idnits 2.17.1 draft-ietf-dots-requirements-22.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 -- however, there's a paragraph with a matching beginning. Boilerplate error? (The document does seem to have the reference to RFC 2119 which the ID-Checklist requires). -- The document date (March 23, 2019) is 1859 days in the past. Is this intentional? Checking references for intended status: Informational ---------------------------------------------------------------------------- ** Obsolete normative reference: RFC 793 (Obsoleted by RFC 9293) == Outdated reference: A later version (-18) exists of draft-ietf-dots-architecture-12 == Outdated reference: A later version (-25) exists of draft-ietf-dots-use-cases-17 == Outdated reference: A later version (-17) exists of draft-ietf-intarea-frag-fragile-09 -- Obsolete informational reference (is this intentional?): RFC 7525 (Obsoleted by RFC 9325) Summary: 1 error (**), 0 flaws (~~), 5 warnings (==), 2 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 DOTS A. Mortensen 3 Internet-Draft Arbor Networks 4 Intended status: Informational T. Reddy 5 Expires: September 24, 2019 McAfee 6 R. Moskowitz 7 Huawei 8 March 23, 2019 10 Distributed Denial of Service (DDoS) Open Threat Signaling Requirements 11 draft-ietf-dots-requirements-22 13 Abstract 15 This document defines the requirements for the Distributed Denial of 16 Service (DDoS) Open Threat Signaling (DOTS) protocols enabling 17 coordinated response to DDoS attacks. 19 Status of This Memo 21 This Internet-Draft is submitted in full conformance with the 22 provisions of BCP 78 and BCP 79. 24 Internet-Drafts are working documents of the Internet Engineering 25 Task Force (IETF). Note that other groups may also distribute 26 working documents as Internet-Drafts. The list of current Internet- 27 Drafts is at https://datatracker.ietf.org/drafts/current/. 29 Internet-Drafts are draft documents valid for a maximum of six months 30 and may be updated, replaced, or obsoleted by other documents at any 31 time. It is inappropriate to use Internet-Drafts as reference 32 material or to cite them other than as "work in progress." 34 This Internet-Draft will expire on September 24, 2019. 36 Copyright Notice 38 Copyright (c) 2019 IETF Trust and the persons identified as the 39 document authors. All rights reserved. 41 This document is subject to BCP 78 and the IETF Trust's Legal 42 Provisions Relating to IETF Documents 43 (https://trustee.ietf.org/license-info) in effect on the date of 44 publication of this document. Please review these documents 45 carefully, as they describe your rights and restrictions with respect 46 to this document. Code Components extracted from this document must 47 include Simplified BSD License text as described in Section 4.e of 48 the Trust Legal Provisions and are provided without warranty as 49 described in the Simplified BSD License. 51 Table of Contents 53 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 2 54 1.1. Context and Motivation . . . . . . . . . . . . . . . . . 2 55 1.2. Terminology . . . . . . . . . . . . . . . . . . . . . . . 3 56 2. Requirements . . . . . . . . . . . . . . . . . . . . . . . . 5 57 2.1. General Requirements . . . . . . . . . . . . . . . . . . 7 58 2.2. Signal Channel Requirements . . . . . . . . . . . . . . . 8 59 2.3. Data Channel Requirements . . . . . . . . . . . . . . . . 13 60 2.4. Security Requirements . . . . . . . . . . . . . . . . . . 14 61 2.5. Data Model Requirements . . . . . . . . . . . . . . . . . 16 62 3. Congestion Control Considerations . . . . . . . . . . . . . . 17 63 3.1. Signal Channel . . . . . . . . . . . . . . . . . . . . . 17 64 3.2. Data Channel . . . . . . . . . . . . . . . . . . . . . . 17 65 4. Security Considerations . . . . . . . . . . . . . . . . . . . 17 66 5. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 18 67 6. Contributors . . . . . . . . . . . . . . . . . . . . . . . . 18 68 7. Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . 19 69 8. References . . . . . . . . . . . . . . . . . . . . . . . . . 19 70 8.1. Normative References . . . . . . . . . . . . . . . . . . 19 71 8.2. Informative References . . . . . . . . . . . . . . . . . 20 72 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 21 74 1. Introduction 76 1.1. Context and Motivation 78 Distributed Denial of Service (DDoS) attacks afflict networks 79 connected to the Internet, plaguing network operators at service 80 providers and enterprises around the world. High-volume attacks 81 saturating inbound links are now common, as attack scale and 82 frequency continue to increase. 84 The prevalence and impact of these DDoS attacks has led to an 85 increased focus on coordinated attack response. However, many 86 enterprises lack the resources or expertise to operate on-premises 87 attack mitigation solutions themselves, or are constrained by local 88 bandwidth limitations. To address such gaps, service providers have 89 begun to offer on-demand traffic scrubbing services, which are 90 designed to separate the DDoS attack traffic from legitimate traffic 91 and forward only the latter. 93 Today, these services offer proprietary interfaces for subscribers to 94 request attack mitigation. Such proprietary interfaces tie a 95 subscriber to a service and limit the abilities of network elements 96 that would otherwise be capable of participating in attack 97 mitigation. As a result of signaling interface incompatibility, 98 attack responses may be fragmented or otherwise incomplete, leaving 99 operators in the attack path unable to assist in the defense. 101 A standardized method to coordinate a real-time response among 102 involved operators will increase the speed and effectiveness of DDoS 103 attack mitigation, and reduce the impact of these attacks. This 104 document describes the required characteristics of protocols that 105 enable attack response coordination and mitigation of DDoS attacks. 107 DDoS Open Threat Signaling (DOTS) communicates the need for defensive 108 action in anticipation of or in response to an attack, but does not 109 dictate the implementation of these actions. The DOTS use cases are 110 discussed in [I-D.ietf-dots-use-cases] and the DOTS architecture is 111 discussed in [I-D.ietf-dots-architecture]. 