idnits 2.17.1 draft-ietf-roll-security-threats-09.txt: Checking boilerplate required by RFC 5378 and the IETF Trust (see https://trustee.ietf.org/license-info): ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/1id-guidelines.txt: ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/checklist : ---------------------------------------------------------------------------- No issues found here. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year == The document seems to lack the recommended RFC 2119 boilerplate, even if it appears to use RFC 2119 keywords. (The document does seem to have the reference to RFC 2119 which the ID-Checklist requires). -- The document date (August 13, 2014) is 3536 days in the past. Is this intentional? Checking references for intended status: Informational ---------------------------------------------------------------------------- == Outdated reference: A later version (-03) exists of draft-gilger-smart-object-security-workshop-02 == Outdated reference: A later version (-06) exists of draft-kelsey-intarea-mesh-link-establishment-05 -- Obsolete informational reference (is this intentional?): RFC 5751 (Obsoleted by RFC 8551) -- Obsolete informational reference (is this intentional?): RFC 6824 (Obsoleted by RFC 8684) Summary: 0 errors (**), 0 flaws (~~), 4 warnings (==), 3 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Routing Over Low-Power and Lossy Networks T. Tsao 3 Internet-Draft R. Alexander 4 Intended status: Informational Cooper Power Systems 5 Expires: February 14, 2015 M. Dohler 6 CTTC 7 V. Daza 8 A. Lozano 9 Universitat Pompeu Fabra 10 M. Richardson, Ed. 11 Sandelman Software Works 12 August 13, 2014 14 A Security Threat Analysis for Routing Protocol for Low-power and lossy 15 networks (RPL) 16 draft-ietf-roll-security-threats-09 18 Abstract 20 This document presents a security threat analysis for the Routing 21 Protocol for Low-power and lossy networks (RPL, ROLL). The 22 development builds upon previous work on routing security and adapts 23 the assessments to the issues and constraints specific to low-power 24 and lossy networks. A systematic approach is used in defining and 25 evaluating the security threats. Applicable countermeasures are 26 application specific and are addressed in relevant applicability 27 statements. 29 Status of This Memo 31 This Internet-Draft is submitted in full conformance with the 32 provisions of BCP 78 and BCP 79. 34 Internet-Drafts are working documents of the Internet Engineering 35 Task Force (IETF). Note that other groups may also distribute 36 working documents as Internet-Drafts. The list of current Internet- 37 Drafts is at http://datatracker.ietf.org/drafts/current/. 39 Internet-Drafts are draft documents valid for a maximum of six months 40 and may be updated, replaced, or obsoleted by other documents at any 41 time. It is inappropriate to use Internet-Drafts as reference 42 material or to cite them other than as "work in progress." 44 This Internet-Draft will expire on February 14, 2015. 46 Copyright Notice 47 Copyright (c) 2014 IETF Trust and the persons identified as the 48 document authors. All rights reserved. 50 This document is subject to BCP 78 and the IETF Trust's Legal 51 Provisions Relating to IETF Documents 52 (http://trustee.ietf.org/license-info) in effect on the date of 53 publication of this document. Please review these documents 54 carefully, as they describe your rights and restrictions with respect 55 to this document. Code Components extracted from this document must 56 include Simplified BSD License text as described in Section 4.e of 57 the Trust Legal Provisions and are provided without warranty as 58 described in the Simplified BSD License. 60 Table of Contents 62 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 3 63 2. Relationship to other documents . . . . . . . . . . . . . . . 4 64 3. Terminology . . . . . . . . . . . . . . . . . . . . . . . . . 4 65 4. Considerations on RPL Security . . . . . . . . . . . . . . . 5 66 4.1. Routing Assets and Points of Access . . . . . . . . . . . 6 67 4.2. The ISO 7498-2 Security Reference Model . . . . . . . . . 8 68 4.3. Issues Specific to or Amplified in LLNs . . . . . . . . . 9 69 4.4. RPL Security Objectives . . . . . . . . . . . . . . . . . 12 70 5. Threat Sources . . . . . . . . . . . . . . . . . . . . . . . 13 71 6. Threats and Attacks . . . . . . . . . . . . . . . . . . . . . 13 72 6.1. Threats due to failures to Authenticate . . . . . . . . . 14 73 6.1.1. Node Impersonation . . . . . . . . . . . . . . . . . 14 74 6.1.2. Dummy Node . . . . . . . . . . . . . . . . . . . . . 14 75 6.1.3. Node Resource Spam . . . . . . . . . . . . . . . . . 14 76 6.2. Threats and Attacks on Confidentiality . . . . . . . . . 15 77 6.2.1. Routing Exchange Exposure . . . . . . . . . . . . . . 15 78 6.2.2. Routing Information (Routes and Network Topology) 79 Exposure . . . . . . . . . . . . . . . . . . . . . . 15 80 6.3. Threats and Attacks on Integrity . . . . . . . . . . . . 16 81 6.3.1. Routing Information Manipulation . . . . . . . . . . 16 82 6.3.2. Node Identity Misappropriation . . . . . . . . . . . 17 83 6.4. Threats and Attacks on Availability . . . . . . . . . . . 17 84 6.4.1. Routing Exchange Interference or Disruption . . . . . 17 85 6.4.2. Network Traffic Forwarding Disruption . . . . . . . . 17 86 6.4.3. Communications Resource Disruption . . . . . . . . . 19 87 6.4.4. Node Resource Exhaustion . . . . . . . . . . . . . . 19 88 7. Countermeasures . . . . . . . . . . . . . . . . . . . . . . . 20 89 7.1. Confidentiality Attack Countermeasures . . . . . . . . . 20 90 7.1.1. Countering Deliberate Exposure Attacks . . . . . . . 20 91 7.1.2. Countering Passive Wiretapping Attacks . . . . . . . 21 92 7.1.3. Countering Traffic Analysis . . . . . . . . . . . . . 22 93 7.1.4. Countering Remote Device Access Attacks . . . . . . . 22 94 7.2. Integrity Attack Countermeasures . . . . . . . . . . . . 23 95 7.2.1. Countering Unauthorized Modification Attacks . . . . 23 96 7.2.2. Countering Overclaiming and Misclaiming Attacks . . . 24 97 7.2.3. Countering Identity (including Sybil) Attacks . . . . 24 98 7.2.4. Countering Routing Information Replay Attacks . . . . 24 99 7.2.5. Countering Byzantine Routing Information Attacks . . 25 100 7.3. Availability Attack Countermeasures . . . . . . . . . . . 26 101 7.3.1. Countering HELLO Flood Attacks and ACK Spoofing 102 Attacks . . . . . . . . . . . . . . . . . . . . . . . 26 103 7.3.2. Countering Overload Attacks . . . . . . . . . . . . . 27 104 7.3.3. Countering Selective Forwarding Attacks . . . . . . . 28 105 7.3.4. Countering Sinkhole Attacks . . . . . . . . . . . . . 29 106 7.3.5. Countering Wormhole Attacks . . . . . . . . . . . . . 30 107 8. RPL Security Features . . . . . . . . . . . . . . . . . . . . 30 108 8.1. Confidentiality Features . . . . . . . . . . . . . . . . 31 109 8.2. Integrity Features . . . . . . . . . . . . . . . . . . . 32 110 8.3. Availability Features . . . . . . . . . . . . . . . . . . 32 111 8.4. Key Management . . . . . . . . . . . . . . . . . . . . . 33 112 9. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 33 113 10. Security Considerations . . . . . . . . . . . . . . . . . . . 33 114 11. Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . 33 115 12. References . . . . . . . . . . . . . . . . . . . . . . . . . 34 116 12.1. Normative References . . . . . . . . . . . . . . . . . . 34 117 12.2. Informative References . . . . . . . . . . . . . . . . . 34 118 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 37 120 1. Introduction 122 In recent times, networked electronic devices have found an 123 increasing number of applications in various fields. Yet, for 124 reasons ranging from operational application to economics, these 125 wired and wireless devices are often supplied with minimum physical 126 resources; the constraints include those on computational resources 127 (RAM, clock speed, storage), communication resources (duty cycle, 128 packet size, etc.), but also form factors that may rule out user 129 access interfaces (e.g., the housing of a small stick-on switch), or 130 simply safety considerations (e.g., with gas meters). As a 131 consequence, the resulting networks are more prone to loss of traffic 132 and other vulnerabilities. The proliferation of these low-power and 133 lossy networks (LLNs), however, are drawing efforts to examine and 134 address their potential networking challenges. Securing the 135 establishment and maintenance of network connectivity among these 136 deployed devices becomes one of these key challenges. 138 This document presents a threat analysis for securing the Routing 139 Protocol for LLNs (RPL). The process requires two steps. First, the 140 analysis will be used to identify pertinent security issues. The 141 second step is to identify necessary countermeasures to secure RPL. 142 As there are multiple ways to solve the problem and the specific 143 tradeoffs are deployment specific, the specific countermeasure to be 144 used is detailed in applicability statements. 146 This document uses [ISO.7498-2.1988]] model, which describes 147 Authentication, Access Control, Data Confidentiality, Data Integrity, 148 and Non-Repudiation security services and to which Availability is 149 added. 151 Many of the issues in this document were also covered in The IAB 152 Smart Object Workshop [RFC6574], and The IAB Smart Object Security 153 Workshop [I-D.gilger-smart-object-security-workshop]. 155 All of this document concerns itself with securing the control plane 156 traffic. As such it does not address authorization or authentication 157 of application traffic. RPL uses multicast as part of its protocol, 158 and therefore mechanisms which RPL uses to secure this traffic might 159 also be applicable to MPL control traffic as well: the important part 160 is that the threats are similar. 162 2. Relationship to other documents 164 ROLL has specified a set of routing protocols for Lossy and Low- 165 resource Networks (LLN) [RFC6550]. A number of applicability texts 166 describes a subset of these protocols and the conditions which make 167 the subset the correct choice. The text recommends and motivates the 168 accompanying parameter value ranges. Multiple applicability domains 169 are recognized including: Building and Home, and Advanced Metering 170 Infrastructure. The applicability domains distinguish themselves in 171 the way they are operated, their performance requirements, and the 172 most probable network structures. Each applicability statement 173 identifies the distinguishing properties according to a common set of 174 subjects described in as many sections. 