| 1 |
<pre>Internet Engineering Task Force (IETF) S. Kent |
| 2 |
Request for Comments: 7132 BBN |
| 3 |
Category: Informational A. Chi |
| 4 |
ISSN: 2070-1721 UNC-CH |
| 5 |
February 2014 |
| 6 |
|
| 7 |
|
| 8 |
<span class="h1">Threat Model for BGP Path Security</span> |
| 9 |
|
| 10 |
Abstract |
| 11 |
|
| 12 |
This document describes a threat model for the context in which |
| 13 |
External Border Gateway Protocol (EBGP) path security mechanisms will |
| 14 |
be developed. The threat model includes an analysis of the Resource |
| 15 |
Public Key Infrastructure (RPKI) and focuses on the ability of an |
| 16 |
Autonomous System (AS) to verify the authenticity of the AS path info |
| 17 |
received in a BGP update. We use the term "PATHSEC" to refer to any |
| 18 |
BGP path security technology that makes use of the RPKI. PATHSEC |
| 19 |
will secure BGP, consistent with the inter-AS security focus of the |
| 20 |
RPKI. |
| 21 |
|
| 22 |
The document characterizes classes of potential adversaries that are |
| 23 |
considered to be threats and examines classes of attacks that might |
| 24 |
be launched against PATHSEC. It does not revisit attacks against |
| 25 |
unprotected BGP, as that topic has already been addressed in the |
| 26 |
BGP-4 standard. It concludes with a brief discussion of residual |
| 27 |
vulnerabilities. |
| 28 |
|
| 29 |
Status of This Memo |
| 30 |
|
| 31 |
This document is not an Internet Standards Track specification; it is |
| 32 |
published for informational purposes. |
| 33 |
|
| 34 |
This document is a product of the Internet Engineering Task Force |
| 35 |
(IETF). It represents the consensus of the IETF community. It has |
| 36 |
received public review and has been approved for publication by the |
| 37 |
Internet Engineering Steering Group (IESG). Not all documents |
| 38 |
approved by the IESG are a candidate for any level of Internet |
| 39 |
Standard; see <a href="./rfc5741#section-2">Section 2 of RFC 5741</a>. |
| 40 |
|
| 41 |
Information about the current status of this document, any errata, |
| 42 |
and how to provide feedback on it may be obtained at |
| 43 |
<a href="http://www.rfc-editor.org/info/rfc7132">http://www.rfc-editor.org/info/rfc7132</a>. |
| 44 |
|
| 45 |
|
| 46 |
|
| 47 |
|
| 48 |
|
| 49 |
|
| 50 |
|
| 51 |
|
| 52 |
<span class="grey">Kent & Chi Informational [Page 1]</span></pre> |
| 53 |
<hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-2" ></span> |
| 54 |
<span class="grey"><a href="./rfc7132">RFC 7132</a> Threat Model for BGP Path Security February 2014</span> |
| 55 |
|
| 56 |
|
| 57 |
Copyright Notice |
| 58 |
|
| 59 |
Copyright (c) 2014 IETF Trust and the persons identified as the |
| 60 |
document authors. All rights reserved. |
| 61 |
|
| 62 |
This document is subject to <a href="https://www.rfc-editor.org/bcp/bcp78">BCP 78</a> and the IETF Trust's Legal |
| 63 |
Provisions Relating to IETF Documents |
| 64 |
(<a href="http://trustee.ietf.org/license-info">http://trustee.ietf.org/license-info</a>) in effect on the date of |
| 65 |
publication of this document. Please review these documents |
| 66 |
carefully, as they describe your rights and restrictions with respect |
| 67 |
to this document. Code Components extracted from this document must |
| 68 |
include Simplified BSD License text as described in Section 4.e of |
| 69 |
the Trust Legal Provisions and are provided without warranty as |
| 70 |
described in the Simplified BSD License. |
| 71 |
|
| 72 |
Table of Contents |
| 73 |
|
| 74 |
<a href="#section-1">1</a>. Introduction . . . . . . . . . . . . . . . . . . . . . . . . <a href="#page-2">2</a> |
| 75 |
<a href="#section-2">2</a>. Terminology . . . . . . . . . . . . . . . . . . . . . . . . . <a href="#page-4">4</a> |
| 76 |
<a href="#section-3">3</a>. Threat Characterization . . . . . . . . . . . . . . . . . . . <a href="#page-6">6</a> |
| 77 |
<a href="#section-4">4</a>. Attack Characterization . . . . . . . . . . . . . . . . . . . <a href="#page-8">8</a> |
| 78 |
<a href="#section-4.1">4.1</a>. Active Wiretapping of Sessions between Routers . . . . . <a href="#page-8">8</a> |
| 79 |
<a href="#section-4.2">4.2</a>. Attacks on a BGP Router . . . . . . . . . . . . . . . . . <a href="#page-9">9</a> |
| 80 |
4.3. Attacks on Network Operator Management Computers (Non-CA |
| 81 |
Computers) . . . . . . . . . . . . . . . . . . . . . . . <a href="#page-11">11</a> |
| 82 |
<a href="#section-4.4">4.4</a>. Attacks on a Repository Publication Point . . . . . . . . <a href="#page-12">12</a> |
| 83 |
<a href="#section-4.5">4.5</a>. Attacks on an RPKI CA . . . . . . . . . . . . . . . . . . <a href="#page-14">14</a> |
| 84 |
<a href="#section-5">5</a>. Residual Vulnerabilities . . . . . . . . . . . . . . . . . . <a href="#page-16">16</a> |
| 85 |
<a href="#section-6">6</a>. Security Considerations . . . . . . . . . . . . . . . . . . . <a href="#page-18">18</a> |
| 86 |
<a href="#section-7">7</a>. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . <a href="#page-18">18</a> |
| 87 |
<a href="#section-8">8</a>. Informative References . . . . . . . . . . . . . . . . . . . <a href="#page-18">18</a> |
| 88 |
|
| 89 |
<span class="h2"><a class="selflink" id="section-1" href="#section-1">1</a>. Introduction</span> |
| 90 |
|
| 91 |
This document describes the security context in which PATHSEC is |
| 92 |
intended to operate. The term "PATHSEC" (for path security) refers |
| 93 |
to any design used to preserve the integrity and authenticity of the |
| 94 |
AS_PATH attribute carried in a BGP update message [<a href="./rfc4271" title=""A Border Gateway Protocol 4 (BGP-4)"">RFC4271</a>]. The |
| 95 |
security context used throughout this document is established by the |
| 96 |
Secure Inter-Domain Routing (SIDR) working group charter [<a href="#ref-SIDR-CH" title=""Secure Inter-Domain Routing: Charter for Working Group"">SIDR-CH</a>]. |
| 97 |
The charter requires that solutions that afford PATHSEC make use of |
| 98 |
the Resource Public Key Infrastructure (RPKI) [<a href="./rfc6480" title=""An Infrastructure to Support Secure Internet Routing"">RFC6480</a>]. It also |
| 99 |
calls for protecting only the information required to verify that a |
| 100 |
received route traversed the Autonomous Systems (ASes) in question, |
| 101 |
and that the Network Layer Reachability Information (NLRI) in the |
| 102 |
route is what was advertised. |
| 103 |
|
| 104 |
|
| 105 |
|
| 106 |
|
| 107 |
|
| 108 |
<span class="grey">Kent & Chi Informational [Page 2]</span></pre> |
| 109 |
<hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-3" ></span> |
| 110 |
<span class="grey"><a href="./rfc7132">RFC 7132</a> Threat Model for BGP Path Security February 2014</span> |
| 111 |
|
| 112 |
|
| 113 |
Thus, the goal of PATHSEC is to enable a BGP speaker to verify that |
| 114 |
the ASes enumerated in this path attribute represent the sequence of |
| 115 |
ASes that the NLRI traversed. The term "PATHSEC" is thus consistent |
| 116 |
with the goal described above. (Other SIDR documents use the term |
| 117 |
"BGPSEC" to refer to a specific design; we avoid use of that term |
| 118 |
here.) |
| 119 |
|
| 120 |
This document discusses classes of potential adversaries that are |
| 121 |
considered to be threats, and classes of attacks that might be |
| 122 |
launched against PATHSEC. Because PATHSEC will rely on the RPKI, |
| 123 |
threats and attacks against the RPKI are included. This model also |
| 124 |
takes into consideration classes of attacks that are enabled by the |
| 125 |
use of PATHSEC (e.g., based on use of the RPKI). |
| 126 |
|
| 127 |
The motivation for developing PATHSEC, i.e., residual security |
| 128 |
concerns for BGP, is well described in several documents, including |
| 129 |
"BGP Security Vulnerabilities Analysis" [<a href="./rfc4272" title=""BGP Security Vulnerabilities Analysis"">RFC4272</a>] and "Design and |
| 130 |
Analysis of the Secure Border Gateway Protocol (S-BGP)" [<a href="#ref-Kent2000">Kent2000</a>]. |
| 131 |
All of these documents note that BGP does not include mechanisms that |
| 132 |
allow an AS to verify the legitimacy and authenticity of BGP route |
| 133 |
advertisements. (BGP now mandates support for mechanisms to secure |
| 134 |