113 1.2. Terminology 115 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 116 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this 117 document are to be interpreted as described in BCP14 [RFC2119] 118 [RFC8174], when, and only when, they appear in all capitals. These 119 capitalized words are used to signify the requirements for DOTS 120 protocols design. 122 This document adopts the following terms: 124 DDoS: A distributed denial-of-service attack, in which traffic 125 originating from multiple sources is directed at a target on a 126 network. DDoS attacks are intended to cause a negative impact on 127 the availability and/or other functionality of an attack target. 128 Denial-of-service considerations are discussed in detail in 129 [RFC4732]. 131 DDoS attack target: A network connected entity that is the target of 132 a DDoS attack. Potential targets include (but are not limited to) 133 network elements, network links, servers, and services. 135 DDoS attack telemetry: Collected measurements and behavioral 136 characteristics defining the nature of a DDoS attack. 138 Countermeasure: An action or set of actions focused on recognizing 139 and filtering out specific types of DDoS attack traffic while 140 passing legitimate traffic to the attack target. Distinct 141 countermeasures can be layered to defend against attacks combining 142 multiple DDoS attack types. 144 Mitigation: A set of countermeasures enforced against traffic 145 destined for the target or targets of a detected or reported DDoS 146 attack, where countermeasure enforcement is managed by an entity 147 in the network path between attack sources and the attack target. 148 Mitigation methodology is out of scope for this document. 150 Mitigator: An entity, typically a network element, capable of 151 performing mitigation of a detected or reported DDoS attack. The 152 means by which this entity performs these mitigations and how they 153 are requested of it are out of scope for this document. The 154 mitigator and DOTS server receiving a mitigation request are 155 assumed to belong to the same administrative entity. 157 DOTS client: A DOTS-aware software module responsible for requesting 158 attack response coordination with other DOTS-aware elements. 160 DOTS server: A DOTS-aware software module handling and responding to 161 messages from DOTS clients. The DOTS server enables mitigation on 162 behalf of the DOTS client, if requested, by communicating the DOTS 163 client's request to the mitigator and returning selected mitigator 164 feedback to the requesting DOTS client. 166 DOTS agent: Any DOTS-aware software module capable of participating 167 in a DOTS signal or data channel. It can be a DOTS client, DOTS 168 server, or, as a logical agent, a DOTS gateway. 170 DOTS gateway: A DOTS-aware software module resulting from the 171 logical concatenation of the functionality of a DOTS server and a 172 DOTS client into a single DOTS agent. This functionality is 173 analogous to a Session Initiation Protocol (SIP) [RFC3261] Back- 174 to-Back User Agent (B2BUA) [RFC7092]. A DOTS gateway has a 175 client-facing side, which behaves as a DOTS server for downstream 176 clients, and a server-facing side, which performs the role of DOTS 177 client for upstream DOTS servers. Client-domain DOTS gateways are 178 DOTS gateways that are in the DOTS client's domain, while server- 179 domain DOTS gateways denote DOTS gateways that are in the DOTS 180 server's domain. A DOTS gateway may terminate multiple discrete 181 DOTS client connections and may aggregate these into a single or 182 multiple connections. DOTS gateways are described further in 183 [I-D.ietf-dots-architecture]. 185 Signal channel: A bidirectional, mutually authenticated 186 communication channel between DOTS agents that is resilient even 187 in conditions leading to severe packet loss, such as a volumetric 188 DDoS attack causing network congestion. 190 DOTS signal: A status/control message transmitted over the 191 authenticated signal channel between DOTS agents, used to indicate 192 the client's need for mitigation, or to convey the status of any 193 requested mitigation. 195 Heartbeat: A message transmitted between DOTS agents over the signal 196 channel, used as a keep-alive and to measure peer health. 198 Data channel: A bidirectional, mutually authenticated communication 199 channel between two DOTS agents used for infrequent but reliable 200 bulk exchange of data not easily or appropriately communicated 201 through the signal channel. Reliable bulk data exchange may not 202 function well or at all during attacks causing network congestion. 203 The data channel is not expected to operate in such conditions. 205 Filter: A specification of a matching network traffic flow or set of 206 flows. The filter will typically have a policy associated with 207 it, e.g., rate-limiting or discarding matching traffic [RFC4949]. 209 Drop-list: A list of filters indicating sources from which traffic 210 should be blocked, regardless of traffic content. 212 Accept-list: A list of filters indicating sources from which traffic 213 should always be allowed, regardless of contradictory data gleaned 214 in a detected attack. 216 Multi-homed DOTS client: A DOTS client exchanging messages with 217 multiple DOTS servers, each in a separate administrative domain. 219 2. Requirements 221 The expected layout and interactions amongst DOTS entities is 222 described in the DOTS Architecture [I-D.ietf-dots-architecture]. 224 The goal of the DOTS requirements specification is to specify the 225 requirements for DOTS signal channel and data channel protocols that 226 have different application and transport layer requirements. This 227 section describes the required features and characteristics of the 228 DOTS protocols. 