176 The common set of security threats herein are referred to by the 177 applicability statements, and that series of documents describes the 178 preferred security settings and solutions within the applicability 179 statement conditions. This applicability statements may recommend 180 more light weight security solutions and specify the conditions under 181 which these solutions are appropriate. 183 3. Terminology 185 This document adopts the terminology defined in [RFC6550], in 186 [RFC4949], and in [RFC7102]. 188 The terms control plane and forwarding plane are used consistently 189 with section 1 of [RFC6192]. 191 The term DODAG is from [RFC6550]. 193 EAP-TLS is defined in [RFC5216]. 195 PANA is defined in [RFC5191]. 197 CCM mode is defined in [RFC3610]. 199 The terms SSID, ESSID and PAN refer to network identifiers, defined 200 in [IEEE.802.11] and [IEEE.802.15.4]. 202 Although this is not a protocol specification, the key words "MUST", 203 "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", 204 "RECOMMENDED", "NOT RECOMMENDED", "MAY", and "OPTIONAL" in this 205 document are to be interpreted as described in [RFC2119] in order to 206 clarify and emphasize the guidance and directions to implementers and 207 deployers of LLN nodes that utilize RPL. 209 4. Considerations on RPL Security 211 Routing security, in essence, ensures that the routing protocol 212 operates correctly. It entails implementing measures to ensure 213 controlled state changes on devices and network elements, both based 214 on external inputs (received via communications) or internal inputs 215 (physical security of device itself and parameters maintained by the 216 device, including, e.g., clock). State changes would thereby involve 217 not only authorization of injector's actions, authentication of 218 injectors, and potentially confidentiality of routing data, but also 219 proper order of state changes through timeliness, since seriously 220 delayed state changes, such as commands or updates of routing tables, 221 may negatively impact system operation. A security assessment can 222 therefore begin with a focus on the assets [RFC4949] that may be the 223 target of the state changes and the access points in terms of 224 interfaces and protocol exchanges through which such changes may 225 occur. In the case of routing security, the focus is directed 226 towards the elements associated with the establishment and 227 maintenance of network connectivity. 229 This section sets the stage for the development of the analysis by 230 applying the systematic approach proposed in [Myagmar2005] to the 231 routing security, while also drawing references from other reviews 232 and assessments found in the literature, particularly, [RFC4593] and 233 [Karlof2003]. The subsequent subsections begin with a focus on the 234 elements of a generic routing process that is used to establish 235 routing assets and points of access to the routing functionality. 236 Next, the [ISO.7498-2.1988] security model is briefly described. 237 Then, consideration is given to issues specific to or amplified in 238 LLNs. This section concludes with the formulation of a set of 239 security objectives for RPL. 241 4.1. Routing Assets and Points of Access 243 An asset is an important system resource (including information, 244 process, or physical resource), the access to, corruption or loss of 245 which adversely affects the system. In the control plane context, an 246 asset is information about the network, processes used to manage and 247 manipulate this data, and the physical devices on which this data is 248 stored and manipulated. The corruption or loss of these assets may 249 adversely impact the control plane of the network. Within the same 250 context, a point of access is an interface or protocol that 251 facilitates interaction between control plane assets. Identifying 252 these assets and points of access will provide a basis for 253 enumerating the attack surface of the control plane. 255 A level-0 data flow diagram [Yourdon1979] is used here to identify 256 the assets and points of access within a generic routing process. 257 The use of a data flow diagram allows for a clear and concise model 258 of the way in which routing nodes interact and process information, 259 and hence provides a context for threats and attacks. The goal of 260 the model is to be as detailed as possible so that corresponding 261 assets, points of access, and process in an individual routing 262 protocol can be readily identified. 264 Figure 1 shows that nodes participating in the routing process 265 transmit messages to discover neighbors and to exchange routing 266 information; routes are then generated and stored, which may be 267 maintained in the form of the protocol forwarding table. The nodes 268 use the derived routes for making forwarding decisions. 270 ................................................... 271 : : 272 : : 273 |Node_i|<------->(Routing Neighbor _________________ : 274 : Discovery)------------>Neighbor Topology : 275 : -------+--------- : 277 : | : 278 |Node_j|<------->(Route/Topology +--------+ : 279 : Exchange) | : 280 : | V ______ : 281 : +---->(Route Generation)--->Routes : 282 : ---+-- : 283 : | : 284 : Routing on a Node Node_k | : 285 ................................................... 286 | 287 |Forwarding | 288 |On Node_l|<-------------------------------------------+ 290 Notation: 292 (Proc) A process Proc 294 ________ 295 topology A structure storing neighbor adjacency (parent/child) 296 -------- 297 ________ 298 routes A structure storing the forwarding information base (FIB) 299 -------- 301 |Node_n| An external entity Node_n 303 -------> Data flow 305 Figure 1: Data Flow Diagram of a Generic Routing Process 307 It is seen from Figure 1 that 309 o Assets include 311 * routing and/or topology information; 313 * route generation process; 315 * communication channel resources (bandwidth); 317 * node resources (computing capacity, memory, and remaining 318 energy); 320 * node identifiers (including node identity and ascribed 321 attributes such as relative or absolute node location). 323 o Points of access include 325 * neighbor discovery; 327 * route/topology exchange; 329 * node physical interfaces (including access to data storage). 331 A focus on the above list of assets and points of access enables a 332 more directed assessment of routing security; for example, it is 333 readily understood that some routing attacks are in the form of 334 attempts to misrepresent routing topology. Indeed, the intention of 335 the security threat analysis is to be comprehensive. Hence, some of 336 the discussion which follows is associated with assets and points of 337 access that are not directly related to routing protocol design but 338 nonetheless provided for reference since they do have direct 339 consequences on the security of routing. 341 4.2. The ISO 7498-2 Security Reference Model 343 At the conceptual level, security within an information system in 344 general and applied to RPL in particular is concerned with the 345 primary issues of authentication, access control, data 346 confidentiality, data integrity, and non-repudiation. In the context 347 of RPL: 349 Authentication 350 Authentication involves the mutual authentication of the 351 routing peers prior to exchanging route information (i.e., peer 352 authentication) as well as ensuring that the source of the 353 route data is from the peer (i.e., data origin authentication). 354 [RFC5548] points out that LLNs can be drained by 355 unauthenticated peers before configuration. [RFC5673] requires 356 availability of open and untrusted side channels for new 357 joiners, and it requires strong and automated authentication so 358 that networks can automatically accept or reject new joiners. 360 Access Control 361 Access Control provides protection against unauthorized use of 362 the asset, and deals with the authorization of a node. 364 Confidentiality 365 Confidentiality involves the protection of routing information 366 as well as routing neighbor maintenance exchanges so that only 367 authorized and intended network entities may view or access it. 368 Because LLNs are most commonly found on a publicly accessible 369 shared medium, e.g., air or wiring in a building, and sometimes 370 formed ad hoc, confidentiality also extends to the neighbor 371 state and database information within the routing device since 372 the deployment of the network creates the potential for 373 unauthorized access to the physical devices themselves. 375 Integrity 376 Integrity entails the protection of routing information and 377 routing neighbor maintenance exchanges, as well as derived 378 information maintained in the database, from unauthorized 379 modification, insertions, deletions or replays. to be addressed 380 beyond the routing protocol. 382 Non-repudiation 383 Non-repudiation is the assurance that the transmission and/or 384 reception of a message cannot later be denied. The service of 385 non-repudiation applies after-the-fact and thus relies on the 386 logging or other capture of on-going message exchanges and 387 signatures. Applied to routing, non-repudiation is not an 388 issue because it does not apply to routing protocols, which are 389 machine-to-machine protocols. Further, with the LLN 390 application domains as described in [RFC5867] and [RFC5548], 391 proactive measures are much more critical than retrospective 392 protections. Finally, given the significant practical limits 393 to on-going routing transaction logging and storage and 394 individual device digital signature verification for each 395 exchange, non-repudiation in the context of routing is an 396 unsupportable burden that bears no further considered as an RPL 397 security issue. 399 It is recognized that, besides those security issues captured in the 400 ISO 7498-2 model, availability, is a security requirement: 402 Availability 403 Availability ensures that routing information exchanges and 404 forwarding services need to be available when they are required 405 for the functioning of the serving network. Availability will 406 apply to maintaining efficient and correct operation of routing 407 and neighbor discovery exchanges (including needed information) 408 and forwarding services so as not to impair or limit the 409 network's central traffic flow function 411 It should be emphasized here that for RPL security the above 412 requirements must be complemented by the proper security policies and 413 enforcement mechanisms to ensure that security objectives are met by 414 a given RPL implementation. 