peer-to-peer communication, i.e., for the links that connect BGP |
| 135 |
routers. There are several secure protocol options to address this |
| 136 |
security concern, e.g., IPsec [<a href="./rfc4301" title=""Security Architecture for the Internet Protocol"">RFC4301</a>] and TCP Authentication Option |
| 137 |
(TCP-AO) [<a href="./rfc5925" title=""The TCP Authentication Option"">RFC5925</a>]. This document briefly notes the need to address |
| 138 |
this aspect of BGP security, but focuses on application layer BGP |
| 139 |
security issues that must be addressed by PATHSEC.) |
| 140 |
|
| 141 |
<a href="./rfc4272">RFC 4272</a> [<a href="./rfc4272" title=""BGP Security Vulnerabilities Analysis"">RFC4272</a>] succinctly notes: |
| 142 |
|
| 143 |
BGP speakers themselves can inject bogus routing information, |
| 144 |
either by masquerading as any other legitimate BGP speaker, or by |
| 145 |
distributing unauthorized routing information as themselves. |
| 146 |
Historically, misconfigured and faulty routers have been |
| 147 |
responsible for widespread disruptions in the Internet. The |
| 148 |
legitimate BGP peers have the context and information to produce |
| 149 |
believable, yet bogus, routing information, and therefore have the |
| 150 |
opportunity to cause great damage. The cryptographic protections |
| 151 |
of [TCPMD5] and operational protections cannot exclude the bogus |
| 152 |
information arising from a legitimate peer. The risk of |
| 153 |
disruptions caused by legitimate BGP speakers is real and cannot |
| 154 |
be ignored. |
| 155 |
|
| 156 |
PATHSEC is intended to address the concerns cited above, to provide |
| 157 |
significantly improved path security, which builds upon the route |
| 158 |
origination validation capability offered by use of the RPKI |
| 159 |
[<a href="./rfc6810" title=""The Resource Public Key Infrastructure (RPKI) to Router Protocol"">RFC6810</a>]. Specifically, the RPKI enables relying parties (RPs) to |
| 160 |
determine if the origin AS for a path was authorized to advertise the |
| 161 |
|
| 162 |
|
| 163 |
|
| 164 |
<span class="grey">Kent & Chi Informational [Page 3]</span></pre> |
| 165 |
<hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-4" ></span> |
| 166 |
<span class="grey"><a href="./rfc7132">RFC 7132</a> Threat Model for BGP Path Security February 2014</span> |
| 167 |
|
| 168 |
|
| 169 |
prefix contained in a BGP update message. This security feature is |
| 170 |
enabled by the use of two types of digitally signed data: a PKI |
| 171 |
[<a href="./rfc6487" title=""A Profile for X.509 PKIX Resource Certificates"">RFC6487</a>] that associates one or more prefixes with the public key(s) |
| 172 |
of an address space holder, and Route Origin Authorizations (ROAs) |
| 173 |
[<a href="./rfc6482" title=""A Profile for Route Origin Authorizations (ROAs)"">RFC6482</a>] that allow a prefix holder to specify one or more ASes that |
| 174 |
are authorized to originate routes for a prefix. |
| 175 |
|
| 176 |
The security model adopted for PATHSEC does not assume an "oracle" |
| 177 |
that can see all of the BGP inputs and outputs associated with every |
| 178 |
AS or every BGP router. Instead, the model is based on a local |
| 179 |
notion of what constitutes legitimate, authorized behavior by the BGP |
| 180 |
routers associated with an AS. This is an AS-centric model of secure |
| 181 |
operation, consistent with the AS-centric model that BGP employs for |
| 182 |
routing. This model forms the basis for the discussion that follows. |
| 183 |
|
| 184 |
This document begins with a brief set of definitions relevant to the |
| 185 |
subsequent sections. It then discusses classes of adversaries that |
| 186 |
are perceived as viable threats against routing in the public |
| 187 |
Internet. It continues to explore a range of attacks that might be |
| 188 |
effected by these adversaries against both path security and the |
| 189 |
infrastructure upon which PATHSEC relies. It concludes with a brief |
| 190 |
review of residual vulnerabilities, i.e., vulnerabilities that are |
| 191 |
not addressed by use of the RPKI and that appear likely to be outside |
| 192 |
the scope of PATHSEC mechanisms. |
| 193 |
|
| 194 |
<span class="h2"><a class="selflink" id="section-2" href="#section-2">2</a>. Terminology</span> |
| 195 |
|
| 196 |
The following security and routing terminology definitions are |
| 197 |
employed in this document. |
| 198 |
|
| 199 |
Adversary: An adversary is an entity (e.g., a person or an |
| 200 |
organization) that is perceived as malicious, relative to the |
| 201 |
security policy of a system. The decision to characterize an |
| 202 |
entity as an adversary is made by those responsible for the |
| 203 |
security of a system. Often, one describes classes of adversaries |
| 204 |
with similar capabilities or motivations rather than specific |
| 205 |
individuals or organizations. |
| 206 |
|
| 207 |
Attack: An attack is an action that attempts to violate the security |
| 208 |
policy of a system, e.g., by exploiting a vulnerability. There is |
| 209 |
often a many-to-one mapping of attacks to vulnerabilities because |
| 210 |
many different attacks may be used to exploit a vulnerability. |
| 211 |
|
| 212 |
Autonomous System (AS): An AS is a set of one or more IP networks |
| 213 |
operated by a single administrative entity. |
| 214 |
|
| 215 |
AS Number (ASN): An ASN is a 2- or 4-byte number issued by a |
| 216 |
registry to identify an AS in BGP. |
| 217 |
|
| 218 |
|
| 219 |
|
| 220 |
<span class="grey">Kent & Chi Informational [Page 4]</span></pre> |
| 221 |
<hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-5" ></span> |
| 222 |
<span class="grey"><a href="./rfc7132">RFC 7132</a> Threat Model for BGP Path Security February 2014</span> |
| 223 |
|
| 224 |
|
| 225 |
Certification Authority (CA): An entity that issues digital |
| 226 |
certificates (e.g., X.509 certificates) and vouches for the |
| 227 |
binding between the data items in a certificate. |
| 228 |
|
| 229 |
Countermeasure: A countermeasure is a procedure or technique that |
| 230 |
thwarts an attack, preventing it from being successful. Often, |
| 231 |
countermeasures are specific to attacks or classes of attacks. |
| 232 |
|
| 233 |
Border Gateway Protocol (BGP): A path vector protocol used to convey |
| 234 |
"reachability" information among ASes in support of inter-domain |
| 235 |
routing. |
| 236 |
|
| 237 |
False (Route) Origination: If a network operator originates a route |
| 238 |
for a prefix that the operator does not hold (and that has not |
| 239 |
been authorized to originate by the prefix holder), this is termed |
| 240 |
false route origination. |
| 241 |
|
| 242 |
Internet Service Provider (ISP): An organization managing (and |
| 243 |
typically selling) Internet services to other organizations or |
| 244 |
individuals. |
| 245 |
|
| 246 |
Internet Number Resources (INRs): IPv4 or IPv6 address space and |
| 247 |
ASNs. |
| 248 |
|
| 249 |
Internet Registry: An organization that manages the allocation or |
| 250 |
distribution of INRs. This encompasses the Internet Assigned |
| 251 |
Number Authority (IANA), Regional Internet Registries (RIRs), |
| 252 |
National Internet Registries (NIRs), and Local Internet Registries |
| 253 |
(LIRs) (network operators). |
| 254 |
|
| 255 |
Man in the Middle (MITM): A MITM is an entity that is able to |
| 256 |
examine and modify traffic between two (or more) parties on a |
| 257 |
communication path. |
| 258 |
|
| 259 |
Network Operator: An entity that manages an AS and thus emits (E)BGP |
| 260 |
updates, e.g., an ISP. |
| 261 |
|
| 262 |
Network Operations Center (NOC): A network operator employs a set of |
| 263 |
equipment and a staff to manage a network, typically on a 24/7 |
| 264 |
basis. The equipment and staff are often referred to as the NOC |
| 265 |
for the network. |
| 266 |
|
| 267 |
Prefix: A prefix is an IP address and a mask used to specify a set |
| 268 |
of addresses that are grouped together for purposes of routing. |
| 269 |
|
| 270 |
Public Key Infrastructure (PKI): A PKI is a collection of hardware, |
| 271 |
software, people, policies, and procedures used to create, manage, |
| 272 |
distribute, store, and revoke digital certificates. |
| 273 |
|
