RFC 9319




Internet Engineering Task Force (IETF)                          Y. Gilad
Request for Comments: 9319                Hebrew University of Jerusalem
BCP: 185                                                     S. Goldberg
Category: Best Current Practice                        Boston University
ISSN: 2070-1721                                                K. Sriram
                                                                USA NIST
                                                             J. Snijders
                                                                  Fastly
                                                             B. Maddison
                                               Workonline Communications
                                                            October 2022


The Use of maxLength in the Resource Public Key Infrastructure (RPKI)

Abstract



   This document recommends ways to reduce the forged-origin hijack
   attack surface by prudently limiting the set of IP prefixes that are
   included in a Route Origin Authorization (ROA).  One recommendation
   is to avoid using the maxLength attribute in ROAs except in some
   specific cases.  The recommendations complement and extend those in
   RFC 7115.  This document also discusses the creation of ROAs for
   facilitating the use of Distributed Denial of Service (DDoS)
   mitigation services.  Considerations related to ROAs and RPKI-based
   Route Origin Validation (RPKI-ROV) in the context of destination-
   based Remotely Triggered Discard Route (RTDR) (elsewhere referred to
   as "Remotely Triggered Black Hole") filtering are also highlighted.

Status of This Memo



   This memo documents an Internet Best Current Practice.

   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
   BCPs 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/rfc9319.

Copyright Notice



   Copyright (c) 2022 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 Revised BSD License text as described in Section 4.e of the
   Trust Legal Provisions and are provided without warranty as described
   in the Revised BSD License.

Table of Contents



   1.  Introduction
     1.1.  Requirements
     1.2.  Documentation Prefixes
   2.  Suggested Reading
   3.  Forged-Origin Sub-Prefix Hijack
   4.  Measurements of the RPKI
   5.  Recommendations about Minimal ROAs and maxLength
     5.1.  Facilitating Ad Hoc Routing Changes and DDoS Mitigation
     5.2.  Defensive De-aggregation in Response to Prefix Hijacks
   6.  Considerations for RTDR Filtering Scenarios
   7.  User Interface Design Recommendations
   8.  Operational Considerations
   9.  Security Considerations
   10. IANA Considerations
   11. References
     11.1.  Normative References
     11.2.  Informative References
   Acknowledgments

   Authors' Addresses



1.  Introduction



   The Resource Public Key Infrastructure (RPKI) [RFC6480] uses Route
   Origin Authorizations (ROAs) to create a cryptographically verifiable
   mapping from an IP prefix to a set of Autonomous Systems (ASes) that
   are authorized to originate that prefix.  Each ROA contains a set of
   IP prefixes and the AS number of one of the ASes authorized to
   originate all the IP prefixes in the set [RFC6482].  The ROA is
   cryptographically signed by the party that holds a certificate for
   the set of IP prefixes.

   The ROA format also supports a maxLength attribute.  According to
   [RFC6482], "When present, the maxLength specifies the maximum length
   of the IP address prefix that the AS is authorized to advertise."
   Thus, rather than requiring the ROA to list each prefix that the AS
   is authorized to originate, the maxLength attribute provides a
   shorthand that authorizes an AS to originate a set of IP prefixes.

   However, measurements of RPKI deployments have found that the use of
   the maxLength attribute in ROAs tends to lead to security problems.
   In particular, measurements taken in June 2017 showed that of the
   prefixes specified in ROAs that use the maxLength attribute, 84% were
   vulnerable to a forged-origin sub-prefix hijack [GSG17].  The forged-
   origin prefix or sub-prefix hijack involves inserting the legitimate
   AS as specified in the ROA as the origin AS in the AS_PATH; the
   hijack can be launched against any IP prefix/sub-prefix that has a
   ROA.  Consider a prefix/sub-prefix that has a ROA that is unused
   (i.e., not announced in BGP by a legitimate AS).  A forged-origin
   hijack involving such a prefix/sub-prefix can propagate widely
   throughout the Internet.  On the other hand, if the prefix/sub-prefix
   were announced by the legitimate AS, then the propagation of the
   forged-origin hijack is somewhat limited because of its increased
   AS_PATH length relative to the legitimate announcement.  Of course,
   forged-origin hijacks are harmful in both cases, but the extent of
   harm is greater for unannounced prefixes.  See Section 3 for detailed
   discussion.