416 4.3. Issues Specific to or Amplified in LLNs 417 The requirements work detailed in Urban Requirements ([RFC5548]), 418 Industrial Requirements ([RFC5673]), Home Automation ([RFC5826], and 419 Building Automation ([RFC5867]) have identified specific issues and 420 constraints of routing in LLNs. The following is a list of 421 observations from those requirements and evaluation of their impact 422 on routing security considerations. 424 Limited energy, memory, and processing node resources 425 As a consequence of these constraints, there is an even more 426 critical need than usual for a careful study of trade-offs on 427 which and what level of security services are to be afforded 428 during the system design process. The chosen security 429 mechanisms also needs to work within these constraints. 430 Synchronization of security states with sleepy nodes is yet 431 another issue. A non-rechargeable battery powered node may 432 well be limited in energy for it's lifetime: once exchausted, 433 it may well never function again. 435 Large scale of rolled out network 436 The possibly numerous nodes to be deployed make manual on-site 437 configuration unlikely. For example, an urban deployment can 438 see several hundreds of thousands of nodes being installed by 439 many installers with a low level of expertise. Nodes may be 440 installed and not activated for many years, and additional 441 nodes may be added later on, which may be from old inventory. 442 The lifetime of the network is measured in decades, and this 443 complicates the operation of key management. 445 Autonomous operations 446 Self-forming and self-organizing are commonly prescribed 447 requirements of LLNs. In other words, a routing protocol 448 designed for LLNs needs to contain elements of ad hoc 449 networking and in most cases cannot rely on manual 450 configuration for initialization or local filtering rules. 451 Network topology/ownership changes, partitioning or merging, as 452 well as node replacement, can all contribute to complicating 453 the operations of key management. 455 Highly directional traffic 456 Some types of LLNs see a high percentage of their total traffic 457 traverse between the nodes and the LLN Border Routers (LBRs) 458 where the LLNs connect to non-LLNs. The special routing status 459 of and the greater volume of traffic near the LBRs have routing 460 security consequences as a higher valued attack target. In 461 fact, when Point-to-MultiPoint (P2MP) and MultiPoint-to-Point 462 (MP2P) traffic represents a majority of the traffic, routing 463 attacks consisting of advertising incorrect preferred routes 464 can cause serious damage. 466 While it might seem that nodes higher up in the cyclic graph 467 (i.e. those with lower rank) should be secured in a stronger 468 fashion, it is not in general easy to predict which nodes will 469 occupy those positions until after deployment. Issues of 470 redundancy and inventory control suggests that any node might 471 wind up in such a sensitive attack position, so all nodes need 472 to be equally secure. 474 In addition, even if it were possible to predict which nodes 475 will occupy positions of lower rank and provision them with 476 stronger security mechanisms, in the absense of a strong 477 authorization model, any node could advertise an incorrect 478 preferred route. 480 Unattended locations and limited physical security 481 Many applications have the nodes deployed in unattended or 482 remote locations; furthermore, the nodes themselves are often 483 built with minimal physical protection. These constraints 484 lower the barrier of accessing the data or security material 485 stored on the nodes through physical means. 487 Support for mobility 488 On the one hand, only a limited number of applications require 489 the support of mobile nodes, e.g., a home LLN that includes 490 nodes on wearable health care devices or an industry LLN that 491 includes nodes on cranes and vehicles. On the other hand, if a 492 routing protocol is indeed used in such applications, it will 493 clearly need to have corresponding security mechanisms. 495 Additionally nodes may appear to move from one side of a wall 496 to another without any actual motion involved, the result of 497 changes to electromagnetic properties, such as opening and 498 closing of a metal door. 500 Support for multicast and anycast 501 Support for multicast and anycast is called out chiefly for 502 large-scale networks. Since application of these routing 503 mechanisms in autonomous operations of many nodes is new, the 504 consequence on security requires careful consideration. 506 The above list considers how an LLN's physical constraints, size, 507 operations, and variety of application areas may impact security. 508 However, it is the combinations of these factors that particularly 509 stress the security concerns. For instance, securing routing for a 510 large number of autonomous devices that are left in unattended 511 locations with limited physical security presents challenges that are 512 not found in the common circumstance of administered networked 513 routers. The following subsection sets up the security objectives 514 for the routing protocol designed by the ROLL WG. 516 4.4. RPL Security Objectives 518 This subsection applies the ISO 7498-2 model to routing assets and 519 access points, taking into account the LLN issues, to develop a set 520 of RPL security objectives. 522 Since the fundamental function of a routing protocol is to build 523 routes for forwarding packets, it is essential to ensure that: 525 o routing/topology information integrity remains intact during 526 transfer and in storage; 528 o routing/topology information is used by authorized entities; 530 o routing/topology information is available when needed. 532 In conjunction, it is necessary to be assured that 534 o authorized peers authenticate themselves during the routing 535 neighbor discovery process; 537 o the routing/topology information received is generated according 538 to the protocol design. 540 However, when trust cannot be fully vested through authentication of 541 the principals alone, i.e., concerns of insider attack, assurance of 542 the truthfulness and timeliness of the received routing/topology 543 information is necessary. With regard to confidentiality, protecting 544 the routing/topology information from unauthorized exposure may be 545 desirable in certain cases but is in itself less pertinent in general 546 to the routing function. 548 One of the main problems of synchronizing security states of sleepy 549 nodes, as listed in the last subsection, lies in difficulties in 550 authentication; these nodes may not have received in time the most 551 recent update of security material. Similarly, the issues of minimal 552 manual configuration, prolonged rollout and delayed addition of 553 nodes, and network topology changes also complicate key management. 555 Hence, routing in LLNs needs to bootstrap the authentication process 556 and allow for flexible expiration scheme of authentication 557 credentials. 559 The vulnerability brought forth by some special-function nodes, e.g., 560 LBRs, requires the assurance, particularly in a security context, 562 o of the availability of communication channels and node resources; 564 o that the neighbor discovery process operates without undermining 565 routing availability. 567 There are other factors which are not part of RPL but directly 568 affecting its function. These factors include weaker barrier of 569 accessing the data or security material stored on the nodes through 570 physical means; therefore, the internal and external interfaces of a 571 node need to be adequate for guarding the integrity, and possibly the 572 confidentiality, of stored information, as well as the integrity of 573 routing and route generation processes. 575 Each individual system's use and environment will dictate how the 576 above objectives are applied, including the choices of security 577 services as well as the strengths of the mechanisms that must be 578 implemented. The next two sections take a closer look at how the RPL 579 security objectives may be compromised and how those potential 580 compromises can be countered. 582 5. Threat Sources 584 [RFC4593] provides a detailed review of the threat sources: outsiders 585 and byzantine. RPL has the same threat sources. 587 6. Threats and Attacks 589 This section outlines general categories of threats under the ISO 590 7498-2 model and highlights the specific attacks in each of these 591 categories for RPL. As defined in [RFC4949], a threat is "a 592 potential for violation of security, which exists when there is a 593 circumstance, capability, action, or event that could breach security 594 and cause harm." 596 An attack is "an assault on system security that derives from an 597 intelligent threat, i.e., an intelligent act that is a deliberate 598 attempt (especially in the sense of a method or technique) to evade 599 security services and violate the security policy of a system." 601 The subsequent subsections consider the threats and the attacks that 602 can cause security breaches under the ISO 7498-2 model to the routing 603 assets and via the routing points of access identified in 604 Section 4.1. The assessment steps through the security concerns of 605 each routing asset and looks at the attacks that can exploit routing 606 points of access. The threats and attacks identified are based on 607 the routing model analysis and associated review of the existing 608 literature. The source of the attacks is assumed to be from either 609 inside or outside attackers. While some attackers inside the network 610 will be using compromised nodes, and therefore are only able to do 611 what an ordinary node can ("node-equivalent"), other attacks may not 612 limited in memory, CPU, power consumption or long term storage. 613 Moore's law favours the attacker with access to the latest 614 capabilities, while the defenders will remain in place for years to 615 decades. 