| 274 |
|
| 275 |
|
| 276 |
<span class="grey">Kent & Chi Informational [Page 5]</span></pre> |
| 277 |
<hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-6" ></span> |
| 278 |
<span class="grey"><a href="./rfc7132">RFC 7132</a> Threat Model for BGP Path Security February 2014</span> |
| 279 |
|
| 280 |
|
| 281 |
Relying Parties (RPs): An RP is an entity that makes use of signed |
| 282 |
products from a PKI, i.e., it relies on signed data that is |
| 283 |
verified using certificates and Certificate Revocation Lists |
| 284 |
(CRLs) from a PKI. |
| 285 |
|
| 286 |
RPKI Repository System: The RPKI repository system consists of a |
| 287 |
distributed set of loosely synchronized databases. |
| 288 |
|
| 289 |
Resource PKI (RPKI): A PKI operated by the entities that manage INRs |
| 290 |
and that issue X.509 certificates (and CRLs) that attest to the |
| 291 |
holdings of INRs. |
| 292 |
|
| 293 |
RPKI Signed Object: An RPKI signed object is a data object |
| 294 |
encapsulated with Cryptographic Message Syntax (CMS) that complies |
| 295 |
with the format and semantics defined in [<a href="./rfc6488" title=""Signed Object Template for the Resource Public Key Infrastructure (RPKI)"">RFC6488</a>]. |
| 296 |
|
| 297 |
Route: In the Internet, a route is a prefix and an associated |
| 298 |
sequence of ASNs that indicates a path via which traffic destined |
| 299 |
for the prefix can be directed. (The route includes the origin |
| 300 |
AS.) |
| 301 |
|
| 302 |
Route Leak: A route leak is said to occur when AS-A advertises |
| 303 |
routes that it has received from AS-B to the neighbors of AS-A, |
| 304 |
but AS-A is not viewed as a transit provider for the prefixes in |
| 305 |
the route. |
| 306 |
|
| 307 |
Threat: A threat is a motivated, capable adversary. An adversary |
| 308 |
that is not motivated to launch an attack is not a threat. An |
| 309 |
adversary that is motivated but not capable of launching an attack |
| 310 |
also is not a threat. |
| 311 |
|
| 312 |
Vulnerability: A vulnerability is a flaw or weakness in a system's |
| 313 |
design, implementation, or operation and management that could be |
| 314 |
exploited to violate the security policy of a system. |
| 315 |
|
| 316 |
<span class="h2"><a class="selflink" id="section-3" href="#section-3">3</a>. Threat Characterization</span> |
| 317 |
|
| 318 |
As noted in <a href="#section-2">Section 2</a> above, a threat is defined as a motivated, |
| 319 |
capable adversary. The following classes of threats represent |
| 320 |
classes of adversaries viewed as relevant to this environment. |
| 321 |
|
| 322 |
Network Operators: A network operator may be a threat. An |
| 323 |
operator may be motivated to cause BGP routers it controls to emit |
| 324 |
update messages with inaccurate routing info, e.g., to cause |
| 325 |
traffic to flow via paths that are economically advantageous for |
| 326 |
the operator. Such updates might cause traffic to flow via paths |
| 327 |
that would otherwise be rejected as less advantageous by other |
| 328 |
network operators. Because an operator controls the BGP routers |
| 329 |
|
| 330 |
|
| 331 |
|
| 332 |
<span class="grey">Kent & Chi Informational [Page 6]</span></pre> |
| 333 |
<hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-7" ></span> |
| 334 |
<span class="grey"><a href="./rfc7132">RFC 7132</a> Threat Model for BGP Path Security February 2014</span> |
| 335 |
|
| 336 |
|
| 337 |
in its network, it is in a position to modify their operation in |
| 338 |
arbitrary ways. Routers managed by a network operator are |
| 339 |
vehicles for mounting MITM attacks on both control and data plane |
| 340 |
traffic. If an operator participates in the RPKI, it will have at |
| 341 |
least one CA resource certificate and may be able to generate an |
| 342 |
arbitrary number of subordinate CA certificates and ROAs. It will |
| 343 |
be authorized to populate (and may even host) its own repository |
| 344 |
publication point. If it implements PATHSEC, and if PATHSEC makes |
| 345 |
use of certificates associated with routers or ASes, it will have |
| 346 |
the ability to issue such certificates for itself. If PATHSEC |
| 347 |
digitally signs updates, it will be able to do so in a fashion |
| 348 |
that will be accepted by PATHSEC-enabled neighbors. |
| 349 |
|
| 350 |
Hackers: Hackers are considered a threat. A hacker might assume |
| 351 |
control of network management computers and routers controlled by |
| 352 |
operators, including operators that implement PATHSEC. In such |
| 353 |
cases, hackers would be able to act as rogue network operators |
| 354 |
(see above). It is assumed that hackers generally do not have the |
| 355 |
capability to effect MITM attacks on most links between networks |
| 356 |
(links used to transmit BGP and subscriber traffic). A hacker |
| 357 |
might be recruited, without his/her knowledge, by criminals or by |
| 358 |
nations, to act on their behalf. Hackers may be motivated by a |
| 359 |
desire for "bragging rights", for profit, or to express support |
| 360 |
for a cause ("hacktivists" [<a href="#ref-Sam04" title=""Hacktivism and the Future of Political Participation"">Sam04</a>]). We view hackers as possibly |
| 361 |
distinct from criminals in that the former are presumed to effect |
| 362 |
attacks only remotely (not via a physical presence associated with |
| 363 |
a target) and not necessarily for monetary gain. Some hackers may |
| 364 |
commit criminal acts (depending on the jurisdiction), and thus |
| 365 |
there is a potential for overlap between this adversary group and |
| 366 |
criminals. |
| 367 |
|
| 368 |
Criminals: Criminals may be a threat. Criminals might persuade |
| 369 |
(via threats or extortion) a network operator to act as a rogue |
| 370 |
operator (see above) and thus be able to effect a wide range of |
| 371 |
attacks. Criminals might persuade the staff of a |
| 372 |
telecommunications provider to enable MITM attacks on links |
| 373 |
between routers. Motivations for criminals may include the |
| 374 |
ability to extort money from network operators or network operator |
| 375 |
clients, e.g., by adversely affecting routing for these network |
| 376 |
operators or their clients. Criminals also may wish to manipulate |
| 377 |
routing to conceal the sources of spam, DoS attacks, or other |
| 378 |
criminal activities. |
| 379 |
|
| 380 |
Registries: Any registry in the RPKI could be a threat. Staff at |
| 381 |
the registry are capable of manipulating repository content or |
| 382 |
mismanaging the RPKI certificates that they issue. These actions |
| 383 |
could adversely affect a network operator or a client of a network |
| 384 |
|
| 385 |
|
| 386 |
|
| 387 |
|
| 388 |
<span class="grey">Kent & Chi Informational [Page 7]</span></pre> |
| 389 |
<hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-8" ></span> |
| 390 |
<span class="grey"><a href="./rfc7132">RFC 7132</a> Threat Model for BGP Path Security February 2014</span> |
| 391 |
|
| 392 |
|
| 393 |
operator. The staff could be motivated to do this based on |
| 394 |
political pressure from the nation in which the registry operates |
| 395 |
(see below) or due to criminal influence (see above). |
| 396 |
|
| 397 |
Nations: A nation may be a threat. A nation may control one or |
| 398 |
more network operators that operate in the nation, and thus can |
| 399 |
cause them to act as rogue network operators. A nation may have a |
| 400 |
technical active wiretapping capability (e.g., within its |
| 401 |
territory) that enables it to effect MITM attacks on inter-network |
| 402 |
traffic. (This capability may be facilitated by control or |
| 403 |
influence over a telecommunications provider operating within the |
| 404 |
nation.) It may have an ability to attack and take control of |
| 405 |
routers or management network computers of network operators in |
| 406 |
other countries. A nation may control a registry (e.g., an RIR) |
| 407 |
that operates within its territory and might force that registry |
| 408 |
to act in a rogue capacity. National threat motivations include |
| 409 |
the desire to control the flow of traffic to/from the nation or to |
| 410 |
divert traffic destined for other nations (for passive or active |