230 The goal of DOTS protocols is to enable and manage mitigation on 231 behalf of a network domain or resource which is or may become the 232 focus of a DDoS attack. An active DDoS attack against the entity 233 controlling the DOTS client need not be present before establishing a 234 communication channel between DOTS agents. Indeed, establishing a 235 relationship with peer DOTS agents during normal network conditions 236 provides the foundation for more rapid attack response against future 237 attacks, as all interactions setting up DOTS, including any business 238 or service level agreements, are already complete. Reachability 239 information of peer DOTS agents is provisioned to a DOTS client using 240 a variety of manual or dynamic methods. Once a relationship between 241 DOTS agents is established, regular communication between DOTS 242 clients and servers enables a common understanding of the DOTS 243 agents' health and activity. 245 The DOTS protocol must at a minimum make it possible for a DOTS 246 client to request aid mounting a defense against a suspected attack. 247 This defense could be coordinated by a DOTS server and include 248 signaling within or between domains as requested by local operators. 249 DOTS clients should similarly be able to withdraw aid requests. DOTS 250 requires no justification from DOTS clients for requests for help, 251 nor do DOTS clients need to justify withdrawing help requests: the 252 decision is local to the DOTS clients' domain. Multi-homed DOTS 253 clients must be able to select the appropriate DOTS server(s) to 254 which a mitigation request is to be sent. The method for selecting 255 the appropriate DOTS server in a multi-homed environment is out of 256 scope for this document. 258 DOTS protocol implementations face competing operational goals when 259 maintaining this bidirectional communication stream. On the one 260 hand, DOTS must include measures to ensure message confidentiality, 261 integrity, authenticity, and replay protection to keep the protocols 262 from becoming additional vectors for the very attacks it is meant to 263 help fight off. On the other hand, the protocol must be resilient 264 under extremely hostile network conditions, providing continued 265 contact between DOTS agents even as attack traffic saturates the 266 link. Such resiliency may be developed several ways, but 267 characteristics such as small message size, asynchronous, redundant 268 message delivery and minimal connection overhead (when possible given 269 local network policy) will tend to contribute to the robustness 270 demanded by a viable DOTS protocol. Operators of peer DOTS-enabled 271 domains may enable quality- or class-of-service traffic tagging to 272 increase the probability of successful DOTS signal delivery, but DOTS 273 does not require such policies be in place, and should be viable in 274 their absence. 276 The DOTS server and client must also have some standardized method of 277 defining the scope of any mitigation, as well as managing other 278 mitigation-related configuration. 280 Finally, DOTS should be sufficiently extensible to meet future needs 281 in coordinated attack defense, although this consideration is 282 necessarily superseded by the other operational requirements. 284 2.1. General Requirements 286 GEN-001 Extensibility: Protocols and data models developed as part 287 of DOTS MUST be extensible in order to keep DOTS adaptable to 288 proprietary DDoS defenses. Future extensions MUST be backward 289 compatible. Implementations of older protocol versions MUST 290 ignore optional information added to DOTS messages as part of 291 newer protocol versions. Implementations of older protocol 292 versions MUST reject DOTS messages carrying mandatory information 293 as part of newer protocol versions. 295 GEN-002 Resilience and Robustness: The signaling protocol MUST be 296 designed to maximize the probability of signal delivery even under 297 the severely constrained network conditions caused by attack 298 traffic. Additional means to enhance the resilience of DOTS 299 protocols, including when multiple DOTS servers are provisioned to 300 the DOTS clients, SHOULD be considered. The protocol MUST be 301 resilient, that is, continue operating despite message loss and 302 out-of-order or redundant message delivery. In support of 303 signaling protocol robustness, DOTS signals SHOULD be conveyed 304 over transport and application protocols not susceptible to Head 305 of Line Blocking. These requirements are at SHOULD strength to 306 handle middle-boxes and firewall traversal. 308 GEN-003 Bulk Data Exchange: Infrequent bulk data exchange between 309 DOTS agents can also significantly augment attack response 310 coordination, permitting such tasks as population of drop- or 311 accept-listed source addresses; address or prefix group aliasing; 312 exchange of incident reports; and other hinting or configuration 313 supplementing attack response. 315 As the resilience requirements for the DOTS signal channel mandate 316 small signal message size, a separate, secure data channel 317 utilizing a reliable transport protocol MUST be used for bulk data 318 exchange. However, reliable bulk data exchange may not be 319 possible during attacks causing network congestion. 321 GEN-004 Mitigation Hinting: DOTS clients may have access to attack 322 details which can be used to inform mitigation techniques. 323 Example attack details might include locally collected 324 fingerprints for an on-going attack, or anticipated or active 325 attack focal points based on other threat intelligence. DOTS 326 clients MAY send mitigation hints derived from attack details to 327 DOTS servers, in the full understanding that the DOTS server MAY 328 ignore mitigation hints. Mitigation hints MUST be transmitted 329 across the signal channel, as the data channel may not be 330 functional during an attack. DOTS server handling of mitigation 331 hints is implementation-specific. 333 GEN-005 Loop Handling: In certain scenarios, typically involving 334 misconfiguration of DNS or routing policy, it may be possible for 335 communication between DOTS agents to loop. Signal and data 336 channel implementations should be prepared to detect and terminate 337 such loops to prevent service disruption. 