617 6.1. Threats due to failures to Authenticate 619 6.1.1. Node Impersonation 621 If an attacker can join a network using any identity, then it may be 622 able to assume the role of a legitimate (and existing node). It may 623 be able to report false readings (in metering applications), or 624 provide inappropriate control messages (in control systems involving 625 actuators) if the security of the application is implied by the 626 security of the routing system. 628 Even in systems where there application layer security, the ability 629 to impersonate a node would permit an attacker to direct traffic to 630 itself. This may permit various on-path attacks which would 631 otherwise be difficult, such replaying, delaying, or duplicating 632 (application) control messages. 634 6.1.2. Dummy Node 636 If an attacker can join a network using any identify, then it can 637 pretend to be a legitimate node, receiving any service legitimate 638 nodes receive. It may also be able to report false readings (in 639 metering applications), or provide inappropriate authorizations (in 640 control systems involving actuators), or perform any other attacks 641 that are facilitated by being able to direct traffic towards itself. 643 6.1.3. Node Resource Spam 645 If an attacker can join a network with any identify, then it can 646 continously do so with new (random) identities. This act may drain 647 down the resources of the network (battery, RAM, bandwidth). This 648 may cause legitimate nodes of the network to be unable to 649 communicate. 651 6.2. Threats and Attacks on Confidentiality 653 The assessment in Section 4.2 indicates that there are attacks 654 against the confidentiality of routing information at all points of 655 access. This threat may result in disclosure, as described in 656 Section 3.1.2 of [RFC4593], and may involve a disclosure of routing 657 information. 659 6.2.1. Routing Exchange Exposure 661 Routing exchanges include both routing information as well as 662 information associated with the establishment and maintenance of 663 neighbor state information. As indicated in Section 4.1, the 664 associated routing information assets may also include device 665 specific resource information, such as available memory, remaining 666 power, etc., that may be metrics of the routing protocol. 668 The routing exchanges will contain reachability information, which 669 would identify the relative importance of different nodes in the 670 network. Nodes higher up in the DODAG, to which more streams of 671 information flow, would be more interesting targets for other 672 attacks, and routing exchange exposures can identify them. 674 6.2.2. Routing Information (Routes and Network Topology) Exposure 676 Routes (which may be maintained in the form of the protocol 677 forwarding table) and neighbor topology information are the products 678 of the routing process that are stored within the node device 679 databases. 681 The exposure of this information will allow attackers to gain direct 682 access to the configuration and connectivity of the network thereby 683 exposing routing to targeted attacks on key nodes or links. Since 684 routes and neighbor topology information is stored within the node 685 device, attacks on the confidentiality of the information will apply 686 to the physical device including specified and unspecified internal 687 and external interfaces. 689 The forms of attack that allow unauthorized access or disclosure of 690 the routing information will include: 692 o Physical device compromise; 694 o Remote device access attacks (including those occurring through 695 remote network management or software/field upgrade interfaces). 697 Both of these attack vectors are considered a device specific issue, 698 and are out of scope for RPL to defend against. In some 699 applications, physical device compromise may be a real threat and it 700 may be necessary to provide for other devices to securely detect a 701 compromised device and react quickly to exclude it. 703 6.3. Threats and Attacks on Integrity 705 The assessment in Section 4.2 indicates that information and identity 706 assets are exposed to integrity threats from all points of access. 707 In other words, the integrity threat space is defined by the 708 potential for exploitation introduced by access to assets available 709 through routing exchanges and the on-device storage. 711 6.3.1. Routing Information Manipulation 713 Manipulation of routing information that range from neighbor states 714 to derived routes will allow unauthorized sources to influence the 715 operation and convergence of the routing protocols and ultimately 716 impact the forwarding decisions made in the network. 718 Manipulation of topology and reachability information will allow 719 unauthorized sources to influence the nodes with which routing 720 information is exchanged and updated. The consequence of 721 manipulating routing exchanges can thus lead to sub-optimality and 722 fragmentation or partitioning of the network by restricting the 723 universe of routers with which associations can be established and 724 maintained. 726 A sub-optimal network may use too much power and/or may congest some 727 routes leading to premature failure of a node, and a denial of 728 service on the entire network. 730 In addition, being able to attract network traffic can make a 731 blackhole attack more damaging. 733 The forms of attack that allow manipulation to compromise the content 734 and validity of routing information include 736 o Falsification, including overclaiming and misclaiming (claiming 737 routes to devices which the device can not in fact reach); 739 o Routing information replay; 741 o Byzantine (internal) attacks that permit corruption of routing 742 information in the node even where the node continues to be a 743 validated entity within the network (see, for example, [RFC4593] 744 for further discussions on Byzantine attacks); 746 o Physical device compromise or remote device access attacks. 748 6.3.2. Node Identity Misappropriation 750 Falsification or misappropriation of node identity between routing 751 participants opens the door for other attacks; it can also cause 752 incorrect routing relationships to form and/or topologies to emerge. 753 Routing attacks may also be mounted through less sophisticated node 754 identity misappropriation in which the valid information broadcast or 755 exchanged by a node is replayed without modification. The receipt of 756 seemingly valid information that is however no longer current can 757 result in routing disruption, and instability (including failure to 758 converge). Without measures to authenticate the routing participants 759 and to ensure the freshness and validity of the received information 760 the protocol operation can be compromised. The forms of attack that 761 misuse node identity include 763 o Identity attacks, including Sybil attacks (see [Sybil2002]) in 764 which a malicious node illegitimately assumes multiple identities; 766 o Routing information replay. 768 6.4. Threats and Attacks on Availability 770 The assessment in Section 4.2 indicates that the process and 771 resources assets are exposed to threats against availability; attacks 772 in this category may exploit directly or indirectly information 773 exchange or forwarding (see [RFC4732] for a general discussion). 775 6.4.1. Routing Exchange Interference or Disruption 777 Interference is the threat action and disruption is threat 778 consequence that allows attackers to influence the operation and 779 convergence of the routing protocols by impeding the routing 780 information exchange. 782 The forms of attack that allow interference or disruption of routing 783 exchange include: 785 o Routing information replay; 787 o ACK spoofing; 789 o Overload attacks. (Section 7.3.2) 791 In addition, attacks may also be directly conducted at the physical 792 layer in the form of jamming or interfering. 794 6.4.2. Network Traffic Forwarding Disruption 795 The disruption of the network traffic forwarding capability will 796 undermine the central function of network routers and the ability to 797 handle user traffic. This affects the availability of the network 798 because of the potential to impair the primary capability of the 799 network. 801 In addition to physical layer obstructions, the forms of attack that 802 allows disruption of network traffic forwarding include [Karlof2003] 804 o Selective forwarding attacks; 806 |Node_1|--(msg1|msg2|msg3)-->|Attacker|--(msg1|msg3)-->|Node_2| 808 Figure 2: Selective forwarding example 810 o Wormhole attacks; 812 |Node_1|-------------Unreachable---------x|Node_2| 813 | ^ 814 | Private Link | 815 '-->|Attacker_1|===========>|Attacker_2|--' 817 Figure 3: Wormhole Attacks 819 o Sinkhole attacks. 821 |Node_1| |Node_4| 822 | | 823 `--------. | 824 Falsify as \ | 825 Good Link \ | | 826 To Node_5 \ | | 827 \ V V 828 |Node_2|-->|Attacker|--Not Forwarded---x|Node_5| 829 ^ ^ \ 830 | | \ Falsify as 831 | | \Good Link 832 / | To Node_5 833 ,-------' | 834 | | 835 |Node_3| |Node_i| 837 Figure 4: sinkhole attack example 839 These attacks are generally done to both control plane and forwarding 840 plane traffic. A system that prevents control plane traffic (RPL 841 messages) from being diverted in these ways will also prevent actual 842 data from being diverted. 844 6.4.3. Communications Resource Disruption 846 Attacks mounted against the communication channel resource assets 847 needed by the routing protocol can be used as a means of disrupting 848 its operation. However, while various forms of Denial of Service 849 (DoS) attacks on the underlying transport subsystem will affect 850 routing protocol exchanges and operation (for example physical layer 851 RF jamming in a wireless network or link layer attacks), these 852 attacks cannot be countered by the routing protocol. As such, the 853 threats to the underlying transport network that supports routing is 854 considered beyond the scope of the current document. Nonetheless, 855 attacks on the subsystem will affect routing operation and so must be 856 directly addressed within the underlying subsystem and its 857 implemented protocol layers. 859 6.4.4. Node Resource Exhaustion 861 A potential threat consequence can arise from attempts to overload 862 the node resource asset by initiating exchanges that can lead to the 863 exhaustion of processing, memory, or energy resources. The 864 establishment and maintenance of routing neighbors opens the routing 865 process to engagement and potential acceptance of multiple 866 neighboring peers. Association information must be stored for each 867 peer entity and for the wireless network operation provisions made to 868 periodically update and reassess the associations. An introduced 869 proliferation of apparent routing peers can therefore have a negative 870 impact on node resources. 