   For this reason, this document recommends that, whenever possible,
   operators SHOULD use "minimal ROAs" that authorize only those IP
   prefixes that are actually originated in BGP, and no other prefixes.
   Further, it recommends ways to reduce the forged-origin attack
   surface by prudently limiting the address space that is included in
   ROAs.  One recommendation is to avoid using the maxLength attribute
   in ROAs except in some specific cases.  The recommendations
   complement and extend those in [RFC7115].  The document also
   discusses the creation of ROAs for facilitating the use of DDoS
   mitigation services.  Considerations related to ROAs and RPKI-ROV in
   the context of destination-based Remotely Triggered Discard Route
   (RTDR) (elsewhere referred to as "Remotely Triggered Black Hole")
   filtering are also highlighted.

   Please note that the term "RPKI-based Route Origin Validation" and
   the corresponding acronym "RPKI-ROV" that are used in this document
   mean the same as the term "Prefix Origin Validation" used in
   [RFC6811].

   One ideal place to implement the ROA-related recommendations is in
   the user interfaces for configuring ROAs.  Recommendations for
   implementors of such user interfaces are provided in Section 7.

   The practices described in this document require no changes to the
   RPKI specifications and will not increase the number of signed ROAs
   in the RPKI because ROAs already support lists of IP prefixes
   [RFC6482].

1.1.  Requirements



   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.

1.2.  Documentation Prefixes



   The documentation prefixes recommended in [RFC5737] are insufficient
   for use as example prefixes in this document.  Therefore, this
   document uses the address space defined in [RFC1918] for constructing
   example prefixes.

   Note that although the examples in this document are presented using
   IPv4 prefixes, all the analysis thereof and the recommendations made
   are equally valid for the equivalent IPv6 cases.

2.  Suggested Reading



   It is assumed that the reader understands BGP [RFC4271], RPKI
   [RFC6480], ROAs [RFC6482], RPKI-ROV [RFC6811], and BGPsec [RFC8205].

3.  Forged-Origin Sub-Prefix Hijack



   A detailed description and discussion of forged-origin sub-prefix
   hijacks are presented here, especially considering the case when the
   sub-prefix is not announced in BGP.  The forged-origin sub-prefix
   hijack is relevant to a scenario in which:

   (1)  the RPKI [RFC6480] is deployed, and

   (2)  routers use RPKI-ROV to drop invalid routes [RFC6811], but

   (3)  BGPsec [RFC8205] (or any similar method to validate the
        truthfulness of the BGP AS_PATH attribute) is not deployed.

   Note that this set of assumptions accurately describes a substantial
   and growing number of large Internet networks at the time of writing.

   The forged-origin sub-prefix hijack [RFC7115] [GCHSS] is described
   here using a running example.

   Consider the IP prefix 192.168.0.0/16, which is allocated to an
   organization that also operates AS 64496.  In BGP, AS 64496
   originates the IP prefix 192.168.0.0/16 as well as its sub-prefix
   192.168.225.0/24.  Therefore, the RPKI should contain a ROA
   authorizing AS 64496 to originate these two IP prefixes.

   Suppose, however, the organization issues and publishes a ROA
   including a maxLength value of 24:

      ROA:(192.168.0.0/16-24, AS 64496)

   We refer to the above as a "loose ROA" since it authorizes AS 64496
   to originate any sub-prefix of 192.168.0.0/16 up to and including
   length /24, rather than only those prefixes that are intended to be
   announced in BGP.

   Because AS 64496 only originates two prefixes in BGP (192.168.0.0/16
   and 192.168.225.0/24), all other prefixes authorized by the loose ROA
   (for instance, 192.168.0.0/24) are vulnerable to the following
   forged-origin sub-prefix hijack [RFC7115] [GCHSS]:

      The hijacker AS 64511 sends a BGP announcement "192.168.0.0/24: AS
      64511, AS 64496", falsely claiming that AS 64511 is a neighbor of
      AS 64496 and that AS 64496 originates the IP prefix
      192.168.0.0/24.  In fact, the IP prefix 192.168.0.0/24 is not
      originated by AS 64496.