339 2.2. Signal Channel Requirements 341 SIG-001 Use of Common Transport Protocols: DOTS MUST operate over 342 common widely deployed and standardized transport protocols. 343 While connectionless transport such as the User Datagram Protocol 344 (UDP) [RFC0768] SHOULD be used for the signal channel, the 345 Transmission Control Protocol (TCP) [RFC0793] MAY be used if 346 necessary due to network policy or middlebox capabilities or 347 configurations. 349 SIG-002 Sub-MTU Message Size: To avoid message fragmentation and the 350 consequently decreased probability of message delivery over a 351 congested link, signaling protocol message size MUST be kept under 352 signaling Path Maximum Transmission Unit (PMTU), including the 353 byte overhead of any encapsulation, transport headers, and 354 transport- or message-level security. If the total message size 355 exceeds the path MTU, the DOTS agent MUST split the message into 356 separate messages; for example, the list of mitigation scope types 357 could be split into multiple lists and each list conveyed in a new 358 message. 360 DOTS agents can attempt to learn PMTU using the procedures 361 discussed in [I-D.ietf-intarea-frag-fragile]. If the PMTU cannot 362 be discovered, DOTS agents MUST assume a PMTU of 1280 bytes, as 363 IPv6 requires that every link in the Internet have an MTU of 1280 364 octets or greater as specified in [RFC8200]. If IPv4 support on 365 legacy or otherwise unusual networks is a consideration and the 366 PMTU is unknown, DOTS implementations MAY assume on a PMTU of 576 367 bytes for IPv4 datagrams, as every IPv4 host must be capable of 368 receiving a packet whose length is equal to 576 bytes as discussed 369 in [RFC0791] and [RFC1122]. 371 SIG-003 Bidirectionality: To support peer health detection, to 372 maintain an active signal channel, and to increase the probability 373 of signal delivery during an attack, the signal channel MUST be 374 bidirectional, with client and server transmitting signals to each 375 other at regular intervals, regardless of any client request for 376 mitigation. The bidirectional signal channel MUST support 377 unidirectional messaging to enable notifications between DOTS 378 agents. 380 SIG-004 Channel Health Monitoring: DOTS agents MUST support exchange 381 of heartbeat messages over the signal channel to monitor channel 382 health. These keepalives serve the purpose to maintain any on- 383 path NAT or Firewall bindings to avoid cryptographic handshake for 384 new mitigation requests. The heartbeat interval during active 385 mitigation could be negotiable based on NAT/Firewall 386 characteristics. Absent information about the NAT/Firewall 387 characteristics, DOTS agent needs to ensure its on-path NAT or 388 Firewall bindings do not expire, by using the keep-alive frequency 389 discussed in Section 3.5 of [RFC8085]. 391 To support scenarios in which loss of heartbeat is used to trigger 392 mitigation, and to keep the channel active, DOTS servers MUST 393 solicit heartbeat exchanges after successful mutual 394 authentication. When DOTS agents are exchanging heartbeats and no 395 mitigation request is active, either agent MAY request changes to 396 the heartbeat rate. For example, a DOTS server might want to 397 reduce heartbeat frequency or cease heartbeat exchanges when an 398 active DOTS client has not requested mitigation, in order to 399 control load. 401 Following mutual authentication, a signal channel MUST be 402 considered active until a DOTS agent explicitly ends the session. 403 When no attack traffic is present, the signal channel MUST be 404 considered active until either DOTS agent fails to receive 405 heartbeats from the other peer after a mutually agreed upon 406 retransmission procedure has been exhausted. Peer DOTS agents 407 MUST regularly send heartbeats to each other while a mitigation 408 request is active. Because heartbeat loss is much more likely 409 during volumetric attack, DOTS agents SHOULD avoid signal channel 410 termination when mitigation is active and heartbeats are not 411 received by either DOTS agent for an extended period. The 412 exception circumstances to terminate the signal channel session 413 during active mitigation are discussed below: 415 * To handle possible DOTS server restart or crash, the DOTS 416 clients MAY attempt to establish a new signal channel session, 417 but MUST continue to send heartbeats on the current session so 418 that the DOTS server knows the session is still alive. If the 419 new session is successfully established, the DOTS client can 420 terminate the current session. 422 * DOTS servers are assumed to have the ability to monitor the 423 attack, using feedback from the mitigator and other available 424 sources, and MAY use the absence of attack traffic and lack of 425 client heartbeats as an indication the signal channel is 426 defunct. 428 SIG-005 Channel Redirection: In order to increase DOTS operational 429 flexibility and scalability, DOTS servers SHOULD be able to 430 redirect DOTS clients to another DOTS server at any time. DOTS 431 clients MUST NOT assume the redirection target DOTS server shares 432 security state with the redirecting DOTS server. DOTS clients are 433 free to attempt abbreviated security negotiation methods supported 434 by the protocol, such as DTLS session resumption, but MUST be 435 prepared to negotiate new security state with the redirection 436 target DOTS server. The redirection DOTS server and redirecting 437 DOTS server MUST belong to the same administrative domain. 439 Due to the increased likelihood of packet loss caused by link 440 congestion during an attack, DOTS servers SHOULD NOT redirect 441 while mitigation is enabled during an active attack against a 442 target in the DOTS client's domain. 444 SIG-006 Mitigation Requests and Status: Authorized DOTS clients MUST 445 be able to request scoped mitigation from DOTS servers. DOTS 446 servers MUST send status to the DOTS clients about mitigation 447 requests. If a DOTS server rejects an authorized request for 448 mitigation, the DOTS server MUST include a reason for the 449 rejection in the status message sent to the client. 