872 Node resources may also be unduly consumed by attackers attempting 873 uncontrolled topology peering or routing exchanges, routing replays, 874 or the generating of other data traffic floods. Beyond the 875 disruption of communications channel resources, these consequences 876 may be able to exhaust node resources only where the engagements are 877 able to proceed with the peer routing entities. Routing operation 878 and network forwarding functions can thus be adversely impacted by 879 node resources exhaustion that stems from attacks that include: 881 o Identity (including Sybil) attacks (see [Sybil2002]); 883 o Routing information replay attacks; 885 o HELLO-type flood attacks; 887 o Overload attacks. (Section 7.3.2) 889 7. Countermeasures 891 By recognizing the characteristics of LLNs that may impact routing, 892 this analysis provides the basis for understanding the capabilities 893 within RPL used to deter the identified attacks and mitigate the 894 threats. The following subsections consider such countermeasures by 895 grouping the attacks according to the classification of the ISO 896 7498-2 model so that associations with the necessary security 897 services are more readily visible. 899 7.1. Confidentiality Attack Countermeasures 901 Attacks to disclosure routing information may be mounted at the level 902 of the routing information assets, at the points of access associated 903 with routing exchanges between nodes, or through device interface 904 access. To gain access to routing/topology information, the attacker 905 may rely on a compromised node that deliberately exposes the 906 information during the routing exchange process, may rely on passive 907 wiretapping or traffic analysis, or may attempt access through a 908 component or device interface of a tampered routing node. 910 7.1.1. Countering Deliberate Exposure Attacks 912 A deliberate exposure attack is one in which an entity that is party 913 to the routing process or topology exchange allows the routing/ 914 topology information or generated route information to be exposed to 915 an unauthorized entity. 917 For instance, due to mis-configuration or inappropriate enabling of a 918 diagnostic interface, an entity might be copying ("bridging") traffic 919 from a secured ESSID/PAN to an unsecured interface. 921 A prerequisite to countering this attack is to ensure that the 922 communicating nodes are authenticated prior to data encryption 923 applied in the routing exchange. Authentication ensures that the 924 nodes are who they claim to be even though it does not provide an 925 indication of whether the node has been compromised. 927 To mitigate the risk of deliberate exposure, the process that 928 communicating nodes use to establish session keys must be peer-to- 929 peer (i.e., between the routing initiating and responding nodes). As 930 is pointed out in [RFC4107], automatic key management is critical for 931 good security. This helps ensure that neither node is exchanging 932 routing information with another peer without the knowledge of both 933 communicating peers. For a deliberate exposure attack to succeed, 934 the comprised node will need to be more overt and take independent 935 actions in order to disclose the routing information to 3rd party. 937 Note that the same measures which apply to securing routing/topology 938 exchanges between operational nodes must also extend to field tools 939 and other devices used in a deployed network where such devices can 940 be configured to participate in routing exchanges. 942 7.1.2. Countering Passive Wiretapping Attacks 944 A passive wiretap attack seeks to breach routing confidentiality 945 through passive, direct analysis and processing of the information 946 exchanges between nodes. 948 Passive wiretap attacks can be directly countered through the use of 949 data encryption for all routing exchanges. Only when a validated and 950 authenticated node association is completed will routing exchange be 951 allowed to proceed using established session keys and an agreed 952 encryption algorithm. The mandatory to implement CCM mode AES-128 953 method, is described in [RFC3610], and is believed to be secure 954 against a brute force attack by even the most well-equipped 955 adversary. 957 The significant challenge for RPL is in the provisioning of the key, 958 which in some modes of RFC6550 is used network-wide. RFC6550 does 959 not solve this problem, and it is the subject of significant future 960 work: see, for instance: [AceCharterProposal], [SolaceProposal], 961 [SmartObjectSecurityWorkshop]. 963 A number of deployments, such as [ZigBeeIP] specify no layer-3/RPL 964 encryption or authentication and rely upon similiar security at 965 layer-2. These networks are immune to outside wiretapping attacks, 966 but are vulnerable to passive (and active) routing attacks through 967 compromises of nodes. (see Section 8.2). 969 Section 10.9 of [RFC6550] specifies AES-128 in CCM mode with a 32-bit 970 MAC. 972 Section 5.6 Zigbee IP [ZigBeeIP] specifies use of CCM, with PANA and 973 EAP-TLS for key management. 975 7.1.3. Countering Traffic Analysis 977 Traffic analysis provides an indirect means of subverting 978 confidentiality and gaining access to routing information by allowing 979 an attacker to indirectly map the connectivity or flow patterns 980 (including link-load) of the network from which other attacks can be 981 mounted. The traffic analysis attack on an LLN, especially one 982 founded on shared medium, is passive and relies on the ability to 983 read the immutable source/destination layer-2 and/or layer-3 routing 984 information that must remain unencrypted to permit network routing. 986 One way in which passive traffic analysis attacks can be muted is 987 through the support of load balancing that allows traffic to a given 988 destination to be sent along diverse routing paths. RPL does not 989 generally support multi-path routing within a single DODAG. Multiple 990 DODAGs are supported in the protocol, and an implementation could 991 make use of that. RPL does not have any inherent or standard way to 992 guarantee that the different DODAGs would have significantly diverse 993 paths. Having the diverse DODAGs routed at different border routers 994 might work in some instances, and this could be combined with a 995 multipath technology like MPTCP ([RFC6824]. It is unlikely that it 996 will be affordable in many LLNs, as few deployments will have memory 997 space for more than a few sets of DODAG tables. 999 Another approach to countering passive traffic analysis could be for 1000 nodes to maintain constant amount of traffic to different 1001 destinations through the generation of arbitrary traffic flows; the 1002 drawback of course would be the consequent overhead and energy 1003 expenditure. 1005 The only means of fully countering a traffic analysis attack is 1006 through the use of tunneling (encapsulation) where encryption is 1007 applied across the entirety of the original packet source/destination 1008 addresses. Deployments which use layer-2 security that includes 1009 encryption already do this for all traffic. 1011 7.1.4. Countering Remote Device Access Attacks 1012 Where LLN nodes are deployed in the field, measures are introduced to 1013 allow for remote retrieval of routing data and for software or field 1014 upgrades. These paths create the potential for a device to be 1015 remotely accessed across the network or through a provided field 1016 tool. In the case of network management a node can be directly 1017 requested to provide routing tables and neighbor information. 1019 To ensure confidentiality of the node routing information against 1020 attacks through remote access, any local or remote device requesting 1021 routing information must be authenticated, and must be authorized for 1022 that access. Since remote access is not invoked as part of a routing 1023 protocol, security of routing information stored on the node against 1024 remote access will not be addressable as part of the routing 1025 protocol. 1027 7.2. Integrity Attack Countermeasures 1029 Integrity attack countermeasures address routing information 1030 manipulation, as well as node identity and routing information 1031 misuse. Manipulation can occur in the form of falsification attack 1032 and physical compromise. To be effective, the following development 1033 considers the two aspects of falsification, namely, the unauthorized 1034 modifications and the overclaiming and misclaiming content. The 1035 countering of physical compromise was considered in the previous 1036 section and is not repeated here. With regard to misuse, there are 1037 two types of attacks to be deterred, identity attacks and replay 1038 attacks. 1040 7.2.1. Countering Unauthorized Modification Attacks 1042 Unauthorized modifications may occur in the form of altering the 1043 message being transferred or the data stored. Therefore, it is 1044 necessary to ensure that only authorized nodes can change the portion 1045 of the information that is allowed to be mutable, while the integrity 1046 of the rest of the information is protected, e.g., through well- 1047 studied cryptographic mechanisms. 1049 Unauthorized modifications may also occur in the form of insertion or 1050 deletion of messages during protocol changes. Therefore, the 1051 protocol needs to ensure the integrity of the sequence of the 1052 exchange sequence. 1054 The countermeasure to unauthorized modifications needs to: 1056 o implement access control on storage; 1058 o provide data integrity service to transferred messages and stored 1059 data; 1061 o include sequence number under integrity protection. 1063 7.2.2. Countering Overclaiming and Misclaiming Attacks 1065 Both overclaiming and misclaiming aim to introduce false routes or a 1066 false topology that would not occur otherwise, while there are not 1067 necessarily unauthorized modifications to the routing messages or 1068 information. In order to counter overclaiming, the capability to 1069 determine unreasonable routes or topology is required. 1071 The counter to overclaiming and misclaiming may employ: 1073 o comparison with historical routing/topology data; 1075 o designs which restrict realizable network topologies. 