      The hijacker's BGP announcement is valid according to the RPKI
      since the ROA (192.168.0.0/16-24, AS 64496) authorizes AS 64496 to
      originate BGP routes for 192.168.0.0/24.

      Because AS 64496 does not actually originate a route for
      192.168.0.0/24, the hijacker's route is the only route for
      192.168.0.0/24.  Longest-prefix-match routing ensures that the
      hijacker's route to the sub-prefix 192.168.0.0/24 is always
      preferred over the legitimate route to 192.168.0.0/16 originated
      by AS 64496.

   Thus, the hijacker's route propagates through the Internet, and
   traffic destined for IP addresses in 192.168.0.0/24 will be delivered
   to the hijacker.

   The forged-origin sub-prefix hijack would have failed if a minimal
   ROA as described in Section 5 was used instead of the loose ROA.  In
   this example, a minimal ROA would be:

      ROA:(192.168.0.0/16, 192.168.225.0/24, AS 64496)

   This ROA is "minimal" because it includes only those IP prefixes that
   AS 64496 originates in BGP, but no other IP prefixes [RFC6907].

   The minimal ROA renders AS 64511's BGP announcement invalid because:

   (1)  this ROA "covers" the attacker's announcement (since
        192.168.0.0/24 is a sub-prefix of 192.168.0.0/16), and

   (2)  there is no ROA "matching" the attacker's announcement (there is
        no ROA for AS 64511 and IP prefix 192.168.0.0/24) [RFC6811].

   If routers ignore invalid BGP announcements, the minimal ROA above
   ensures that the sub-prefix hijack will fail.

   Thus, if a minimal ROA had been used, the attacker would be forced to
   launch a forged-origin prefix hijack in order to attract traffic as
   follows:

      The hijacker AS 64511 sends a BGP announcement "192.168.0.0/16: AS
      64511, AS 64496", falsely claiming that AS 64511 is a neighbor of
      AS 64496.

   This forged-origin prefix hijack is significantly less damaging than
   the forged-origin sub-prefix hijack:

      AS 64496 legitimately originates 192.168.0.0/16 in BGP, so the
      hijacker AS 64511 is not presenting the only route to
      192.168.0.0/16.

      Moreover, the path originated by AS 64511 is one hop longer than
      the path originated by the legitimate origin AS 64496.

   As discussed in [LSG16], this means that the hijacker will attract
   less traffic than it would have in the forged-origin sub-prefix
   hijack where the hijacker presents the only route to the hijacked
   sub-prefix.

   In summary, a forged-origin sub-prefix hijack has the same impact as
   a regular sub-prefix hijack, despite the increased AS_PATH length of
   the illegitimate route.  A forged-origin sub-prefix hijack is also
   more damaging than the forged-origin prefix hijack.

4.  Measurements of the RPKI



   Network measurements taken in June 2017 showed that 12% of the IP
   prefixes authorized in ROAs have a maxLength value longer than their
   prefix length.  Of these, the vast majority (84%) were non-minimal,
   as they included sub-prefixes that are not announced in BGP by the
   legitimate AS and were thus vulnerable to forged-origin sub-prefix
   hijacks.  See [GSG17] for details.

   These measurements suggest that operators commonly misconfigure the
   maxLength attribute and unwittingly open themselves up to forged-
   origin sub-prefix hijacks.  That is, they are exposing a much larger
   attack surface for forged-origin hijacks than necessary.

5.  Recommendations about Minimal ROAs and maxLength



   Operators SHOULD use minimal ROAs whenever possible.  A minimal ROA
   contains only those IP prefixes that are actually originated by an AS
   in BGP and no other IP prefixes.  See Section 3 for an example.

   In general, operators SHOULD avoid using the maxLength attribute in
   their ROAs, since its inclusion will usually make the ROA non-
   minimal.