451 DOTS servers MUST regularly send mitigation status updates to 452 authorized DOTS clients which have requested and been granted 453 mitigation. If unreliable transport is used for the signal 454 channel protocol, due to the higher likelihood of packet loss 455 during a DDoS attack, DOTS servers needs to send mitigation status 456 multiple times at regular intervals following the the data 457 transmission guidelines discussed in Section 3.1.3 of [RFC8085]. 459 When DOTS client-requested mitigation is active, DOTS server 460 status messages MUST include the following mitigation metrics: 462 * Total number of packets blocked by the mitigation 464 * Current number of packets per second blocked 466 * Total number of bytes blocked 468 * Current number of bytes per second blocked 470 DOTS clients MAY take these metrics into account when determining 471 whether to ask the DOTS server to cease mitigation. 473 A DOTS client MAY withdraw a mitigation request at any time, 474 regardless of whether mitigation is currently active. The DOTS 475 server MUST immediately acknowledge a DOTS client's request to 476 stop mitigation. 478 To protect against route or DNS flapping caused by a client 479 rapidly toggling mitigation, and to dampen the effect of 480 oscillating attacks, DOTS servers MAY allow mitigation to continue 481 for a limited period after acknowledging a DOTS client's 482 withdrawal of a mitigation request. During this period, DOTS 483 server status messages SHOULD indicate that mitigation is active 484 but terminating. DOTS clients MAY reverse the mitigation 485 termination during this active-but-terminating period with a new 486 mitigation request for the same scope. The DOTS server MUST treat 487 this request as a mitigation lifetime extension (see SIG-007 488 below). 490 The initial active-but-terminating period is implementation- and 491 deployment- specific, but SHOULD be sufficiently long to absorb 492 latency incurred by route propagation. If a DOTS client refreshes 493 the mitigation before the active-but-terminating period elapses, 494 the DOTS server MAY increase the active-but-terminating period up 495 to a maximum of 300 seconds (5 minutes). After the active-but- 496 terminating period elapses, the DOTS server MUST treat the 497 mitigation as terminated, as the DOTS client is no longer 498 responsible for the mitigation. 500 SIG-007 Mitigation Lifetime: DOTS servers MUST support mitigations 501 for a negotiated time interval, and MUST terminate a mitigation 502 when the lifetime elapses. DOTS servers also MUST support renewal 503 of mitigation lifetimes in mitigation requests from DOTS clients, 504 allowing clients to extend mitigation as necessary for the 505 duration of an attack. 507 DOTS servers MUST treat a mitigation terminated due to lifetime 508 expiration exactly as if the DOTS client originating the 509 mitigation had asked to end the mitigation, including the active- 510 but-terminating period, as described above in SIG-005. 512 DOTS clients MUST include a mitigation lifetime in all mitigation 513 requests. 515 DOTS servers SHOULD support indefinite mitigation lifetimes, 516 enabling architectures in which the mitigator is always in the 517 traffic path to the resources for which the DOTS client is 518 requesting protection. DOTS clients MUST be prepared to not be 519 granted mitigations with indefinite lifetimes. DOTS servers MAY 520 refuse mitigations with indefinite lifetimes, for policy reasons. 522 The reasons themselves are out of scope. If the DOTS server does 523 not grant a mitigation request with an indefinite mitigation 524 lifetime, it MUST set the lifetime to a value that is configured 525 locally. That value MUST be returned in a reply to the requesting 526 DOTS client. 528 SIG-008 Mitigation Scope: DOTS clients MUST indicate desired 529 mitigation scope. The scope type will vary depending on the 530 resources requiring mitigation. All DOTS agent implementations 531 MUST support the following required scope types: 533 * IPv4 prefixes [RFC4632] 535 * IPv6 prefixes [RFC4291][RFC5952] 537 * Domain names [RFC1035] 539 The following mitigation scope types are OPTIONAL: 541 * Uniform Resource Identifiers [RFC3986] 543 DOTS servers MUST be able to resolve domain names and (when 544 supported) URIs. How name resolution is managed on the DOTS 545 server is implementation-specific. 547 DOTS agents MUST support mitigation scope aliases, allowing DOTS 548 clients and servers to refer to collections of protected resources 549 by an opaque identifier created through the data channel, direct 550 configuration, or other means. Domain name and URI mitigation 551 scopes may be thought of as a form of scope alias, in which the 552 addresses to which the domain name or URI resolve represent the 553 full scope of the mitigation. 555 If there is additional information available narrowing the scope 556 of any requested attack response, such as targeted port range, 557 protocol, or service, DOTS clients SHOULD include that information 558 in client mitigation requests. DOTS clients MAY also include 559 additional attack details. DOTS servers MAY ignore such 560 supplemental information when enabling countermeasures on the 561 mitigator. 563 As an active attack evolves, DOTS clients MUST be able to adjust 564 as necessary the scope of requested mitigation by refining the 565 scope of resources requiring mitigation. 567 A DOTS client may obtain the mitigation scope through direct 568 provisioning or through implementation-specific methods of 569 discovery. DOTS clients MUST support at least one mechanism to 570 obtain mitigation scope. 572 SIG-009 Mitigation Efficacy: When a mitigation request is active, 573 DOTS clients MUST be able to transmit a metric of perceived 574 mitigation efficacy to the DOTS server. DOTS servers MAY use the 575 efficacy metric to adjust countermeasures activated on a mitigator 576 on behalf of a DOTS client. 