| 411 |
wiretapping, including DoS). |
| 412 |
|
| 413 |
<span class="h2"><a class="selflink" id="section-4" href="#section-4">4</a>. Attack Characterization</span> |
| 414 |
|
| 415 |
This section describes classes of attacks that may be effected |
| 416 |
against Internet routing (relative to the context described in |
| 417 |
<a href="#section-1">Section 1</a>). Attacks are classified based on the target of the |
| 418 |
attack, an element of the routing system, or the routing security |
| 419 |
infrastructure on which PATHSEC relies. In general, attacks of |
| 420 |
interest are ones that attempt to violate the integrity or |
| 421 |
authenticity of BGP traffic or that violate the authorizations |
| 422 |
associated with entities participating in the RPKI. Attacks that |
| 423 |
violate the implied confidentiality of routing traffic, e.g., passive |
| 424 |
wiretapping attacks, are not considered a requirement for BGP |
| 425 |
security (see [<a href="./rfc4272" title=""BGP Security Vulnerabilities Analysis"">RFC4272</a>]). |
| 426 |
|
| 427 |
<span class="h3"><a class="selflink" id="section-4.1" href="#section-4.1">4.1</a>. Active Wiretapping of Sessions between Routers</span> |
| 428 |
|
| 429 |
An adversary may attack the BGP (TCP) session that connects a pair of |
| 430 |
BGP speakers. An active attack against a BGP (TCP) session can be |
| 431 |
effected by directing traffic to a BGP speaker from some remote |
| 432 |
point, or by being positioned as a MITM on the link that carries BGP |
| 433 |
session traffic. Remote attacks can be effected by any adversary. A |
| 434 |
MITM attack requires access to the link. Modern transport networks |
| 435 |
may be as complex as the packet networks that utilize them for inter- |
| 436 |
AS links. Thus, these transport networks may present significant |
| 437 |
attack surfaces. Nonetheless, only some classes of adversaries are |
| 438 |
assumed to be capable of MITM attacks against a BGP session. MITM |
| 439 |
attacks may be directed against BGP and PATHSEC-protected BGP, or |
| 440 |
against TCP or IP. Such attacks include replay of selected BGP |
| 441 |
|
| 442 |
|
| 443 |
|
| 444 |
<span class="grey">Kent & Chi Informational [Page 8]</span></pre> |
| 445 |
<hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-9" ></span> |
| 446 |
<span class="grey"><a href="./rfc7132">RFC 7132</a> Threat Model for BGP Path Security February 2014</span> |
| 447 |
|
| 448 |
|
| 449 |
messages, selective modification of BGP messages, and DoS attacks |
| 450 |
against BGP routers. [<a href="./rfc4272" title=""BGP Security Vulnerabilities Analysis"">RFC4272</a>] describes several countermeasures for |
| 451 |
such attacks, and thus this document does not further address such |
| 452 |
attacks. |
| 453 |
|
| 454 |
<span class="h3"><a class="selflink" id="section-4.2" href="#section-4.2">4.2</a>. Attacks on a BGP Router</span> |
| 455 |
|
| 456 |
An adversary may attack a BGP router, whether or not it implements |
| 457 |
PATHSEC. Any adversary that controls routers legitimately, or that |
| 458 |
can assume control of a router, is assumed to be able to effect the |
| 459 |
types of attacks described below. Note that any router behavior that |
| 460 |
can be ascribed to a local routing policy decision is not considered |
| 461 |
to be an attack. This is because such behavior could be explained as |
| 462 |
a result of local policy settings and thus is beyond the scope of |
| 463 |
what PATHSEC can detect as unauthorized behavior. Thus, for example, |
| 464 |
a router may fail to propagate some or all route withdrawals or |
| 465 |
effect "route leaks". (These behaviors are not precluded by the |
| 466 |
specification for BGP and might be the result of a local policy that |
| 467 |
is not publicly disclosed. As a result, they are not considered |
| 468 |
attacks. See <a href="#section-5">Section 5</a> for additional discussion.) |
| 469 |
|
| 470 |
Attacks on a router are equivalent to active wiretapping attacks (in |
| 471 |
the most general sense) that manipulate (forge, tamper with, or |
| 472 |
suppress) data contained in BGP updates. The list below illustrates |
| 473 |
attacks of this type. |
| 474 |
|
| 475 |
AS Insertion: A router might insert one or more ASNs, other than |
| 476 |
its own ASN, into an update message. This violates the BGP spec |
| 477 |
and thus is considered an attack. |
| 478 |
|
| 479 |
False (Route) Origination: A router might originate a route for a |
| 480 |
prefix when the AS that the router represents is not authorized to |
| 481 |
originate routes for that prefix. This is an attack, but it is |
| 482 |
addressed by the use of the RPKI [<a href="./rfc6480" title=""An Infrastructure to Support Secure Internet Routing"">RFC6480</a>]. |
| 483 |
|
| 484 |
Secure Path Downgrade: A router might remove AS_PATH data from a |
| 485 |
PATHSEC-protected update that it receives when forwarding this |
| 486 |
update to a PATHSEC-enabled neighbor. This behavior violates the |
| 487 |
PATHSEC security goals and thus is considered an attack. |
| 488 |
|
| 489 |
Invalid AS_PATH Data Insertion: A router might emit a PATHSEC- |
| 490 |
protected update with "bad" data (such as a signature), i.e., |
| 491 |
PATHSEC data that cannot be validated by other PATHSEC routers. |
| 492 |
Such behavior is assumed to violate the PATHSEC goals and thus is |
| 493 |
considered an attack. |
| 494 |
|
| 495 |
|
| 496 |
|
| 497 |
|
| 498 |
|
| 499 |
|
| 500 |
<span class="grey">Kent & Chi Informational [Page 9]</span></pre> |
| 501 |
<hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-10" ></span> |
| 502 |
<span class="grey"><a href="./rfc7132">RFC 7132</a> Threat Model for BGP Path Security February 2014</span> |
| 503 |
|
| 504 |
|
| 505 |
Stale Path Announcement: If PATHSEC-secured announcements can |
| 506 |
expire, such an announcement may be propagated with PATHSEC data |
| 507 |
that is "expired". This behavior would violate the PATHSEC goals |
| 508 |
and is considered a type of replay attack. |
| 509 |
|
| 510 |
Premature Path Announcement Expiration: If a PATHSEC-secured |
| 511 |
announcement has an associated expiration time, a router might |
| 512 |
emit a PATHSEC-secured announcement with an expiry time that is |
| 513 |
very short. Unless the PATHSEC protocol specification mandates a |
| 514 |
minimum expiry time, this is not an attack. However, if such a |
| 515 |
time is mandated, this behavior becomes an attack. BGP speakers |
| 516 |
along a path generally cannot determine if an expiry time is |
| 517 |
"suspiciously short" since they cannot know how long a route may |
| 518 |
have been held by an earlier AS, prior to being released. |
| 519 |
|
| 520 |
MITM Attack: A cryptographic key used for point-to-point security |
| 521 |
(e.g., TCP-AO, TLS, or IPsec) between two BGP routers might be |
| 522 |
compromised (e.g., by extraction from a router). This would |
| 523 |
enable an adversary to effect MITM attacks on the link(s) where |
| 524 |
the key is used. Use of specific security mechanisms to protect |
| 525 |
inter-router links between ASes is outside the scope of PATHSEC. |
| 526 |
|
| 527 |
Compromised Router Private Key: If PATHSEC mechanisms employ |
| 528 |
public key cryptography, e.g., to digitally sign data in an |
| 529 |
update, then a private key associated with a router or an AS might |
| 530 |
be compromised by an attack against the router. An adversary with |
| 531 |
access to this key would be able to generate updates that appear |
| 532 |
to have passed through the AS that this router represents. Such |
| 533 |
updates might be injected on a link between the compromised router |
| 534 |
and its neighbors if that link is accessible to the adversary. If |
| 535 |
the adversary controls another network, it could use this key to |
| 536 |
forge signatures that appear to come from the AS or router(s) in |
| 537 |
question, with some constraints. So, for example, an adversary |
| 538 |
that controls another AS could use a compromised router/AS key to |
| 539 |
issue PATHSEC-signed data that includes the targeted router/AS. |
| 540 |
(Neighbors of the adversary's AS ought not accept a route that |
| 541 |
purports to emanate directly from the targeted AS. So, an |