   One such exception may be when all more specific prefixes permitted
   by the maxLength value are actually announced by the AS in the ROA.
   Another exception is where: (a) the maxLength value is substantially
   larger compared to the specified prefix length in the ROA, and (b) a
   large number of more specific prefixes in that range are announced by
   the AS in the ROA.  In practice, this case should occur rarely (if at
   all).  Operator discretion is necessary in this case.

   This practice requires no changes to the RPKI specifications and need
   not increase the number of signed ROAs in the RPKI because ROAs
   already support lists of IP prefixes [RFC6482].  See [GSG17] for
   further discussion of why this practice will have minimal impact on
   the performance of the RPKI ecosystem.

   Operators that implement these recommendations and have existing ROAs
   published in the RPKI system MUST perform a review of such objects,
   especially where they make use of the maxLength attribute, to ensure
   that the set of included prefixes is "minimal" with respect to the
   current BGP origination and routing policies.  Published ROAs MUST be
   replaced as necessary.  Such an exercise MUST be repeated whenever
   the operator makes changes to either policy.

5.1.  Facilitating Ad Hoc Routing Changes and DDoS Mitigation



   Operational requirements may require that a route for an IP prefix be
   originated on an ad hoc basis, with little or no prior warning.  An
   example of such a situation arises when an operator wishes to make
   use of DDoS mitigation services that use BGP to redirect traffic via
   a "scrubbing center".

   In order to ensure that such ad hoc routing changes are effective, a
   ROA validating the new route should exist.  However, a difficulty
   arises due to the fact that newly created objects in the RPKI are
   made visible to relying parties considerably more slowly than routing
   updates in BGP.

   Ideally, it would not be necessary to pre-create the ROA, which
   validates the ad hoc route, and instead create it "on the fly" as
   required.  However, this is practical only if the latency imposed by
   the propagation of RPKI data is guaranteed to be within acceptable
   limits in the circumstances.  For time-critical interventions such as
   responding to a DDoS attack, this is unlikely to be the case.

   Thus, the ROA in question will usually need to be created well in
   advance of the routing intervention, but such a ROA will be non-
   minimal, since it includes an IP prefix that is sometimes (but not
   always) originated in BGP.

   In this case, the ROA SHOULD only include:

   (1)  the set of IP prefixes that are always originated in BGP, and

   (2)  the set of IP prefixes that are sometimes, but not always,
        originated in BGP.

   The ROA SHOULD NOT include any IP prefixes that the operator knows
   will not be originated in BGP.  In general, the ROA SHOULD NOT make
   use of the maxLength attribute unless doing so has no impact on the
   set of included prefixes.

   The running example is now extended to illustrate one situation where
   it is not possible to issue a minimal ROA.

   Consider the following scenario prior to the deployment of RPKI.
   Suppose AS 64496 announced 192.168.0.0/16 and has a contract with a
   DDoS mitigation service provider that holds AS 64500.  Further,
   assume that the DDoS mitigation service contract applies to all IP
   addresses covered by 192.168.0.0/22.  When a DDoS attack is detected
   and reported by AS 64496, AS 64500 immediately originates
   192.168.0.0/22, thus attracting all the DDoS traffic to itself.  The
   traffic is scrubbed at AS 64500 and then sent back to AS 64496 over a
   backhaul link.  Notice that, during a DDoS attack, the DDoS
   mitigation service provider AS 64500 originates a /22 prefix that is
   longer than AS 64496's /16 prefix, so all the traffic (destined to
   addresses in 192.168.0.0/22) that normally goes to AS 64496 goes to
   AS 64500 instead.  In some deployments, the origination of the /22
   route is performed by AS 64496 and announced only to AS 64500, which
   then announces transit for that prefix.  This variation does not
   change the properties considered here.

   First, suppose the RPKI only had the minimal ROA for AS 64496, as
   described in Section 3.  However, if there is no ROA authorizing AS
   64500 to announce the /22 prefix, then the DDoS mitigation (and
   traffic scrubbing) scheme would not work.  That is, if AS 64500
   originates the /22 prefix in BGP during DDoS attacks, the
   announcement would be invalid [RFC6811].