578 SIG-010 Conflict Detection and Notification: Multiple DOTS clients 579 controlled by a single administrative entity may send conflicting 580 mitigation requests as a result of misconfiguration, operator 581 error, or compromised DOTS clients. DOTS servers in the same 582 administrative domain attempting to honor conflicting requests may 583 flap network route or DNS information, degrading the networks 584 attempting to participate in attack response with the DOTS 585 clients. DOTS servers in a single administrative domain SHALL 586 detect such conflicting requests, and SHALL notify the DOTS 587 clients in conflict. The notification MUST indicate the nature 588 and scope of the conflict, for example, the overlapping prefix 589 range in a conflicting mitigation request. 591 SIG-011 Network Address Translator Traversal: DOTS clients may be 592 deployed behind a Network Address Translator (NAT), and need to 593 communicate with DOTS servers through the NAT. DOTS protocols 594 MUST therefore be capable of traversing NATs. 596 If UDP is used as the transport for the DOTS signal channel, all 597 considerations in "Middlebox Traversal Guidelines" in [RFC8085] 598 apply to DOTS. Regardless of transport, DOTS protocols MUST 599 follow established best common practices established in BCP 127 600 for NAT traversal [RFC4787][RFC6888][RFC7857]. 602 2.3. Data Channel Requirements 604 The data channel is intended to be used for bulk data exchanges 605 between DOTS agents. Unlike the signal channel, the data channel is 606 not expected to be constructed to deal with attack conditions. As 607 the primary function of the data channel is data exchange, a reliable 608 transport is required in order for DOTS agents to detect data 609 delivery success or failure. 611 The data channel provides a protocol for DOTS configuration, 612 management. For example, a DOTS client may submit to a DOTS server a 613 collection of prefixes it wants to refer to by alias when requesting 614 mitigation, to which the server would respond with a success status 615 and the new prefix group alias, or an error status and message in the 616 event the DOTS client's data channel request failed. 618 DATA-001 Reliable transport: Messages sent over the data channel 619 MUST be delivered reliably, in order sent. 621 DATA-003 Resource Configuration: To help meet the general and signal 622 channel requirements in Section 2.1 and Section 2.2, DOTS server 623 implementations MUST provide an interface to configure resource 624 identifiers, as described in SIG-008. DOTS server implementations 625 MAY expose additional configurability. Additional configurability 626 is implementation-specific. 628 DATA-004 Policy management: DOTS servers MUST provide methods for 629 DOTS clients to manage drop- and accept-lists of traffic destined 630 for resources belonging to a client. 632 For example, a DOTS client should be able to create a drop- or 633 accept-list entry, retrieve a list of current entries from either 634 list, update the content of either list, and delete entries as 635 necessary. 637 How a DOTS server authorizes DOTS client management of drop- and 638 accept-list entries is implementation-specific. 640 2.4. Security Requirements 642 DOTS must operate within a particularly strict security context, as 643 an insufficiently protected signal or data channel may be subject to 644 abuse, enabling or supplementing the very attacks DOTS purports to 645 mitigate. 647 SEC-001 Peer Mutual Authentication: DOTS agents MUST authenticate 648 each other before a DOTS signal or data channel is considered 649 valid. The method of authentication is not specified in this 650 document, but should follow current IETF best practices [RFC7525] 651 with respect to any cryptographic mechanisms to authenticate the 652 remote peer. 654 SEC-002 Message Confidentiality, Integrity and Authenticity: DOTS 655 protocols MUST take steps to protect the confidentiality, 656 integrity and authenticity of messages sent between client and 657 server. While specific transport- and message-level security 658 options are not specified, the protocols MUST follow current IETF 659 best practices [RFC7525] for encryption and message 660 authentication. Client-domain DOTS gateways are more trusted than 661 DOTS clients, while server-domain DOTS gateways and DOTS servers 662 share the same level of trust. A security mechanism at the 663 transport layer TLS/DTLS is thus adequate to provide security 664 between peer DOTS agents. 666 In order for DOTS protocols to remain secure despite advancements 667 in cryptanalysis and traffic analysis, DOTS agents MUST support 668 secure negotiation of the terms and mechanisms of protocol 669 security, subject to the interoperability and signal message size 670 requirements in Section 2.2. 672 While the interfaces between downstream DOTS server and upstream 673 DOTS client within a DOTS gateway are implementation-specific, 674 those interfaces nevertheless MUST provide security equivalent to 675 that of the signal channels bridged by gateways in the signaling 676 path. For example, when a DOTS gateway consisting of a DOTS 677 server and DOTS client is running on the same logical device, the 678 two DOTS agents could be implemented within the same process 679 security boundary. 681 SEC-003 Data privacy and integrity: Transmissions over the DOTS 682 protocols are likely to contain operationally or privacy-sensitive 683 information or instructions from the remote DOTS agent. Theft, 684 modification, or replay of message transmissions could lead to 685 information leaks or malicious transactions on behalf of the 686 sending agent (see Section 4 below). Consequently data sent over 687 the DOTS protocols MUST be encrypted using secure transports TLS/ 688 DTLS. DOTS servers MUST enable means to prevent leaking 689 operationally or privacy-sensitive data. Although administrative 690 entities participating in DOTS may detail what data may be 691 revealed to third-party DOTS agents, such considerations are not 692 in scope for this document. 