1077 RPL includes no specific mechanisms in the protocol to counter 1078 overclaims or misclaims. An implementation could have specific 1079 heuristics implemented locally. 1081 7.2.3. Countering Identity (including Sybil) Attacks 1083 Identity attacks, sometimes simply called spoofing, seek to gain or 1084 damage assets whose access is controlled through identity. In 1085 routing, an identity attacker can illegitimately participate in 1086 routing exchanges, distribute false routing information, or cause an 1087 invalid outcome of a routing process. 1089 A perpetrator of Sybil attacks assumes multiple identities. The 1090 result is not only an amplification of the damage to routing, but 1091 extension to new areas, e.g., where geographic distribution is 1092 explicitly or implicitly an asset to an application running on the 1093 LLN, for example, the LBR in a P2MP or MP2P LLN. 1095 RPL includes specific public key based authentication at layer-3 that 1096 provide for authorization. Many deployments use layer-2 security 1097 that includes admission controls at layer-2 using mechanisms such as 1098 PANA. 1100 7.2.4. Countering Routing Information Replay Attacks 1102 In many routing protocols, message replay can result in false 1103 topology and/or routes. This is often counted with some kind of 1104 counter to ensure the freshness of the message. Replay of a current, 1105 literal RPL message are in general idempotent to the topology. An 1106 older (lower DODAGVersionNumber) message, if replayed would be 1107 rejected as being stale. The trickle algorithm further dampens the 1108 effect of any such replay, as if the message was current, then it 1109 would contain the same information as before, and it would cause no 1110 network changes. 1112 Replays may well occur in some radio technologies (not very likely, 1113 802.15.4) as a result of echos or reflections, and so some replays 1114 must be assumed to occur naturally. 1116 Note that for there to be no affect at all, the replay must be done 1117 with the same apparent power for all nodes receiving the replay. A 1118 change in apparent power might change the metrics through changes to 1119 the ETX and therefore might affect the routing even though the 1120 contents of the packet were never changed. Any replay which appears 1121 to be different should be analyzed as a Selective Forwarding Attack, 1122 Sinkhole Attack or Wormhole Attack. 1124 7.2.5. Countering Byzantine Routing Information Attacks 1126 Where a node is captured or compromised but continues to operate for 1127 a period with valid network security credentials, the potential 1128 exists for routing information to be manipulated. This compromise of 1129 the routing information could thus exist in spite of security 1130 countermeasures that operate between the peer routing devices. 1132 Consistent with the end-to-end principle of communications, such an 1133 attack can only be fully addressed through measures operating 1134 directly between the routing entities themselves or by means of 1135 external entities able to access and independently analyze the 1136 routing information. Verification of the authenticity and liveliness 1137 of the routing entities can therefore only provide a limited counter 1138 against internal (Byzantine) node attacks. 1140 For link state routing protocols where information is flooded with, 1141 for example, areas (OSPF [RFC2328]) or levels (ISIS [RFC7142]), 1142 countermeasures can be directly applied by the routing entities 1143 through the processing and comparison of link state information 1144 received from different peers. By comparing the link information 1145 from multiple sources decisions can be made by a routing node or 1146 external entity with regard to routing information validity; see 1147 Chapter 2 of [Perlman1988] for a discussion on flooding attacks. 1149 For distance vector protocols, such as RPL, where information is 1150 aggregated at each routing node it is not possible for nodes to 1151 directly detect Byzantine information manipulation attacks from the 1152 routing information exchange. In such cases, the routing protocol 1153 must include and support indirect communications exchanges between 1154 non-adjacent routing peers to provide a secondary channel for 1155 performing routing information validation. S-RIP [Wan2004] is an 1156 example of the implementation of this type of dedicated routing 1157 protocol security where the correctness of aggregate distance vector 1158 information can only be validated by initiating confirmation 1159 exchanges directly between nodes that are not routing neighbors. 1161 RPL does not provide any direct mechanisms like S-RIP. It does 1162 listen to multiple parents, and may switch parents if it begins to 1163 suspect that it is being lied to. 1165 7.3. Availability Attack Countermeasures 1167 As alluded to before, availability requires that routing information 1168 exchanges and forwarding mechanisms be available when needed so as to 1169 guarantee proper functioning of the network. This may, e.g., include 1170 the correct operation of routing information and neighbor state 1171 information exchanges, among others. We will highlight the key 1172 features of the security threats along with typical countermeasures 1173 to prevent or at least mitigate them. We will also note that an 1174 availability attack may be facilitated by an identity attack as well 1175 as a replay attack, as was addressed in Section 7.2.3 and 1176 Section 7.2.4, respectively. 1178 7.3.1. Countering HELLO Flood Attacks and ACK Spoofing Attacks 1180 HELLO Flood [Karlof2003],[I-D.suhopark-hello-wsn] and ACK Spoofing 1181 attacks are different but highly related forms of attacking an LLN. 1182 They essentially lead nodes to believe that suitable routes are 1183 available even though they are not and hence constitute a serious 1184 availability attack. 1186 A HELLO attack mounted against RPL would involve sending out (or 1187 replaying) DIO messages by the attacker. Lower power LLN nodes might 1188 then attempt to join the DODAG at a lower rank than they would 1189 otherwise. 1191 The most effective method from [I-D.suhopark-hello-wsn] is the verify 1192 bidirectionality. A number of layer-2 links are arranged in 1193 controller/spoke arrangements, and continuously are validating 1194 connectivity at layer 2. 1196 In addition, in order to calculate metrics, the ETX must be computed, 1197 and this involves, in general, sending a number of messages between 1198 nodes which are believed to be adjacent. 1199 [I-D.kelsey-intarea-mesh-link-establishment] is one such protocol. 1201 In order to join the DODAG, a DAO message is sent upwards. In RPL 1202 the DAO is acknowledged by the DAO-ACK message. This clearly checks 1203 bidirectionality at the control plane. 1205 As discussed in section 5.1, [I-D.suhopark-hello-wsn] a receiver with 1206 a sensitive receiver could well hear the DAOs, and even send DAO-ACKs 1207 as well. Such a node is a form of wormhole attack. 1209 These attacks are also all easily defended against using either 1210 layer-2 or layer-3 authentication. Such an attack could only be made 1211 against a completely open network (such as might be used for 1212 provisioning new nodes), or by a compromised node. 1214 7.3.2. Countering Overload Attacks 1216 Overload attacks are a form of DoS attack in that a malicious node 1217 overloads the network with irrelevant traffic, thereby draining the 1218 nodes' energy store more quickly, when the nodes rely on batteries or 1219 energy scavenging. It thus significantly shortens the lifetime of 1220 networks of energy-constrained nodes and constitutes another serious 1221 availability attack. 1223 With energy being one of the most precious assets of LLNs, targeting 1224 its availability is a fairly obvious attack. Another way of 1225 depleting the energy of an LLN node is to have the malicious node 1226 overload the network with irrelevant traffic. This impacts 1227 availability since certain routes get congested which: 1229 o renders them useless for affected nodes and data can hence not be 1230 delivered; 1232 o makes routes longer as shortest path algorithms work with the 1233 congested network; 1235 o depletes battery and energy scavenging nodes more quickly and thus 1236 shortens the network's availability at large. 1238 Overload attacks can be countered by deploying a series of mutually 1239 non-exclusive security measures: 1241 o introduce quotas on the traffic rate each node is allowed to send; 1243 o isolate nodes which send traffic above a certain threshold based 1244 on system operation characteristics; 1246 o allow only trusted data to be received and forwarded. 1248 As for the first one, a simple approach to minimize the harmful 1249 impact of an overload attack is to introduce traffic quotas. This 1250 prevents a malicious node from injecting a large amount of traffic 1251 into the network, even though it does not prevent said node from 1252 injecting irrelevant traffic at all. Another method is to isolate 1253 nodes from the network at the network layer once it has been detected 1254 that more traffic is injected into the network than allowed by a 1255 prior set or dynamically adjusted threshold. Finally, if 1256 communication is sufficiently secured, only trusted nodes can receive 1257 and forward traffic which also lowers the risk of an overload attack. 1259 Receiving nodes that validate signatures and sending nodes that 1260 encrypt messages need to be cautious of cryptographic processing 1261 usage when validating signatures and encrypting messages. Where 1262 feasible, certificates should be validated prior to use of the 1263 associated keys to counter potential resource overloading attacks. 1264 The associated design decision needs to also consider that the 1265 validation process requires resources and thus itself could be 1266 exploited for attacks. Alternatively, resource management limits can 1267 be placed on routing security processing events (see the comment in 1268 Section 6, paragraph 4, of [RFC5751]). 1270 7.3.3. Countering Selective Forwarding Attacks 1272 Selective forwarding attacks are a form of DoS attack which impacts 1273 the availability of the generated routing paths. 1275 A selective forwarding attack may be done by a node involved with the 1276 routing process, or it may be done by what otherwise appears to be a 1277 passive antenna or other RF feature or device, but is in fact an 1278 active (and selective) device. An RF antenna/repeater which is not 1279 selective, is not a threat. 