., "Key words for use in RFCs to Indicate
              Requirement Levels", BCP 14, RFC 2119,
              DOI 10.17487/RFC2119, March 1997,
              <https://www.rfc-editor.org/info/rfc2119>.

   [RFC5905]  Mills, D., Martin, J., Ed., Burbank, J., and W. Kasch,
              "Network Time Protocol Version 4: Protocol and Algorithms
              Specification", RFC 5905, DOI 10.17487/RFC5905, June 2010,
              <https://www.rfc-editor.org/info/rfc5905>.

   [RFC6056]  Larsen, M. and F. Gont, "Recommendations for Transport-
              Protocol Port Randomization", BCP 156, RFC 6056,
              DOI 10.17487/RFC6056, January 2011,
              <https://www.rfc-editor.org/info/rfc6056>.

   [RFC8174]  Leiba, B., "Ambiguity of Uppercase vs Lowercase in RFC
              2119 Key Words", BCP 14, RFC 8174, DOI 10.17487/RFC8174,
              May 2017, <https://www.rfc-editor.org/info/rfc8174>.

7.2.  Informative References



   [NIST-NTP] Sherman, J. and J. Levine, "Usage Analysis of the NIST
              Internet Time Service", Journal of Research of the
              National Institute of Standards and Technology, Volume
              121, DOI 10.6028/jres.121.003, March 2016,
              <https://tf.nist.gov/general/pdf/2818.pdf>.

   [NTP-CHLNG]
              Sommars, S., "Challenges in Time Transfer using the
              Network Time Protocol (NTP)", Proceedings of the 48th
              Annual Precise Time and Time Interval Systems and
              Applications Meeting, pp. 271-290,
              DOI 10.33012/2017.14978, January 2017,
              <http://leapsecond.com/ntp/
              NTP_Paper_Sommars_PTTI2017.pdf>.

   [NTP-DATA-MINIMIZATION]
              Franke, D. and A. Malhotra, "NTP Client Data
              Minimization", Work in Progress, Internet-Draft, draft-
              ietf-ntp-data-minimization-04, 25 March 2019,
              <https://datatracker.ietf.org/doc/html/draft-ietf-ntp-
              data-minimization-04>.

   [NTP-FRAG] Malhotra, A., Cohen, I., Brakke, E., and S. Goldberg,
              "Attacking the Network Time Protocol", NDSS '16,
              DOI 10.14722/ndss.2016.23090, February 2016,
              <https://www.cs.bu.edu/~goldbe/papers/NTPattack.pdf>.

   [NTP-security]
              Malhotra, A., Van Gundy, M., Varia, M., Kennedy, H.,
              Gardner, J., and S. Goldberg, "The Security of NTP's
              Datagram Protocol", Cryptology ePrint Archive Report
              2016/1006, DOI 10.1007/978-3-319-70972-7_23, February
              2017, <https://eprint.iacr.org/2016/1006.pdf>.

   [NTP-VULN] "Network Time Foundation",
              <http://support.ntp.org/bin/view/Main/SecurityNotice>.

   [PEARG-NUMERIC-IDS]
              Gont, F. and I. Arce, "On the Generation of Transient
              Numeric Identifiers", Work in Progress, Internet-Draft,
              draft-irtf-pearg-numeric-ids-generation-07, 2 February
              2021, <https://datatracker.ietf.org/doc/html/draft-irtf-
              pearg-numeric-ids-generation-07>.

   [RFC0792]  Postel, J., "Internet Control Message Protocol", STD 5,
              RFC 792, DOI 10.17487/RFC0792, September 1981,
              <https://www.rfc-editor.org/info/rfc792>.

   [RFC2663]  Srisuresh, P. and M. Holdrege, "IP Network Address
              Translator (NAT) Terminology and Considerations",
              RFC 2663, DOI 10.17487/RFC2663, August 1999,
              <https://www.rfc-editor.org/info/rfc2663>.

   [RFC3715]  Aboba, B. and W. Dixon, "IPsec-Network Address Translation
              (NAT) Compatibility Requirements", RFC 3715,
              DOI 10.17487/RFC3715, March 2004,
              <https://www.rfc-editor.org/info/rfc3715>.

   [RFC4443]  Conta, A., Deering, S., and M. Gupta, Ed., "Internet
              Control Message Protocol (ICMPv6) for the Internet
              Protocol Version 6 (IPv6) Specification", STD 89,
              RFC 4443, DOI 10.17487/RFC4443, March 2006,
              <https://www.rfc-editor.org/info/rfc4443>.

   [RFC4953]  Touch, J., "Defending TCP Against Spoofing Attacks",
              RFC 4953, DOI 10.17487/RFC4953, July 2007,
              <https://www.rfc-editor.org/info/rfc4953>.

   [RFC5927]  Gont, F., "ICMP Attacks against TCP", RFC 5927,
              DOI 10.17487/RFC5927, July 2010,
              <https://www.rfc-editor.org/info/rfc5927>.

   [RFC6335]  Cotton, M., Eggert, L., Touch, J., Westerlund, M., and S.
              Cheshire, "Internet Assigned Numbers Authority (IANA)
              Procedures for the Management of the Service Name and
              Transport Protocol Port Number Registry", BCP 165,
              RFC 6335, DOI 10.17487/RFC6335, August 2011,
              <https://www.rfc-editor.org/info/rfc6335>.

   [VULN-REPORT]
              The MITRE Corporation, "CVE-2019-1133", National
              Vulnerability Database, August 2020,
              <https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-
              2019-11331>.

Acknowledgments

   The authors would like to thank (in alphabetical order) Ivan Arce,
   Roman Danyliw, Dhruv Dhody, Lars Eggert, Todd Glassey, Blake Hudson,
   Benjamin Kaduk, Erik Kline, Watson Ladd, Aanchal Malhotra, Danny
   Mayer, Gary E. Miller, Bjorn Mork, Hal Murray, Francesca Palombini,
   Tomoyuki Sahara, Zaheduzzaman Sarker, Dieter Sibold, Steven Sommars,
   Jean St-Laurent, Kristof Teichel, Brian Trammell, Éric Vyncke, Ulrich
   Windl, and Dan Wing for providing valuable comments on earlier draft
   versions of this document.

   Watson Ladd raised the problem of DDoS mitigation when the NTP well-
   known port is employed as the client port (discussed in Section 3.3
   of this document).

   The authors would like to thank Harlan Stenn for answering questions
   about a popular NTP implementation (see <https://www.nwtime.org>).

   Fernando Gont would like to thank Nelida Garcia and Jorge Oscar Gont
   for their love and support.

Authors' Addresses



   Fernando Gont
   SI6 Networks
   Evaristo Carriego 2644
   1706 Haedo, Provincia de Buenos Aires
   Argentina

   Phone: +54 11 4650 8472
   Email: fgont@si6networks.com
   URI:   https://www.si6networks.com


   Guillermo Gont
   SI6 Networks
   Evaristo Carriego 2644
   1706 Haedo, Provincia de Buenos Aires
   Argentina

   Phone: +54 11 4650 8472
   Email: ggont@si6networks.com
   URI:   https://www.si6networks.com


   Miroslav Lichvar
   Red Hat
   Purkynova 115
   612 00 Brno
   Czech Republic

   Email: mlichvar@redhat.com