694 SEC-004 Message Replay Protection: To prevent a passive attacker 695 from capturing and replaying old messages, and thereby potentially 696 disrupting or influencing the network policy of the receiving DOTS 697 agent's domain, DOTS protocols MUST provide a method for replay 698 detection and prevention. 700 Within the signal channel, messages MUST be uniquely identified 701 such that replayed or duplicated messages can be detected and 702 discarded. Unique mitigation requests MUST be processed at most 703 once. 705 SEC-005 Authorization: DOTS servers MUST authorize all messages from 706 DOTS clients which pertain to mitigation, configuration, 707 filtering, or status. 709 DOTS servers MUST reject mitigation requests with scopes which the 710 DOTS client is not authorized to manage. 712 Likewise, DOTS servers MUST refuse to allow creation, modification 713 or deletion of scope aliases and drop-/accept-lists when the DOTS 714 client is unauthorized. 716 The modes of authorization are implementation-specific. 718 2.5. Data Model Requirements 720 A well-structured DOTS data model is critical to the development of 721 successful DOTS protocols. 723 DM-001 Structure: The data model structure for the DOTS protocol MAY 724 be described by a single module, or be divided into related 725 collections of hierarchical modules and sub-modules. If the data 726 model structure is split across modules, those distinct modules 727 MUST allow references to describe the overall data model's 728 structural dependencies. 730 DM-002 Versioning: To ensure interoperability between DOTS protocol 731 implementations, data models MUST be versioned. How the protocols 732 represent data model versions is not defined in this document. 734 DM-003 Mitigation Status Representation: The data model MUST provide 735 the ability to represent a request for mitigation and the 736 withdrawal of such a request. The data model MUST also support a 737 representation of currently requested mitigation status, including 738 failures and their causes. 740 DM-004 Mitigation Scope Representation: The data model MUST support 741 representation of a requested mitigation's scope. As mitigation 742 scope may be represented in several different ways, per SIG-008 743 above, the data model MUST include extensible representation of 744 mitigation scope. 746 DM-005 Mitigation Lifetime Representation: The data model MUST 747 support representation of a mitigation request's lifetime, 748 including mitigations with no specified end time. 750 DM-006 Mitigation Efficacy Representation: The data model MUST 751 support representation of a DOTS client's understanding of the 752 efficacy of a mitigation enabled through a mitigation request. 754 DM-007 Acceptable Signal Loss Representation: The data model MUST be 755 able to represent the DOTS agent's preference for acceptable 756 signal loss when establishing a signal channel. Measurements of 757 loss might include, but are not restricted to, number of 758 consecutive missed heartbeat messages, retransmission count, or 759 request timeouts. 761 DM-008 Heartbeat Interval Representation: The data model MUST be 762 able to represent the DOTS agent's preferred heartbeat interval, 763 which the client may include when establishing the signal channel, 764 as described in SIG-003. 766 DM-009 Relationship to Transport: The DOTS data model MUST NOT make 767 any assumptions about specific characteristics of any given 768 transport into the data model, but instead, represent the fields 769 in the model explicitly. 771 3. Congestion Control Considerations 773 3.1. Signal Channel 775 As part of a protocol expected to operate over links affected by DDoS 776 attack traffic, the DOTS signal channel MUST NOT contribute 777 significantly to link congestion. To meet the signal channel 778 requirements above, DOTS signal channel implementations SHOULD 779 support connectionless transports. However, some connectionless 780 transports when deployed naively can be a source of network 781 congestion, as discussed in [RFC8085]. Signal channel 782 implementations using such connectionless transports, such as UDP, 783 therefore MUST include a congestion control mechanism. 785 Signal channel implementations using a IETF standard congestion- 786 controlled transport protocol (like TCP) may rely on built-in 787 transport congestion control support. 789 3.2. Data Channel 791 As specified in DATA-001, the data channel requires reliable, in- 792 order message delivery. Data channel implementations using a IETF 793 standard congestion-controlled transport protocol may rely on the 794 transport implementation's built-in congestion control mechanisms. 796 4. Security Considerations 798 This document informs future protocols under development, and so does 799 not have security considerations of its own. However, operators 800 should be aware of potential risks involved in deploying DOTS. DOTS 801 agent impersonation and signal blocking are discussed here. 802 Additional DOTS security considerations may be found in 803 [I-D.ietf-dots-architecture] and DOTS protocol documents. 805 Impersonation of either a DOTS server or a DOTS client could have 806 catastrophic impact on operations in either domain. If an attacker 807 has the ability to impersonate a DOTS client, that attacker can 808 affect policy on the network path to the DOTS client's domain, up to 809 and including instantiation of drop-lists blocking all inbound 810 traffic to networks for which the DOTS client is authorized to 811 request mitigation. 813 Similarly, an impersonated DOTS server may be able to act as a sort 814 of malicious DOTS gateway, intercepting requests from the downstream 815 DOTS client, and modifying them before transmission to the DOTS 816 server to inflict the desired impact on traffic to or from the DOTS 817 client's domain. Among other things, this malicious DOTS gateway 818 might receive and discard mitigation requests from the DOTS client, 819 ensuring no requested mitigation is ever applied. 821 To detect misuse, as detailed in Section 2.4, DOTS implementations 822 require mutual authentication of DOTS agents in order to make agent 823 impersonation more difficult. However, impersonation may still be 824 possible as a result of credential theft, implementation flaws, or 825 compromise of DOTS agents. 