| 542 |
adversary could take a legitimate, protected route that passes |
| 543 |
through the compromised AS, add itself as the next hop, and then |
| 544 |
forward the resulting route to neighbors.) |
| 545 |
|
| 546 |
Withdrawal Suppression Attack: A PATHSEC-protected update may be |
| 547 |
signed and announced, and later withdrawn. An adversary |
| 548 |
controlling intermediate routers could fail to propagate the |
| 549 |
withdrawal. BGP is already vulnerable to behavior of this sort, |
| 550 |
so withdrawal suppression is not characterized as an attack under |
| 551 |
the assumptions upon which this mode is based (i.e., no oracle). |
| 552 |
|
| 553 |
|
| 554 |
|
| 555 |
|
| 556 |
<span class="grey">Kent & Chi Informational [Page 10]</span></pre> |
| 557 |
<hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-11" ></span> |
| 558 |
<span class="grey"><a href="./rfc7132">RFC 7132</a> Threat Model for BGP Path Security February 2014</span> |
| 559 |
|
| 560 |
|
| 561 |
<span class="h3"><a class="selflink" id="section-4.3" href="#section-4.3">4.3</a>. Attacks on Network Operator Management Computers (Non-CA</span> |
| 562 |
<span class="h3"> Computers)</span> |
| 563 |
|
| 564 |
An adversary may choose to attack computers used by a network |
| 565 |
operator to manage its network, especially its routers. Such attacks |
| 566 |
might be effected by an adversary who has compromised the security of |
| 567 |
these computers. This might be effected via remote attacks, |
| 568 |
extortion of network operations staff, etc. If an adversary |
| 569 |
compromises NOC computers, he can execute any management function |
| 570 |
that authorized network operations the staff would have performed. |
| 571 |
Thus, the adversary could modify the local routing policy to change |
| 572 |
preferences, to black-hole certain routes, etc. This type of |
| 573 |
behavior cannot be externally detected as an attack. Externally, |
| 574 |
this appears as a form of rogue operator behavior. (Such behavior |
| 575 |
might be perceived as accidental or malicious by other operators.) |
| 576 |
|
| 577 |
If a network operator participates in the RPKI, an adversary could |
| 578 |
manipulate the RP tools that extract data from the RPKI, causing the |
| 579 |
output of these tools to be corrupted in various ways. For example, |
| 580 |
an attack of this sort could cause the operator to view valid routes |
| 581 |
as not validated, which could alter its routing behavior. |
| 582 |
|
| 583 |
If an adversary invoked the tool used to manage the repository |
| 584 |
publication point for this operator, it could delete any objects |
| 585 |
stored there (certificates, CRLs, manifests, ROAs, or subordinate CA |
| 586 |
certificates). This could affect the routing status of entities that |
| 587 |
have allocations/assignments from this network operator (e.g., by |
| 588 |
deleting their CA certificates). |
| 589 |
|
| 590 |
An adversary could invoke the tool used to request certificate |
| 591 |
revocation, causing router certificates, ROAs, or subordinate CA |
| 592 |
certificates to be revoked. An attack of this sort could affect not |
| 593 |
only this operator but also any operators that receive allocations/ |
| 594 |
assignments from it, e.g., because their CA certificates were |
| 595 |
revoked. |
| 596 |
|
| 597 |
If an operator is PATHSEC-enabled, an attack of this sort could cause |
| 598 |
the affected operator to be viewed as not PATHSEC-enabled, possibly |
| 599 |
making routes it emits less preferable to other operators. |
| 600 |
|
| 601 |
If an adversary invoked a tool used to request ROAs, it could |
| 602 |
effectively reallocate some of the prefixes allocated/assigned to the |
| 603 |
network operator (e.g., by modifying the origin AS in ROAs). This |
| 604 |
might cause other PATHSEC-enabled networks to view the affected |
| 605 |
network as no longer originating routes for these prefixes. Multi- |
| 606 |
homed subscribers of this operator who received an allocation from |
| 607 |
the operator might find that their traffic was routed via other |
| 608 |
connections. |
| 609 |
|
| 610 |
|
| 611 |
|
| 612 |
<span class="grey">Kent & Chi Informational [Page 11]</span></pre> |
| 613 |
<hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-12" ></span> |
| 614 |
<span class="grey"><a href="./rfc7132">RFC 7132</a> Threat Model for BGP Path Security February 2014</span> |
| 615 |
|
| 616 |
|
| 617 |
If the network operator is PATHSEC-enabled, and makes use of |
| 618 |
certificates associated with routers/ASes, an adversary could invoke |
| 619 |
a tool used to request such certificates. The adversary could then |
| 620 |
replace valid certificates for routers/ASes with ones that might be |
| 621 |
rejected by PATHSEC-enabled neighbors. |
| 622 |
|
| 623 |
<span class="h3"><a class="selflink" id="section-4.4" href="#section-4.4">4.4</a>. Attacks on a Repository Publication Point</span> |
| 624 |
|
| 625 |
A critical element of the RPKI is the repository system. An |
| 626 |
adversary might attack a repository, or a publication point within a |
| 627 |
repository, to adversely affect routing. |
| 628 |
|
| 629 |
This section considers only those attacks that can be launched by any |
| 630 |
adversary who controls a computer hosting one or more repository |
| 631 |
publication points, without access to the cryptographic keys needed |
| 632 |
to generate valid RPKI-signed products. Such attacks might be |
| 633 |
effected by an insider or an external threat. Because all repository |
| 634 |
objects are digitally signed, attacks of this sort translate into DoS |
| 635 |
attacks against the RPKI RPs. There are a few distinct forms of such |
| 636 |
attacks, as described below. |
| 637 |
|
| 638 |
Note first that the RPKI calls for RPs to cache the data they acquire |
| 639 |
and verify from the repository system [<a href="./rfc6480" title=""An Infrastructure to Support Secure Internet Routing"">RFC6480</a>][RFC6481]. Attacks |
| 640 |
that delete signed products, insert products with "bad" signatures, |
| 641 |
tamper with object signatures, or replace newer objects with older |
| 642 |
(valid) ones, can be detected by RPs (with a few exceptions). RPs |
| 643 |
are expected to make use of local caches. If repository publication |
| 644 |
points are unavailable or the retrieved data is corrupted, an RP can |
| 645 |
revert to using the cached data. This behavior helps insulate RPs |
| 646 |
from the immediate effects of DoS attacks on publication points. |
| 647 |
|
| 648 |
Each RPKI data object has an associated date on which it expires or |
| 649 |
is considered stale (certificates expire and CRLs become stale). |
| 650 |
When an RP uses cached data, how to deal with stale or expired data |
| 651 |
is a local decision. It is common in PKIs to make use of stale |
| 652 |
certificate revocation status data when fresher data is not |
| 653 |
available. Use of expired certificates is less common, although not |
| 654 |
unknown. Each RP will decide, locally, whether to continue to make |
| 655 |
use of or ignore cached RPKI objects that are stale or expired. |
| 656 |
|
| 657 |
If an adversary inserts an object into a publication point, and the |
| 658 |
object has a "bad" signature, the object will not be accepted and |
| 659 |
used by RPs. |
| 660 |
|
| 661 |
If an adversary modifies any signed product at a publication point, |
| 662 |
the signature on the product will fail, causing RPs to not accept it. |
| 663 |
This is equivalent to deleting the object, in many respects. |
| 664 |
|
| 665 |
|
| 666 |
|
| 667 |
|
| 668 |
<span class="grey">Kent & Chi Informational [Page 12]</span></pre> |
| 669 |
<hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-13" ></span> |
| 670 |
<span class="grey"><a href="./rfc7132">RFC 7132</a> Threat Model for BGP Path Security February 2014</span> |
| 671 |
|
| 672 |
|
| 673 |
If an adversary deletes one or more CA certificates, ROAs, or the CRL |
| 674 |
for a publication point, the manifest for that publication point will |
| 675 |
allow an RP to detect this attack. An RP can continue to use the |
| 676 |
last valid instance of the deleted object (as a local policy option), |
| 677 |
thus minimizing the impact of such an attack. |