   Therefore, the RPKI should have two ROAs: one for AS 64496 and one
   for AS 64500.

      ROA:(192.168.0.0/16, 192.168.225.0/24, AS 64496)

      ROA:(192.168.0.0/22, AS 64500)

   Neither ROA uses the maxLength attribute, but the second ROA is not
   "minimal" because it contains a /22 prefix that is not originated by
   anyone in BGP during normal operations.  The /22 prefix is only
   originated by AS 64500 as part of its DDoS mitigation service during
   a DDoS attack.

   Notice, however, that this scheme does not come without risks.
   Namely, all IP addresses in 192.168.0.0/22 are vulnerable to a
   forged-origin sub-prefix hijack during normal operations when the /22
   prefix is not originated.  (The hijacker AS 64511 would send the BGP
   announcement "192.168.0.0/22: AS 64511, AS 64500", falsely claiming
   that AS 64511 is a neighbor of AS 64500 and falsely claiming that AS
   64500 originates 192.168.0.0/22.)

   In some situations, the DDoS mitigation service at AS 64500 might
   want to limit the amount of DDoS traffic that it attracts and scrubs.
   Suppose that a DDoS attack only targets IP addresses in
   192.168.0.0/24.  Then, the DDoS mitigation service at AS 64500 only
   wants to attract the traffic designated for the /24 prefix that is
   under attack, but not the entire /22 prefix.  To allow for this, the
   RPKI should have two ROAs: one for AS 64496 and one for AS 64500.

      ROA:(192.168.0.0/16, 192.168.225.0/24, AS 64496)

      ROA:(192.168.0.0/22-24, AS 64500)

   The second ROA uses the maxLength attribute because it is designed to
   explicitly enable AS 64500 to originate any /24 sub-prefix of
   192.168.0.0/22.

   As before, the second ROA is not "minimal" because it contains
   prefixes that are not originated by anyone in BGP during normal
   operations.  Also, all IP addresses in 192.168.0.0/22 are vulnerable
   to a forged-origin sub-prefix hijack during normal operations when
   the /22 prefix is not originated.

   The use of the maxLength attribute in this second ROA also comes with
   additional risk.  While it permits the DDoS mitigation service at AS
   64500 to originate prefix 192.168.0.0/24 during a DDoS attack in that
   space, it also makes the other /24 prefixes covered by the /22 prefix
   (i.e., 192.168.1.0/24, 192.168.2.0/24, and 192.168.3.0/24) vulnerable
   to forged-origin sub-prefix attacks.

5.2.  Defensive De-aggregation in Response to Prefix Hijacks



   When responding to certain classes of prefix hijack (in particular,
   the forged-origin sub-prefix hijack described above), it may be
   desirable for the victim to perform "defensive de-aggregation", i.e.,
   to begin originating more-specific prefixes in order to compete with
   the hijack routes for selection as the best path in networks that are
   not performing RPKI-ROV [RFC6811].

   In topologies where at least one AS on every path between the victim
   and hijacker filters RPKI-ROV invalid prefixes, it may be the case
   that the existence of a minimal ROA issued by the victim prevents the
   defensive more-specific prefixes from being propagated to the
   networks topologically close to the attacker, thus hampering the
   effectiveness of this response.

   Nevertheless, this document recommends that, where possible, network
   operators publish minimal ROAs even in the face of this risk.  This
   is because:

   *  Minimal ROAs offer the best possible protection against the
      immediate impact of such an attack, rendering the need for such a
      response less likely;

   *  Increasing RPKI-ROV adoption by network operators will, over time,
      decrease the size of the neighborhoods in which this risk exists;
      and

   *  Other methods for reducing the size of such neighborhoods are
      available to potential victims, such as establishing direct
      External BGP (EBGP) adjacencies with networks from whom the
      defensive routes would otherwise be hidden.