827 To detect compromised DOTS agents, DOTS operators should carefully 828 monitor and audit DOTS agents to detect misbehavior and to deter 829 misuse, while employing current secure network communications best 830 practices to reduce attack surface. 832 Blocking communication between DOTS agents has the potential to 833 disrupt the core function of DOTS, which is to request mitigation of 834 active or expected DDoS attacks. The DOTS signal channel is expected 835 to operate over congested inbound links, and, as described in 836 Section 2.2, the signal channel protocol must be designed for minimal 837 data transfer to reduce the incidence of signal loss. 839 5. IANA Considerations 841 This document does not require any IANA action. 843 6. Contributors 845 Mohamed Boucadair 846 Orange 848 mohamed.boucadair@orange.com 850 Flemming Andreasen 851 Cisco Systems, Inc. 853 fandreas@cisco.com 855 Dave Dolson 856 Sandvine 857 ddolson@sandvine.com 859 7. Acknowledgments 861 Thanks to Roman Danyliw, Matt Richardson, Joe Touch, Scott Bradner, 862 Robert Sparks, Brian Weis, Benjamin Kaduk, Eric Rescorla, Alvaro 863 Retana, Suresh Krishnan, Ben Campbell, Mirja Kuehlewind and Jon 864 Shallow for careful reading and feedback. 866 8. References 868 8.1. Normative References 870 [RFC0768] Postel, J., "User Datagram Protocol", STD 6, RFC 768, 871 DOI 10.17487/RFC0768, August 1980, 872 . 874 [RFC0791] Postel, J., "Internet Protocol", STD 5, RFC 791, 875 DOI 10.17487/RFC0791, September 1981, 876 . 878 [RFC0793] Postel, J., "Transmission Control Protocol", STD 7, 879 RFC 793, DOI 10.17487/RFC0793, September 1981, 880 . 882 [RFC1035] Mockapetris, P., "Domain names - implementation and 883 specification", STD 13, RFC 1035, DOI 10.17487/RFC1035, 884 November 1987, . 886 [RFC1122] Braden, R., Ed., "Requirements for Internet Hosts - 887 Communication Layers", STD 3, RFC 1122, 888 DOI 10.17487/RFC1122, October 1989, 889 . 891 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 892 Requirement Levels", BCP 14, RFC 2119, 893 DOI 10.17487/RFC2119, March 1997, 894 . 896 [RFC3986] Berners-Lee, T., Fielding, R., and L. Masinter, "Uniform 897 Resource Identifier (URI): Generic Syntax", STD 66, 898 RFC 3986, DOI 10.17487/RFC3986, January 2005, 899 . 901 [RFC4291] Hinden, R. and S. Deering, "IP Version 6 Addressing 902 Architecture", RFC 4291, DOI 10.17487/RFC4291, February 903 2006, . 905 [RFC4632] Fuller, V. and T. Li, "Classless Inter-domain Routing 906 (CIDR): The Internet Address Assignment and Aggregation 907 Plan", BCP 122, RFC 4632, DOI 10.17487/RFC4632, August 908 2006, . 910 [RFC4787] Audet, F., Ed. and C. Jennings, "Network Address 911 Translation (NAT) Behavioral Requirements for Unicast 912 UDP", BCP 127, RFC 4787, DOI 10.17487/RFC4787, January 913 2007, . 915 [RFC5952] Kawamura, S. and M. Kawashima, "A Recommendation for IPv6 916 Address Text Representation", RFC 5952, 917 DOI 10.17487/RFC5952, August 2010, 918 . 920 [RFC6888] Perreault, S., Ed., Yamagata, I., Miyakawa, S., Nakagawa, 921 A., and H. Ashida, "Common Requirements for Carrier-Grade 922 NATs (CGNs)", BCP 127, RFC 6888, DOI 10.17487/RFC6888, 923 April 2013, . 925 [RFC7857] Penno, R., Perreault, S., Boucadair, M., Ed., Sivakumar, 926 S., and K. Naito, "Updates to Network Address Translation 927 (NAT) Behavioral Requirements", BCP 127, RFC 7857, 928 DOI 10.17487/RFC7857, April 2016, 929 . 931 [RFC8085] Eggert, L., Fairhurst, G., and G. Shepherd, "UDP Usage 932 Guidelines", BCP 145, RFC 8085, DOI 10.17487/RFC8085, 933 March 2017, . 935 [RFC8174] Leiba, B., "Ambiguity of Uppercase vs Lowercase in RFC 936 2119 Key Words", BCP 14, RFC 8174, DOI 10.17487/RFC8174, 937 May 2017, . 939 [RFC8200] Deering, S. and R. Hinden, "Internet Protocol, Version 6 940 (IPv6) Specification", STD 86, RFC 8200, 941 DOI 10.17487/RFC8200, July 2017, 942 . 944 8.2. Informative References 946 [I-D.ietf-dots-architecture] 947 Mortensen, A., Andreasen, F., K, R., Teague, N., Compton, 948 R., and c. christopher_gray3@cable.comcast.com, 949 "Distributed-Denial-of-Service Open Threat Signaling 950 (DOTS) Architecture", draft-ietf-dots-architecture-12 951 (work in progress), February 2019. 953 [I-D.ietf-dots-use-cases] 954 Dobbins, R., Migault, D., Fouant, S., Moskowitz, R., 955 Teague, N., Xia, L., and K. Nishizuka, "Use cases for DDoS 956 Open Threat Signaling", draft-ietf-dots-use-cases-17 (work 957 in progress), January 2019. 959 [I-D.ietf-intarea-frag-fragile] 960 Bonica, R., Baker, F., Huston, G., Hinden, R., Troan, O., 961 and F. Gont, "IP Fragmentation Considered Fragile", draft- 962 ietf-intarea-frag-fragile-09 (work in progress), February 963 2019. 965 [RFC3261] Rosenberg, J., Schulzrinne, H., Camarillo, G., Johnston, 966 A., Peterson, J., Sparks, R., Handley, M., and E. 967 Schooler, "SIP: Session Initiation Protocol", RFC 3261, 968 DOI 10.17487/RFC3261, June 2002, 969 . 971 [RFC7092] Kaplan, H. and V. Pascual, "A Taxonomy of Session 972 Initiation Protocol (SIP) Back-to-Back User Agents", 973 RFC 7092, DOI 10.17487/RFC7092, December 2013, 974 . 976 [RFC4732] Handley, M., Ed., Rescorla, E., Ed., and IAB, "Internet 977 Denial-of-Service Considerations", RFC 4732, 978 DOI 10.17487/RFC4732, December 2006, 979 . 981 [RFC4949] Shirey, R., "Internet Security Glossary, Version 2", 982 FYI 36, RFC 4949, DOI 10.17487/RFC4949, August 2007, 983 . 985 [RFC7525] Sheffer, Y., Holz, R., and P. Saint-Andre, 986 "Recommendations for Secure Use of Transport Layer 987 Security (TLS) and Datagram Transport Layer Security 988 (DTLS)", BCP 195, RFC 7525, DOI 10.17487/RFC7525, May 989 2015, . 991 Authors' Addresses 993 Andrew Mortensen 994 Arbor Networks 995 2727 S. State St 996 Ann Arbor, MI 48104 997 United States 999 Email: andrewmortensen@gmail.com 1000 Tirumaleswar Reddy 1001 McAfee 1002 Embassy Golf Link Business Park 1003 Bangalore, Karnataka 560071 1004 India 1006 Email: TirumaleswarReddy_Konda@McAfee.com 1008 Robert Moskowitz 1009 Huawei 1010 Oak Park, MI 42837 1011 United States 1013 Email: rgm@htt-consult.com