| 678 |
|
| 679 |
If an adversary deletes a manifest (and does not replace it with an |
| 680 |
older instance), RPs are able to detect this action. Such behavior |
| 681 |
should result in the CA (or publication point maintainer) being |
| 682 |
notified of the problem. An RP can continue to use the last valid |
| 683 |
instance of the deleted manifest (a local policy option), thus |
| 684 |
minimizing the impact of such an attack. |
| 685 |
|
| 686 |
If an adversary deletes newly added CA certificates or ROAs, and |
| 687 |
replaces the current manifest with the previous manifest, the |
| 688 |
manifest (and the CRL that it matches) will be "stale" (see |
| 689 |
[<a href="./rfc6486" title=""Manifests for the Resource Public Key Infrastructure (RPKI)"">RFC6486</a>]). This alerts an RP that there may be a problem. The RP |
| 690 |
should use the information from a Ghostbuster Record [<a href="./rfc6493" title=""The Resource Public Key Infrastructure (RPKI) Ghostbusters Record"">RFC6493</a>] to |
| 691 |
contact the entity responsible for the publication point and request |
| 692 |
a remedy to the problem (e.g., republish the missing CA certificates |
| 693 |
and/or ROAs). An RP cannot know the content of the new certificates |
| 694 |
or ROAs that are not present, but it can continue to use what it has |
| 695 |
cached. An attack of this sort will, at least temporarily, cause RPs |
| 696 |
to be unaware of the newly published objects. INRs associated with |
| 697 |
these objects will be treated as unauthenticated. |
| 698 |
|
| 699 |
If a CA revokes a CA certificate or a ROA (via deleting the |
| 700 |
corresponding End Entity (EE) certificate), and the adversary tries |
| 701 |
to reinstate that CA certificate or ROA, the adversary would have to |
| 702 |
rollback the CRL and the manifest to undo this action by the CA. As |
| 703 |
above, this would make the CRL and manifest stale, and this is |
| 704 |
detectable by RPs. An RP cannot know which CA certificates or ROAs |
| 705 |
were deleted. Depending on local policy, the RP might use the cached |
| 706 |
instances of the affected objects and thus be tricked into making |
| 707 |
decisions based on these revoked objects. Here too, the goal is that |
| 708 |
the CA will be notified of the problem (by RPs) and will remedy the |
| 709 |
error. |
| 710 |
|
| 711 |
In the attack scenarios above, when a CRL or manifest is described as |
| 712 |
stale, this means that the next issue date for the CRL or manifest |
| 713 |
has passed. Until the next issue date, an RP will not detect the |
| 714 |
attack. Thus, it behooves CAs to select CRL/manifest lifetimes (the |
| 715 |
two are linked) that represent an acceptable trade-off between risk |
| 716 |
and operational burdens. |
| 717 |
|
| 718 |
Attacks effected by adversaries that are legitimate managers of |
| 719 |
publication points can have much greater effects and are discussed |
| 720 |
below under attacks on or by CAs. |
| 721 |
|
| 722 |
|
| 723 |
|
| 724 |
<span class="grey">Kent & Chi Informational [Page 13]</span></pre> |
| 725 |
<hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-14" ></span> |
| 726 |
<span class="grey"><a href="./rfc7132">RFC 7132</a> Threat Model for BGP Path Security February 2014</span> |
| 727 |
|
| 728 |
|
| 729 |
<span class="h3"><a class="selflink" id="section-4.5" href="#section-4.5">4.5</a>. Attacks on an RPKI CA</span> |
| 730 |
|
| 731 |
Every entity to which INRs have been allocated/assigned is a CA in |
| 732 |
the RPKI. Each CA is nominally responsible for managing the |
| 733 |
repository publication point for the set of signed products that it |
| 734 |
generates. (An INR holder may choose to outsource the operation of |
| 735 |
the RPKI CA function and the associated publication point. In such |
| 736 |
cases, the organization operating on behalf of the INR holder becomes |
| 737 |
the CA from an operational and security perspective. The following |
| 738 |
discussion does not distinguish such outsourced CA operations.) |
| 739 |
|
| 740 |
Note that attacks attributable to a CA may be the result of malice by |
| 741 |
the CA (i.e., the CA is the adversary), or they may result from a |
| 742 |
compromise of the CA. |
| 743 |
|
| 744 |
All of the adversaries listed in <a href="#section-2">Section 2</a> are presumed to be capable |
| 745 |
of launching attacks against the computers used to perform CA |
| 746 |
functions. Some adversaries might effect an attack on a CA by |
| 747 |
violating personnel or physical security controls as well. The |
| 748 |
distinction between the CA as an adversary versus the CA as an attack |
| 749 |
victim is important. Only in the latter case should one expect the |
| 750 |
CA to remedy problems caused by an attack once the attack has been |
| 751 |
detected. (If a CA does not take such action, the effects are the |
| 752 |
same as if the CA is an adversary.) |
| 753 |
|
| 754 |
Note that most of the attacks described below do not require |
| 755 |
disclosure of a CA's private key to an adversary. If the adversary |
| 756 |
can gain control of the computer used to issue certificates, it can |
| 757 |
effect these attacks, even though the private key for the CA remains |
| 758 |
"secure" (i.e., not disclosed to unauthorized parties). However, if |
| 759 |
the CA is not the adversary, and if the CA's private key is not |
| 760 |
compromised, then recovery from these attacks is much easier. This |
| 761 |
motivates use of hardware security modules to protect CA keys, at |
| 762 |
least for higher tiers in the RPKI. |
| 763 |
|
| 764 |
An attack by a CA can result in revocation or replacement of any of |
| 765 |
the certificates that the CA has issued. Revocation of a certificate |
| 766 |
should cause RPs to delete the (formerly) valid certificate (and |
| 767 |
associated signed object, in the case of a revoked EE certificate) |
| 768 |
that they have cached. This would cause repository objects (e.g., CA |
| 769 |
certificates and ROAs) that are verified under that certificate to be |
| 770 |
considered invalid, transitively. As a result, RPs would not |
| 771 |
consider any ROAs or PATHSEC-protected updates to be valid based on |
| 772 |
these certificates, which would make routes dependent on them less |
| 773 |
preferred. Because a CA that revokes a certificate is authorized to |
| 774 |
do so, this sort of attack cannot be detected, intrinsically, by most |
| 775 |
RPs. However, the entities affected by the revocation or replacement |
| 776 |
of CA certificates can be expected to detect the attack and contact |
| 777 |
|
| 778 |
|
| 779 |
|
| 780 |
<span class="grey">Kent & Chi Informational [Page 14]</span></pre> |
| 781 |
<hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-15" ></span> |
| 782 |
<span class="grey"><a href="./rfc7132">RFC 7132</a> Threat Model for BGP Path Security February 2014</span> |
| 783 |
|
| 784 |
|
| 785 |
the CA to effect remediation. If the CA was not the adversary, it |
| 786 |
should be able to issue new certificates and restore the publication |
| 787 |
point. |
| 788 |
|
| 789 |
An adversary that controls the CA for a publication point can publish |
| 790 |
signed products that create more subtle types of DoS attacks against |
| 791 |
RPs. For example, such an attacker could create subordinate CA |
| 792 |
certificates with Subject Information Access (SIA) pointers that lead |
| 793 |
RPs on a "wild goose chase" looking for additional publication points |
| 794 |
and signed products. An attacker could publish certificates with |
| 795 |
very brief validity intervals or CRLs and manifests that become |
| 796 |
"stale" very quickly. This sort of attack would cause RPs to access |
| 797 |
repositories more frequently, and that might interfere with |
| 798 |
legitimate accesses by other RPs. |
| 799 |
|
| 800 |
An attacker with this capability could create very large numbers of |
| 801 |
ROAs to be processed (with prefixes that are consistent with the |
| 802 |
allocation for the CA) and correspondingly large manifests. An |
| 803 |
attacker could create very deep subtrees with many ROAs per |
| 804 |
publication point, etc. All of these types of DoS attacks against |
| 805 |
RPs are feasible within the syntactic and semantic constraints |