6.  Considerations for RTDR Filtering Scenarios



   Considerations related to ROAs and RPKI-ROV [RFC6811] for the case of
   destination-based RTDR (elsewhere referred to as "Remotely Triggered
   Black Hole") filtering are addressed here.  In RTDR filtering, highly
   specific prefixes (greater than /24 in IPv4 and greater than /48 in
   IPv6, or possibly even /32 in IPv4 and /128 in IPv6) are announced in
   BGP.  These announcements are tagged with the well-known BGP
   community defined by [RFC7999].  For the reasons set out above, it is
   obviously not desirable to use a large maxLength value or include any
   such highly specific prefixes in the ROAs to accommodate destination-
   based RTDR filtering.

   As a result, RPKI-ROV [RFC6811] is a poor fit for the validation of
   RTDR routes.  Specification of new procedures to address this use
   case through the use of the RPKI is outside the scope of this
   document.

   Therefore:

   *  Operators SHOULD NOT create non-minimal ROAs (by either creating
      additional ROAs or using the maxLength attribute) for the purpose
      of advertising RTDR routes; and

   *  Operators providing a means for operators of neighboring
      autonomous systems to advertise RTDR routes via BGP MUST NOT make
      the creation of non-minimal ROAs a pre-requisite for its use.

7.  User Interface Design Recommendations



   Most operator interaction with the RPKI system when creating or
   modifying ROAs will occur via a user interface that abstracts the
   underlying encoding, signing, and publishing operations.

   This document recommends that designers and/or providers of such user
   interfaces SHOULD provide warnings to draw the user's attention to
   the risks of creating non-minimal ROAs in general and using the
   maxLength attribute in particular.

   Warnings provided by such a system may vary in nature from generic
   warnings based purely on the inclusion of the maxLength attribute to
   customised guidance based on the observable BGP routing policy of the
   operator in question.  The choices made in this respect are expected
   to be dependent on the target user audience of the implementation.

8.  Operational Considerations



   The recommendations specified in this document (in particular, those
   in Section 5) involve trade-offs between operational agility and
   security.

   Operators adopting the recommended practice of issuing minimal ROAs
   will, by definition, need to make changes to their existing set of
   issued ROAs in order to effect changes to the set of prefixes that
   are originated in BGP.

   Even in the case of routing changes that are planned in advance,
   existing procedures may need to be updated to incorporate changes to
   issued ROAs and may require additional time allowed for those changes
   to propagate.

   Operators are encouraged to carefully review the issues highlighted
   (especially those in Sections 5.1 and 5.2) in light of their specific
   operational requirements.  Failure to do so could, in the worst case,
   result in a self-inflicted denial of service.

   The recommendations made in Section 5 are likely to be more onerous
   for operators utilising large IP address space allocations from which
   many more-specific advertisements are made in BGP.  Operators of such
   networks are encouraged to seek opportunities to automate the
   required procedures in order to minimise manual operational burden.

9.  Security Considerations



   This document makes recommendations regarding the use of RPKI-ROV as
   defined in [RFC6811] and, as such, introduces no additional security
   considerations beyond those specified therein.

10.  IANA Considerations



   This document has no IANA actions.

11.  References



11.1.  Normative References



   [RFC1918]  Rekhter, Y., Moskowitz, B., Karrenberg, D., de Groot, G.
              J., and E. Lear, "Address Allocation for Private
              Internets", BCP 5, RFC 1918, DOI 10.17487/RFC1918,
              February 1996, <https://www.rfc-editor.org/info/rfc1918>.

   [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>.

   [RFC4271]  Rekhter, Y., Ed., Li, T., Ed., and S. Hares, Ed., "A
              Border Gateway Protocol 4 (BGP-4)", RFC 4271,
              DOI 10.17487/RFC4271, January 2006,
              <https://www.rfc-editor.org/info/rfc4271>.

   [RFC6480]  Lepinski, M. and S. Kent, "An Infrastructure to Support
              Secure Internet Routing", RFC 6480, DOI 10.17487/RFC6480,
              February 2012, <https://www.rfc-editor.org/info/rfc6480>.

   [RFC6482]  Lepinski, M., Kent, S., and D. Kong, "A Profile for Route
              Origin Authorizations (ROAs)", RFC 6482,
              DOI 10.17487/RFC6482, February 2012,
              <https://www.rfc-editor.org/info/rfc6482>.