1281 An insider malicious node basically blends neatly in with the network 1282 but then may decide to forward and/or manipulate certain packets. If 1283 all packets are dropped, then this attacker is also often referred to 1284 as a "black hole". Such a form of attack is particularly dangerous 1285 if coupled with sinkhole attacks since inherently a large amount of 1286 traffic is attracted to the malicious node and thereby causing 1287 significant damage. In a shared medium, an outside malicious node 1288 would selectively jam overheard data flows, where the thus caused 1289 collisions incur selective forwarding. 1291 Selective Forwarding attacks can be countered by deploying a series 1292 of mutually non-exclusive security measures: 1294 o multipath routing of the same message over disjoint paths; 1296 o dynamically selecting the next hop from a set of candidates. 1298 The first measure basically guarantees that if a message gets lost on 1299 a particular routing path due to a malicious selective forwarding 1300 attack, there will be another route which successfully delivers the 1301 data. Such a method is inherently suboptimal from an energy 1302 consumption point of view; it is also suboptimal from a network 1303 utilization perspective. The second method basically involves a 1304 constantly changing routing topology in that next-hop routers are 1305 chosen from a dynamic set in the hope that the number of malicious 1306 nodes in this set is negligible. A routing protocol that allows for 1307 disjoint routing paths may also be useful. 1309 7.3.4. Countering Sinkhole Attacks 1311 In sinkhole attacks, the malicious node manages to attract a lot of 1312 traffic mainly by advertising the availability of high-quality links 1313 even though there are none [Karlof2003]. It hence constitutes a 1314 serious attack on availability. 1316 The malicious node creates a sinkhole by attracting a large amount 1317 of, if not all, traffic from surrounding neighbors by advertising in 1318 and outwards links of superior quality. Affected nodes hence eagerly 1319 route their traffic via the malicious node which, if coupled with 1320 other attacks such as selective forwarding, may lead to serious 1321 availability and security breaches. Such an attack can only be 1322 executed by an inside malicious node and is generally very difficult 1323 to detect. An ongoing attack has a profound impact on the network 1324 topology and essentially becomes a problem of flow control. 1326 Sinkhole attacks can be countered by deploying a series of mutually 1327 non-exclusive security measures: 1329 o use geographical insights for flow control; 1331 o isolate nodes which receive traffic above a certain threshold; 1333 o dynamically pick up next hop from set of candidates; 1335 o allow only trusted data to be received and forwarded. 1337 Some LLNs may provide for geolocation services, often derived from 1338 solving triangulation equations from radio delay calculations, such 1339 calculations could in theory be subverted by a sinkhole that 1340 transmitted at precisely the right power in a node to node fashion. 1342 While geographic knowledge could help assure that traffic always went 1343 in the physical direction desired, it would not assure that the 1344 traffic was taking the most efficient route, as the lowest cost real 1345 route might be match the physical topology; such as when different 1346 parts of an LLN are connected by high-speed wired networks. 1348 7.3.5. Countering Wormhole Attacks 1350 In wormhole attacks at least two malicious nodes claim to have a 1351 short path between themselves [Karlof2003]. This changes the 1352 availability of certain routing paths and hence constitutes a serious 1353 security breach. 1355 Essentially, two malicious insider nodes use another, more powerful, 1356 transmitter to communicate with each other and thereby distort the 1357 would-be-agreed routing path. This distortion could involve 1358 shortcutting and hence paralyzing a large part of the network; it 1359 could also involve tunneling the information to another region of the 1360 network where there are, e.g., more malicious nodes available to aid 1361 the intrusion or where messages are replayed, etc. 1363 In conjunction with selective forwarding, wormhole attacks can create 1364 race conditions which impact topology maintenance, routing protocols 1365 as well as any security suits built on "time of check" and "time of 1366 use". 1368 A pure wormhole attack is nearly impossible to detect. A wormhole 1369 which is used in order to subsequently mount another kind of attack 1370 would be defeated by defeating the other attack. A perfect wormhole, 1371 in which there is nothing adverse that occurs to the traffic, would 1372 be difficult to call an attack. The worst thing that a benign 1373 wormhole can do in such a situation is to cease to operate (become 1374 unstable), causing the network to have to recalculate routes. 1376 A highly unstable wormhole is no different than a radio opaque (i.e. 1377 metal) door that opens and closes a lot. RPL includes hysteresis in 1378 its objective functions [RFC6719] in an attempt to deal with frequent 1379 changes to the ETX between nodes. 1381 8. RPL Security Features 1383 The assessments and analysis in Section 6 examined all areas of 1384 threats and attacks that could impact routing, and the 1385 countermeasures presented in Section 7 were reached without confining 1386 the consideration to means only available to routing. This section 1387 puts the results into perspective; dealing with those threats which 1388 are endemic to this field, those which have been mitigated through 1389 RPL protocol design, and those which require specific decisions to be 1390 made as part of provisioning a network. 1392 The first part of this section, Section 8.1 to Section 8.3, is a 1393 description of RPL security features that address specific threats. 1394 The second part of this section, Section 8.4, discusses issues of 1395 provisioning of security aspects that may impact routing but that 1396 also require considerations beyond the routing protocol, as well as 1397 potential approaches. 1399 RPL employs multicast and so these alternative communications modes 1400 MUST be secured with the same routing security services specified in 1401 this section. Furthermore, irrespective of the modes of 1402 communication, nodes MUST provide adequate physical tamper resistance 1403 commensurate with the particular application domain environment to 1404 ensure the confidentiality, integrity, and availability of stored 1405 routing information. 1407 8.1. Confidentiality Features 1409 With regard to confidentiality, protecting the routing/topology 1410 information from unauthorized disclosure is not directly essential to 1411 maintaining the routing function. Breaches of confidentiality may 1412 lead to other attacks or the focusing of an attacker's resources (see 1413 Section 6.2) but does not of itself directly undermine the operation 1414 of the routing function. However, to protect against, and reduce 1415 consequences from other more direct attacks, routing information 1416 should be protected. Thus, to secure RPL: 1418 o implement payload encryption using layer-3 mechanisms described in 1419 [RFC6550]; 1421 o or: implement layer-2 confidentiality; 1423 Where confidentiality is incorporated into the routing exchanges, 1424 encryption algorithms and key lengths need to be specified in 1425 accordance with the level of protection dictated by the routing 1426 protocol and the associated application domain transport network. 1427 For most networks, this means use of AES128 in CCM mode, but this 1428 needs to be specified clearly in the applicability statement. 1430 In terms of the life time of the keys, the opportunity to 1431 periodically change the encryption key increases the offered level of 1432 security for any given implementation. However, where strong 1433 cryptography is employed, physical, procedural, and logical data 1434 access protection considerations may have more significant impact on 1435 cryptoperiod selection than algorithm and key size factors. 1436 Nevertheless, in general, shorter cryptoperiods, during which a 1437 single key is applied, will enhance security. 1439 Given the mandatory protocol requirement to implement routing node 1440 authentication as part of routing integrity (see Section 8.2), key 1441 exchanges may be coordinated as part of the integrity verification 1442 process. This provides an opportunity to increase the frequency of 1443 key exchange and shorten the cryptoperiod as a complement to the key 1444 length and encryption algorithm required for a given application 1445 domain. 1447 8.2. Integrity Features 1449 The integrity of routing information provides the basis for ensuring 1450 that the function of the routing protocol is achieved and maintained. 1451 To protect integrity, RPL must either run using only the Secure 1452 versions of the messages, or must run over a layer-2 that uses 1453 channel binding between node identity and transmissions. 1455 Some layer-2 security mechanisms use a single key for the entire 1456 network, and these networks can not provide significant amount of 1457 integrity protection, as any node that has that key may impersonate 1458 any other node. This mode of operation is likely acceptable when an 1459 entire deployment is under the control of a single administrative 1460 entity. 1462 Other layer-2 security mechanisms form a unique session key for every 1463 pair of nodes that needs to communicate; this is often called a per- 1464 link key. Such networks can provide a strong degree of origin 1465 authentication and integrity on unicast messages. 1467 However, some RPL messages are broadcast, and even when per-node 1468 layer-2 security mechanisms are used, the integrity and origin 1469 authentication of broadcast messages can not be as trusted due to the 1470 proliferation of the key used to secure them. 1472 RPL has two specific options which are broadcast in RPL Control 1473 Messages: the DODAG Information Object (DIO), and the DODAG 1474 Information Solicitation (DIS). The purpose of the DIS is to cause 1475 potential parents to reply with a DIO, so the integrity of the DIS is 1476 not of great concern. The DIS may also be unicast. 