| 806 |
established for RPKI certificates, CRLs, and signed objects. |
| 807 |
|
| 808 |
An attack that results in revocation and replacement (e.g., key |
| 809 |
rollover or certificate renewal) of a CA certificate would cause RPs |
| 810 |
to replace the old, valid certificate with the new one. This new |
| 811 |
certificate might contain a public key that does not correspond to |
| 812 |
the private key held by the certificate subject. That would cause |
| 813 |
objects signed by that subject to be rejected as invalid, and prevent |
| 814 |
the affected subject from being able to sign new objects. As above, |
| 815 |
RPs would not consider any ROAs issued under the affected CA |
| 816 |
certificate to be valid, and updates based on router certificates |
| 817 |
issued by the affected CA would be rejected. This would make routes |
| 818 |
dependent on these signed products less preferred. However, the |
| 819 |
constraints imposed by the use of extensions detailed in [<a href="./rfc3779" title=""X.509 Extensions for IP Addresses and AS Identifiers"">RFC3779</a>] |
| 820 |
prevent a compromised CA from issuing (valid) certificates with INRs |
| 821 |
outside the scope of the CA, thus limiting the impact of the attack. |
| 822 |
|
| 823 |
An adversary that controls a CA could issue CA certificates with |
| 824 |
overlapping INRs to different entities when no transfer of INRs is |
| 825 |
intended. This could cause confusion for RPs as conflicting ROAs |
| 826 |
could be issued by the distinct (subordinate) CAs. |
| 827 |
|
| 828 |
An adversary could replace a CA certificate, use the corresponding |
| 829 |
private key to issue new signed products, and then publish them at a |
| 830 |
publication point controlled by the attacker. This would effectively |
| 831 |
transfer the affected INRs to the adversary or to a third party of |
| 832 |
his choosing. The result would be to cause RPs to view the entity |
| 833 |
|
| 834 |
|
| 835 |
|
| 836 |
<span class="grey">Kent & Chi Informational [Page 15]</span></pre> |
| 837 |
<hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-16" ></span> |
| 838 |
<span class="grey"><a href="./rfc7132">RFC 7132</a> Threat Model for BGP Path Security February 2014</span> |
| 839 |
|
| 840 |
|
| 841 |
that controls the private key in question as the legitimate INR |
| 842 |
holder. Again, the constraints imposed by the use of the extensions |
| 843 |
in <a href="./rfc3779">RFC 3779</a> prevent a compromised CA from issuing (valid) |
| 844 |
certificates with INRs outside the scope of the CA, thus limiting the |
| 845 |
impact of the attack. |
| 846 |
|
| 847 |
Finally, an entity that manages a repository publication point can |
| 848 |
inadvertently act as an attacker (an example of Walt Kelly's most |
| 849 |
famous "Pogo" quote [<a href="#ref-Kelly70" title=""We Have Met The Enemy and He Is Us: Pogo Earth Day Poster"">Kelly70</a>]). For example, a CA might fail to |
| 850 |
replace its own certificate in a timely fashion (well before it |
| 851 |
expires). It might fail to issue its CRL and manifest prior to |
| 852 |
expiration, creating stale instances of these products that cause |
| 853 |
concern for RPs. A CA with many subordinate CAs (e.g., an RIR or |
| 854 |
NIR) might fail to distribute the expiration times for the CA |
| 855 |
certificates that it issues. A network with many ROAs might do the |
| 856 |
same for the EE certificates associated with the ROAs it generates. |
| 857 |
A CA could rollover its key but fail to reissue subordinate CA |
| 858 |
certificates under its new key. Poor planning with regard to rekey |
| 859 |
intervals for managed CAs could impose undue burdens for RPs, despite |
| 860 |
a lack of malicious intent. All of these examples of mismanagement |
| 861 |
could adversely affect RPs, despite the absence of malicious intent. |
| 862 |
|
| 863 |
<span class="h2"><a class="selflink" id="section-5" href="#section-5">5</a>. Residual Vulnerabilities</span> |
| 864 |
|
| 865 |
The RPKI, upon which PATHSEC relies, has several residual |
| 866 |
vulnerabilities that were discussed in the preceding text (Sections |
| 867 |
4.4 and 4.5). These vulnerabilities are of two principle forms: |
| 868 |
|
| 869 |
o The RPKI repository system may be attacked in ways that make its |
| 870 |
contents unavailable, not current, or inconsistent. The principle |
| 871 |
defense against most forms of DoS attacks is the use of a local |
| 872 |
cache by each RP. The local cache ensures availability of |
| 873 |
previously acquired RPKI data in the event that a repository is |
| 874 |
inaccessible or if the repository contents are deleted |
| 875 |
(maliciously). Nonetheless, the system cannot ensure that every |
| 876 |
RP will always have access to up-to-date RPKI data. An RP, when |
| 877 |
it detects a problem with acquired repository data, has two |
| 878 |
options: |
| 879 |
|
| 880 |
1. The RP may choose to make use of its local cache, employing |
| 881 |
local configuration settings that tolerate expired or stale |
| 882 |
objects. (Such behavior is, nominally, always within the |
| 883 |
purview of an RP in PKI.) Using cached, expired, or stale |
| 884 |
data subjects the RP to attacks that take advantage of the |
| 885 |
RP's ignorance of changes to this data. |
| 886 |
|
| 887 |
|
| 888 |
|
| 889 |
|
| 890 |
|
| 891 |
|
| 892 |
<span class="grey">Kent & Chi Informational [Page 16]</span></pre> |
| 893 |
<hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-17" ></span> |
| 894 |
<span class="grey"><a href="./rfc7132">RFC 7132</a> Threat Model for BGP Path Security February 2014</span> |
| 895 |
|
| 896 |
|
| 897 |
2. The RP may chose to purge expired objects. Purging expired |
| 898 |
objects removes the security information associated with the |
| 899 |
real-world INRs to which the objects refer. This is |
| 900 |
equivalent to the affected INRs not having been afforded |
| 901 |
protection via the RPKI. Since use of the RPKI (and PATHSEC) |
| 902 |
is voluntary, there may always be a set of INRs that are not |
| 903 |
protected by these mechanisms. Thus, purging moves the |
| 904 |
affected INRs to the set of non-participating INR holders. |
| 905 |
This more conservative response enables an attacker to move |
| 906 |
INRs from the protected set to the unprotected set. |
| 907 |
|
| 908 |
o Any CA in the RPKI may misbehave within the bounds of the INRs |
| 909 |
allocated to it, e.g., it may issue certificates with duplicate |
| 910 |
resource allocations or revoke certificates inappropriately. This |
| 911 |
vulnerability is intrinsic in any PKI, but its impact is limited |
| 912 |
in the RPKI because of the use of extensions in <a href="./rfc3779">RFC 3779</a>. It is |
| 913 |
anticipated that RPs will deal with such misbehavior through |
| 914 |
administrative means once it is detected. |
| 915 |
|
| 916 |
PATHSEC has a separate set of residual vulnerabilities: |
| 917 |
|
| 918 |
o It has been stated that "route leaks" are viewed as a routing |
| 919 |
security problem by many operators. However, BGP itself does not |
| 920 |
include semantics that preclude what many perceive as route leaks, |
| 921 |
and there is no definition of the term in any RFC. This makes it |
| 922 |
inappropriate to address route leaks in this document. |
| 923 |
Additionally, route leaks are outside the scope of PATHSEC, |
| 924 |
consistent with the security context noted in <a href="#section-1">Section 1</a> of this |
| 925 |
document. If, at a later time, the SIDR security context is |
| 926 |
revised to include route leaks, and an appropriate definition |
| 927 |
exists, this document should be revised. |
| 928 |
|
| 929 |
o PATHSEC is not required to protect all attributes associated with |
| 930 |
an AS_PATH, even though some of these attributes may be employed |
| 931 |
as inputs to routing decisions. Thus, attacks that modify (or |
| 932 |
strip) these other attributes are not prevented/detected by |
| 933 |
PATHSEC. As noted in <a href="#section-1">Section 1</a>, the SIDR security context calls |
| 934 |
for protecting only the information needed to verify that a |
| 935 |
received route traversed the ASes in question, and that the NLRI |
| 936 |