   [RFC6811]  Mohapatra, P., Scudder, J., Ward, D., Bush, R., and R.
              Austein, "BGP Prefix Origin Validation", RFC 6811,
              DOI 10.17487/RFC6811, January 2013,
              <https://www.rfc-editor.org/info/rfc6811>.

   [RFC7115]  Bush, R., "Origin Validation Operation Based on the
              Resource Public Key Infrastructure (RPKI)", BCP 185,
              RFC 7115, DOI 10.17487/RFC7115, January 2014,
              <https://www.rfc-editor.org/info/rfc7115>.

   [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>.

11.2.  Informative References



   [GCHSS]    Gilad, Y., Cohen, A., Herzberg, A., Schapira, M., and H.
              Shulman, "Are We There Yet? On RPKI's Deployment and
              Security", NDSS 2017, February 2017,
              <https://eprint.iacr.org/2016/1010.pdf>.

   [GSG17]    Gilad, Y., Sagga, O., and S. Goldberg, "MaxLength
              Considered Harmful to the RPKI", CoNEXT '17,
              DOI 10.1145/3143361.3143363, December 2017,
              <https://eprint.iacr.org/2016/1015.pdf>.

   [LSG16]    Lychev, R., Shapira, M., and S. Goldberg, "Rethinking
              security for internet routing", Communications of the ACM,
              DOI 10.1145/2896817, October 2016, <http://cacm.acm.org/
              magazines/2016/10/207763-rethinking-security-for-internet-
              routing/>.

   [RFC5737]  Arkko, J., Cotton, M., and L. Vegoda, "IPv4 Address Blocks
              Reserved for Documentation", RFC 5737,
              DOI 10.17487/RFC5737, January 2010,
              <https://www.rfc-editor.org/info/rfc5737>.

   [RFC6907]  Manderson, T., Sriram, K., and R. White, "Use Cases and
              Interpretations of Resource Public Key Infrastructure
              (RPKI) Objects for Issuers and Relying Parties", RFC 6907,
              DOI 10.17487/RFC6907, March 2013,
              <https://www.rfc-editor.org/info/rfc6907>.

   [RFC7999]  King, T., Dietzel, C., Snijders, J., Doering, G., and G.
              Hankins, "BLACKHOLE Community", RFC 7999,
              DOI 10.17487/RFC7999, October 2016,
              <https://www.rfc-editor.org/info/rfc7999>.

   [RFC8205]  Lepinski, M., Ed. and K. Sriram, Ed., "BGPsec Protocol
              Specification", RFC 8205, DOI 10.17487/RFC8205, September
              2017, <https://www.rfc-editor.org/info/rfc8205>.

Acknowledgments

   The authors would like to thank the following people for their review
   and contributions to this document: Omar Sagga and Aris Lambrianidis.
   Thanks are also due to Matthias Waehlisch, Ties de Kock, Amreesh
   Phokeer, Éric Vyncke, Alvaro Retana, John Scudder, Roman Danyliw,
   Andrew Alston, and Murray Kucherawy for comments and suggestions, to
   Roni Even for the Gen-ART review, to Jean Mahoney for the ART-ART
   review, to Acee Lindem for the Routing Area Directorate review, and
   to Sean Turner for the Security Area Directorate review.

Authors' Addresses



   Yossi Gilad
   Hebrew University of Jerusalem
   Rothburg Family Buildings
   Edmond J. Safra Campus
   Jerusalem 9190416
   Israel
   Email: yossigi@cs.huji.ac.il


   Sharon Goldberg
   Boston University
   111 Cummington St, MCS135
   Boston, MA 02215
   United States of America
   Email: goldbe@cs.bu.edu


   Kotikalapudi Sriram
   USA National Institute of Standards and Technology
   100 Bureau Drive
   Gaithersburg, MD 20899
   United States of America
   Email: kotikalapudi.sriram@nist.gov


   Job Snijders
   Fastly
   Amsterdam
   Netherlands
   Email: job@fastly.com


   Ben Maddison
   Workonline Communications
   114 West St
   Johannesburg
   2196
   South Africa
   Email: benm@workonline.africa