1478 The DIO is a critical piece of routing and carries many critical 1479 parameters. RPL provides for asymmetric authentication at layer 3 of 1480 the RPL Control Message carrying the DIO and this may be warranted in 1481 some deployments. A node could, if it felt that the DIO that it had 1482 received was suspicious, send a unicast DIS message to the node in 1483 question, and that node would reply with a unicast DIS. Those 1484 messages could be protected with the per-link key. 1486 8.3. Availability Features 1488 Availability of routing information is linked to system and network 1489 availability which in the case of LLNs require a broader security 1490 view beyond the requirements of the routing entities. Where 1491 availability of the network is compromised, routing information 1492 availability will be accordingly affected. However, to specifically 1493 assist in protecting routing availability, nodes: 1495 o MAY restrict neighborhood cardinality; 1497 o MAY use multiple paths; 1499 o MAY use multiple destinations; 1501 o MAY choose randomly if multiple paths are available; 1503 o MAY set quotas to limit transmit or receive volume; 1505 o MAY use geographic information for flow control. 1507 8.4. Key Management 1509 The functioning of the routing security services requires keys and 1510 credentials. Therefore, even though not directly a RPL security 1511 requirement, an LLN MUST have a process for initial key and 1512 credential configuration, as well as secure storage within the 1513 associated devices. Anti-tampering SHOULD be a consideration in 1514 physical design. Beyond initial credential configuration, an LLN is 1515 also encouraged to have automatic procedures for the revocation and 1516 replacement of the maintained security credentials. 1518 While RPL has secure modes, but some modes are impractical without 1519 use of public key cryptography believed to be too expensive by many. 1520 RPL layer-3 security will often depend upon existing LLN layer-2 1521 security mechanisms, which provides for node authentication, but 1522 little in the way of node authorization. 1524 9. IANA Considerations 1526 This memo includes no request to IANA. 1528 10. Security Considerations 1530 The analysis presented in this document provides security analysis 1531 and design guidelines with a scope limited to RPL. Security services 1532 are identified as requirements for securing RPL. The specific 1533 mechanisms to be used to deal with each threat is specified in link- 1534 layer and deployment specific applicability statements. 1536 11. Acknowledgments 1538 The authors would like to acknowledge the review and comments from 1539 Rene Struik and JP Vasseur. The authors would also like to 1540 acknowledge the guidance and input provided by the RPL Chairs, David 1541 Culler, and JP Vasseur, and the Area Director Adrian Farrel. 1543 This document started out as a combined threat and solutions 1544 document. As a result of security review, the document was split up 1545 by RPL co-Chair Michael Richardson and security Area Director Sean 1546 Turner as it went through the IETF publication process. The 1547 solutions to the threats are application and layer-2 specific, and 1548 have therefore been moved to the relevant applicability statements. 1550 Ines Robles and Robert Cragie kept track of the many issues that were 1551 raised during the development of this document 1553 12. References 1555 12.1. Normative References 1557 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 1558 Requirement Levels", BCP 14, RFC 2119, March 1997. 1560 [RFC4107] Bellovin, S. and R. Housley, "Guidelines for Cryptographic 1561 Key Management", BCP 107, RFC 4107, June 2005. 1563 [RFC6550] Winter, T., Thubert, P., Brandt, A., Hui, J., Kelsey, R., 1564 Levis, P., Pister, K., Struik, R., Vasseur, JP., and R. 1565 Alexander, "RPL: IPv6 Routing Protocol for Low-Power and 1566 Lossy Networks", RFC 6550, March 2012. 1568 [RFC6719] Gnawali, O. and P. Levis, "The Minimum Rank with 1569 Hysteresis Objective Function", RFC 6719, September 2012. 1571 [RFC7102] Vasseur, JP., "Terms Used in Routing for Low-Power and 1572 Lossy Networks", RFC 7102, January 2014. 1574 [ZigBeeIP] 1575 ZigBee Public Document 15-002r00, "ZigBee IP 1576 Specification", 2013. 1578 12.2. Informative References 1580 [AceCharterProposal] 1581 Li, Kepeng., Ed., "Authentication and Authorization for 1582 Constrained Environment Charter (work-in-progress)", 1583 December 2013, . 1586 [I-D.gilger-smart-object-security-workshop] 1587 Gilger, J. and H. Tschofenig, "Report from the 'Smart 1588 Object Security Workshop', March 23, 2012, Paris, France", 1589 draft-gilger-smart-object-security-workshop-02 (work in 1590 progress), October 2013. 1592 [I-D.kelsey-intarea-mesh-link-establishment] 1593 Kelsey, R., "Mesh Link Establishment", draft-kelsey- 1594 intarea-mesh-link-establishment-05 (work in progress), 1595 February 2013. 1597 [I-D.suhopark-hello-wsn] 1598 Park, S., "Routing Security in Sensor Network: HELLO Flood 1599 Attack and Defense", draft-suhopark-hello-wsn-00 (work in 1600 progress), December 2005. 1602 [IEEE.802.11] 1603 , "Draft Standard for Information Technology - 1604 Telecommunications and information exchange between 1605 systems - Local and metropolitan area networks Specific 1606 requirements - Part 11: Wireless LAN Medium Access Control 1607 (MAC) and Physical Layer (PHY) specifications ", IEEE 1608 802.11-REVma, 2006. 1610 [IEEE.802.15.4] 1611 , "Information technology - Telecommunications and 1612 information exchange between systems - Local and 1613 metropolitan area networks - Specific requirements - Part 1614 15.4: Wireless Medium Access Control (MAC) and Physical 1615 Layer (PHY) Specifications for Low Rate Wireless Personal 1616 Area Networks (LR-WPANs) ", IEEE Std 802.15.4-2006, June 1617 2006, . 1619 [ISO.7498-2.1988] 1620 International Organization for Standardization, 1621 "Information Processing Systems - Open Systems 1622 Interconnection Reference Model - Security Architecture", 1623 ISO Standard 7498-2, 1988. 1625 [Karlof2003] 1626 Karlof, C. and D. Wagner, "Secure routing in wireless 1627 sensor networks: attacks and countermeasures", Elsevier 1628 AdHoc Networks Journal, Special Issue on Sensor Network 1629 Applications and Protocols, 1(2):293-315, September 2003, 1630 . 1633 [Myagmar2005] 1634 Myagmar, S., Lee, AJ., and W. Yurcik, "Threat Modeling as 1635 a Basis for Security Requirements", in Proceedings of the 1636 Symposium on Requirements Engineering for Information 1637 Security (SREIS'05), Paris, France, pp. 94-102, Aug 29, 1638 2005. 1640 [Perlman1988] 1641 Perlman, N., "Network Layer Protocols with Byzantine 1642 Robustness", MIT LCS Tech Report, 429, 1988. 1644 [RFC2328] Moy, J., "OSPF Version 2", STD 54, RFC 2328, April 1998. 1646 [RFC3610] Whiting, D., Housley, R., and N. Ferguson, "Counter with 1647 CBC-MAC (CCM)", RFC 3610, September 2003. 1649 [RFC4593] Barbir, A., Murphy, S., and Y. Yang, "Generic Threats to 1650 Routing Protocols", RFC 4593, October 2006. 1652 [RFC4732] Handley, M., Rescorla, E., IAB, "Internet Denial-of- 1653 Service Considerations", RFC 4732, December 2006. 1655 [RFC4949] Shirey, R., "Internet Security Glossary, Version 2", RFC 1656 4949, August 2007. 1658 [RFC5191] Forsberg, D., Ohba, Y., Patil, B., Tschofenig, H., and A. 1659 Yegin, "Protocol for Carrying Authentication for Network 1660 Access (PANA)", RFC 5191, May 2008. 1662 [RFC5216] Simon, D., Aboba, B., and R. Hurst, "The EAP-TLS 1663 Authentication Protocol", RFC 5216, March 2008. 1665 [RFC5548] Dohler, M., Watteyne, T., Winter, T., and D. Barthel, 1666 "Routing Requirements for Urban Low-Power and Lossy 1667 Networks", RFC 5548, May 2009. 1669 [RFC5673] Pister, K., Thubert, P., Dwars, S., and T. Phinney, 1670 "Industrial Routing Requirements in Low-Power and Lossy 1671 Networks", RFC 5673, October 2009. 1673 [RFC5751] Ramsdell, B. and S. Turner, "Secure/Multipurpose Internet 1674 Mail Extensions (S/MIME) Version 3.2 Message 1675 Specification", RFC 5751, January 2010. 1677 [RFC5826] Brandt, A., Buron, J., and G. Porcu, "Home Automation 1678 Routing Requirements in Low-Power and Lossy Networks", RFC 1679 5826, April 2010. 1681 [RFC5867] Martocci, J., De Mil, P., Riou, N., and W. Vermeylen, 1682 "Building Automation Routing Requirements in Low-Power and 1683 Lossy Networks", RFC 5867, June 2010. 1685 [RFC6192] Dugal, D., Pignataro, C., and R. Dunn, "Protecting the 1686 Router Control Plane", RFC 6192, March 2011. 1688 [RFC6574] Tschofenig, H. and J. Arkko, "Report from the Smart Object 1689 Workshop", RFC 6574, April 2012. 1691 [RFC6824] Ford, A., Raiciu, C., Handley, M., and O. Bonaventure, 1692 "TCP Extensions for Multipath Operation with Multiple 1693 Addresses", RFC 6824, January 2013. 1695 [RFC7142] Shand, M. and L. Ginsberg, "Reclassification of RFC 1142 1696 to Historic", RFC 7142, February 2014. 1698 [SmartObjectSecurityWorkshop] 1699 Klausen, T., Ed., "Workshop on Smart Object Security", 1700 March 2012, . 1703 [SolaceProposal] 1704 Bormann, C., Ed., "Notes from the SOLACE ad-hoc at IETF85 1705 (work-in-progress)", November 2012, . 1708 [Sybil2002] 1709 Douceur, J., "The Sybil Attack", First International 1710 Workshop on Peer-to-Peer Systems , March 2002. 1712 [Wan2004] Wan, T., Kranakis, E., and PC. van Oorschot, "S-RIP: A 1713 Secure Distance Vector Routing Protocol", in Proceedings 1714 of the 2nd International Conference on Applied 1715 Cryptography and Network Security, Yellow Mountain, China, 1716 pp. 103-119, Jun. 8-11 2004. 1718 [Yourdon1979] 1719 Yourdon, E. and L. Constantine, "Structured Design", 1720 Yourdon Press, New York, Chapter 10, pp. 187-222, 1979. 1722 Authors' Addresses 1723 Tzeta Tsao 1724 Cooper Power Systems 1725 910 Clopper Rd. Suite 201S 1726 Gaithersburg, Maryland 20878 1727 USA 1729 Email: tzeta.tsao@cooperindustries.com 1731 Roger K. Alexander 1732 Cooper Power Systems 1733 910 Clopper Rd. Suite 201S 1734 Gaithersburg, Maryland 20878 1735 USA 1737 Email: roger.alexander@cooperindustries.com 1739 Mischa Dohler 1740 CTTC 1741 Parc Mediterrani de la Tecnologia, Av. Canal Olimpic S/N 1742 Castelldefels, Barcelona 08860 1743 Spain 1745 Email: mischa.dohler@cttc.es 1747 Vanesa Daza 1748 Universitat Pompeu Fabra 1749 P/ Circumval.lacio 8, Oficina 308 1750 Barcelona 08003 1751 Spain 1753 Email: vanesa.daza@upf.edu 1755 Angel Lozano 1756 Universitat Pompeu Fabra 1757 P/ Circumval.lacio 8, Oficina 309 1758 Barcelona 08003 1759 Spain 1761 Email: angel.lozano@upf.edu 1762 Michael Richardson (ed) (editor) 1763 Sandelman Software Works 1764 470 Dawson Avenue 1765 Ottawa, ON K1Z5V7 1766 Canada 1768 Email: mcr+ietf@sandelman.ca