in the route is what was advertised. (The AS_PATH data also may |
| 937 |
have traversed ASes within a confederation that are not |
| 938 |
represented. However, these ASes are not externally visible and |
| 939 |
thus do not influence route selection, so their omission in this |
| 940 |
context is not a security concern.) Thus, protection of other |
| 941 |
attributes is outside the scope of this document, as described in |
| 942 |
<a href="#section-1">Section 1</a>. If, at a later time, the SIDR security context is |
| 943 |
revised to include protection of additional BGP attributes, this |
| 944 |
document should be revised. |
| 945 |
|
| 946 |
|
| 947 |
|
| 948 |
<span class="grey">Kent & Chi Informational [Page 17]</span></pre> |
| 949 |
<hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-18" ></span> |
| 950 |
<span class="grey"><a href="./rfc7132">RFC 7132</a> Threat Model for BGP Path Security February 2014</span> |
| 951 |
|
| 952 |
|
| 953 |
o PATHSEC cannot ensure that an AS will withdraw a route when the AS |
| 954 |
no longer has a route for a prefix, as noted in <a href="#section-4.2">Section 4.2</a>. |
| 955 |
PATHSEC may incorporate features to limit the lifetime of an |
| 956 |
advertisement. Such lifetime limits provide an upper bound on the |
| 957 |
time that the failure to withdraw a route will remain effective. |
| 958 |
|
| 959 |
<span class="h2"><a class="selflink" id="section-6" href="#section-6">6</a>. Security Considerations</span> |
| 960 |
|
| 961 |
A threat model is, by definition, a security-centric document. |
| 962 |
Unlike a protocol description, a threat model does not create |
| 963 |
security problems nor does it purport to address security problems. |
| 964 |
This model postulates a set of threats (i.e., motivated, capable |
| 965 |
adversaries) and examines classes of attacks that these threats are |
| 966 |
capable of effecting, based on the motivations ascribed to the |
| 967 |
threats. It describes the impact of these types of attacks on |
| 968 |
PATHSEC, including the RPKI on which PATHSEC relies. It describes |
| 969 |
how the design of the RPKI (and the PATHSEC design goals) address |
| 970 |
classes of attacks, where applicable. It also notes residual |
| 971 |
vulnerabilities. |
| 972 |
|
| 973 |
<span class="h2"><a class="selflink" id="section-7" href="#section-7">7</a>. Acknowledgements</span> |
| 974 |
|
| 975 |
The authors with to thank the members of the SIDR working group for |
| 976 |
the extensive feedback provided during the development of this |
| 977 |
document. |
| 978 |
|
| 979 |
<span class="h2"><a class="selflink" id="section-8" href="#section-8">8</a>. Informative References</span> |
| 980 |
|
| 981 |
[<a id="ref-Kelly70">Kelly70</a>] Kelly, W., "We Have Met The Enemy and He Is Us: Pogo Earth |
| 982 |
Day Poster", April 1970. |
| 983 |
|
| 984 |
[<a id="ref-Kent2000">Kent2000</a>] |
| 985 |
Kent, S., Lynn, C., and K. Seo, "Design and Analysis of |
| 986 |
the Secure Border Gateway Protocol (S-BGP)", IEEE DISCEX |
| 987 |
Conference, June 2000. |
| 988 |
|
| 989 |
[<a id="ref-RFC3779">RFC3779</a>] Lynn, C., Kent, S., and K. Seo, "X.509 Extensions for IP |
| 990 |
Addresses and AS Identifiers", <a href="./rfc3779">RFC 3779</a>, June 2004. |
| 991 |
|
| 992 |
[<a id="ref-RFC4271">RFC4271</a>] Rekhter, Y., Li, T., and S. Hares, "A Border Gateway |
| 993 |
Protocol 4 (BGP-4)", <a href="./rfc4271">RFC 4271</a>, January 2006. |
| 994 |
|
| 995 |
[<a id="ref-RFC4272">RFC4272</a>] Murphy, S., "BGP Security Vulnerabilities Analysis", <a href="./rfc4272">RFC</a> |
| 996 |
<a href="./rfc4272">4272</a>, January 2006. |
| 997 |
|
| 998 |
[<a id="ref-RFC4301">RFC4301</a>] Kent, S. and K. Seo, "Security Architecture for the |
| 999 |
Internet Protocol", <a href="./rfc4301">RFC 4301</a>, December 2005. |
| 1000 |
|
| 1001 |
|
| 1002 |
|
| 1003 |
|
| 1004 |
<span class="grey">Kent & Chi Informational [Page 18]</span></pre> |
| 1005 |
<hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-19" ></span> |
| 1006 |
<span class="grey"><a href="./rfc7132">RFC 7132</a> Threat Model for BGP Path Security February 2014</span> |
| 1007 |
|
| 1008 |
|
| 1009 |
[<a id="ref-RFC5925">RFC5925</a>] Touch, J., Mankin, A., and R. Bonica, "The TCP |
| 1010 |
Authentication Option", <a href="./rfc5925">RFC 5925</a>, June 2010. |
| 1011 |
|
| 1012 |
[<a id="ref-RFC6480">RFC6480</a>] Lepinski, M. and S. Kent, "An Infrastructure to Support |
| 1013 |
Secure Internet Routing", <a href="./rfc6480">RFC 6480</a>, February 2012. |
| 1014 |
|
| 1015 |
[<a id="ref-RFC6481">RFC6481</a>] Huston, G., Loomans, R., and G. Michaelson, "A Profile for |
| 1016 |
Resource Certificate Repository Structure", <a href="./rfc6481">RFC 6481</a>, |
| 1017 |
February 2012. |
| 1018 |
|
| 1019 |
[<a id="ref-RFC6482">RFC6482</a>] Lepinski, M., Kent, S., and D. Kong, "A Profile for Route |
| 1020 |
Origin Authorizations (ROAs)", <a href="./rfc6482">RFC 6482</a>, February 2012. |
| 1021 |
|
| 1022 |
[<a id="ref-RFC6486">RFC6486</a>] Austein, R., Huston, G., Kent, S., and M. Lepinski, |
| 1023 |
"Manifests for the Resource Public Key Infrastructure |
| 1024 |
(RPKI)", <a href="./rfc6486">RFC 6486</a>, February 2012. |
| 1025 |
|
| 1026 |
[<a id="ref-RFC6487">RFC6487</a>] Huston, G., Michaelson, G., and R. Loomans, "A Profile for |
| 1027 |
X.509 PKIX Resource Certificates", <a href="./rfc6487">RFC 6487</a>, February |
| 1028 |
2012. |
| 1029 |
|
| 1030 |
[<a id="ref-RFC6488">RFC6488</a>] Lepinski, M., Chi, A., and S. Kent, "Signed Object |
| 1031 |
Template for the Resource Public Key Infrastructure |
| 1032 |
(RPKI)", <a href="./rfc6488">RFC 6488</a>, February 2012. |
| 1033 |
|
| 1034 |
[<a id="ref-RFC6493">RFC6493</a>] Bush, R., "The Resource Public Key Infrastructure (RPKI) |
| 1035 |
Ghostbusters Record", <a href="./rfc6493">RFC 6493</a>, February 2012. |
| 1036 |
|
| 1037 |
[<a id="ref-RFC6810">RFC6810</a>] Bush, R. and R. Austein, "The Resource Public Key |
| 1038 |
Infrastructure (RPKI) to Router Protocol", <a href="./rfc6810">RFC 6810</a>, |
| 1039 |
January 2013. |
| 1040 |
|
| 1041 |
[<a id="ref-SIDR-CH">SIDR-CH</a>] "Secure Inter-Domain Routing: Charter for Working Group", |
| 1042 |
September 2013, <<a href="http://tools.ietf.org/wg/sidr/charters?item=charter-sidr-2013-09-20.txt">http://tools.ietf.org/wg/sidr/</a> |
| 1043 |
<a href="http://tools.ietf.org/wg/sidr/charters?item=charter-sidr-2013-09-20.txt">charters?item=charter-sidr-2013-09-20.txt</a>>. |
| 1044 |
|
| 1045 |
[<a id="ref-Sam04">Sam04</a>] Samuel, A., "Hacktivism and the Future of Political |
| 1046 |
Participation", Ph.D. dissertation, Harvard University, |
| 1047 |
September 2004, <<a href="http://www.alexandrasamuel.com/dissertation/pdfs/Samuel-Hacktivism-entire.pdf">http://www.alexandrasamuel.com/</a> |
| 1048 |
<a href="http://www.alexandrasamuel.com/dissertation/pdfs/Samuel-Hacktivism-entire.pdf">dissertation/pdfs/Samuel-Hacktivism-entire.pdf</a>>. |
| 1049 |
|
| 1050 |
|
| 1051 |
|
| 1052 |
|
| 1053 |
|
| 1054 |
|
| 1055 |
|
| 1056 |
|
| 1057 |
|
| 1058 |
|
| 1059 |
|
| 1060 |
<span class="grey">Kent & Chi Informational [Page 19]</span></pre> |
| 1061 |
<hr class='noprint'/><!--NewPage--><pre class='newpage'><span id="page-20" ></span> |
| 1062 |
<span class="grey"><a href="./rfc7132">RFC 7132</a> Threat Model for BGP Path Security February 2014</span> |
| 1063 |
|
| 1064 |
|
| 1065 |
Authors' Addresses |
| 1066 |
|
| 1067 |
Stephen Kent |
| 1068 |
BBN Technologies |
| 1069 |
10 Moulton St. |
| 1070 |
Cambridge, MA 02138 |
| 1071 |
USA |
| 1072 |
|
| 1073 |
EMail: [email protected] |
| 1074 |
|
| 1075 |
|
| 1076 |
Andrew Chi |
| 1077 |
University of North Carolina - Chapel Hill |
| 1078 |
c/o Department of Computer Science |
| 1079 |
CB 3175, Sitterson Hall |
| 1080 |
Chapel Hill, NC 27599 |
| 1081 |
USA |
| 1082 |
|
| 1083 |
EMail: [email protected] |
| 1084 |
|
| 1085 |
|
| 1086 |
|
| 1087 |
|
| 1088 |
|
| 1089 |
|
| 1090 |
|
| 1091 |
|
| 1092 |
|
| 1093 |
|
| 1094 |
|
| 1095 |
|
| 1096 |
|
| 1097 |
|
| 1098 |
|
| 1099 |
|
| 1100 |
|
| 1101 |
|
| 1102 |
|
| 1103 |
|
| 1104 |
|
| 1105 |
|
| 1106 |
|
| 1107 |
|
| 1108 |
|
| 1109 |
|
| 1110 |
|
| 1111 |
|
| 1112 |
|
| 1113 |
|
| 1114 |
|
| 1115 |
|
| 1116 |
Kent & Chi Informational [Page 20] |
| 1117 |
</pre> |