idnits 2.17.1 draft-ietf-anima-autonomic-control-plane-00.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 : ---------------------------------------------------------------------------- ** The document seems to lack separate sections for Informative/Normative References. All references will be assumed normative when checking for downward references. ** The document seems to lack a both a reference to RFC 2119 and the recommended RFC 2119 boilerplate, even if it appears to use RFC 2119 keywords. RFC 2119 keyword, line 231: '... 1. The ACP SHOULD provide robust c...' RFC 2119 keyword, line 236: '... 2. The ACP MUST have a separate ad...' RFC 2119 keyword, line 240: '... 3. The ACP MUST use autonomically ...' RFC 2119 keyword, line 245: '... 4. The ACP MUST be generic. Usabl...' RFC 2119 keyword, line 246: '...infrastructure. MUST NOT be tied to a...' (4 more instances...) Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year -- The document date (August 17, 2015) is 3146 days in the past. Is this intentional? Checking references for intended status: Proposed Standard ---------------------------------------------------------------------------- (See RFCs 3967 and 4897 for information about using normative references to lower-maturity documents in RFCs) == Outdated reference: A later version (-04) exists of draft-behringer-anima-reference-model-03 ** Downref: Normative reference to an Informational draft: draft-behringer-anima-reference-model (ref. 'I-D.behringer-anima-reference-model') ** Downref: Normative reference to an Informational draft: draft-behringer-autonomic-control-plane (ref. 'I-D.behringer-autonomic-control-plane') == Outdated reference: A later version (-02) exists of draft-eckert-anima-stable-connectivity-01 ** Downref: Normative reference to an Informational draft: draft-eckert-anima-stable-connectivity (ref. 'I-D.eckert-anima-stable-connectivity') ** Downref: Normative reference to an Informational draft: draft-pritikin-anima-bootstrapping-keyinfra (ref. 'I-D.pritikin-anima-bootstrapping-keyinfra') ** Downref: Normative reference to an Informational RFC: RFC 7575 ** Downref: Normative reference to an Informational RFC: RFC 7576 Summary: 8 errors (**), 0 flaws (~~), 3 warnings (==), 1 comment (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 ANIMA WG M. Behringer, Ed. 3 Internet-Draft S. Bjarnason 4 Intended status: Standards Track Balaji. BL 5 Expires: February 18, 2016 T. Eckert 6 Cisco Systems 7 August 17, 2015 9 An Autonomic Control Plane 10 draft-ietf-anima-autonomic-control-plane-00 12 Abstract 14 Autonomic functions need a control plane to communicate, which 15 depends on some addressing and routing. This Autonomic Control Plane 16 should ideally be self-managing, and as independent as possible of 17 configuration. One application is a "virtual out of band channel" 18 for communications over a network that is not configured or mis- 19 configured. This document describes requirements and implementation 20 options for an "Autonomic Control Plane". 22 Status of This Memo 24 This Internet-Draft is submitted in full conformance with the 25 provisions of BCP 78 and BCP 79. 27 Internet-Drafts are working documents of the Internet Engineering 28 Task Force (IETF). Note that other groups may also distribute 29 working documents as Internet-Drafts. The list of current Internet- 30 Drafts is at http://datatracker.ietf.org/drafts/current/. 32 Internet-Drafts are draft documents valid for a maximum of six months 33 and may be updated, replaced, or obsoleted by other documents at any 34 time. It is inappropriate to use Internet-Drafts as reference 35 material or to cite them other than as "work in progress." 37 This Internet-Draft will expire on February 18, 2016. 39 Copyright Notice 41 Copyright (c) 2015 IETF Trust and the persons identified as the 42 document authors. All rights reserved. 44 This document is subject to BCP 78 and the IETF Trust's Legal 45 Provisions Relating to IETF Documents 46 (http://trustee.ietf.org/license-info) in effect on the date of 47 publication of this document. Please review these documents 48 carefully, as they describe your rights and restrictions with respect 49 to this document. Code Components extracted from this document must 50 include Simplified BSD License text as described in Section 4.e of 51 the Trust Legal Provisions and are provided without warranty as 52 described in the Simplified BSD License. 54 Table of Contents 56 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 2 57 2. Use Cases for an Autonomic Control Plane . . . . . . . . . . 4 58 2.1. An Infrastructure for Autonomic Functions . . . . . . . . 4 59 2.2. Secure Bootstrap over an Unconfigured Network . . . . . . 4 60 2.3. Data Plane Independent Permanent Reachability . . . . . . 4 61 3. Requirements . . . . . . . . . . . . . . . . . . . . . . . . 5 62 4. Overview . . . . . . . . . . . . . . . . . . . . . . . . . . 6 63 5. Self-Creation of an Autonomic Control Plane . . . . . . . . . 7 64 5.1. Preconditions . . . . . . . . . . . . . . . . . . . . . . 7 65 5.2. Adjacency Discovery . . . . . . . . . . . . . . . . . . . 8 66 5.3. Authenticating Neighbors . . . . . . . . . . . . . . . . 8 67 5.4. Capability Negotiation . . . . . . . . . . . . . . . . . 9 68 5.5. Channel Establishment . . . . . . . . . . . . . . . . . . 9 69 5.6. Context Separation . . . . . . . . . . . . . . . . . . . 10 70 5.7. Addressing inside the ACP . . . . . . . . . . . . . . . . 10 71 5.8. Routing in the ACP . . . . . . . . . . . . . . . . . . . 11 72 5.9. Connecting a Controller / NMS system . . . . . . . . . . 11 73 6. Self-Healing Properties . . . . . . . . . . . . . . . . . . . 12 74 7. Self-Protection Properties . . . . . . . . . . . . . . . . . 13 75 8. The Administrator View . . . . . . . . . . . . . . . . . . . 14 76 9. Security Considerations . . . . . . . . . . . . . . . . . . . 14 77 10. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 15 78 11. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . 15 79 12. Change log [RFC Editor: Please remove] . . . . . . . . . . . 15 80 12.1. Initial version . . . . . . . . . . . . . . . . . . . . 15 81 12.2. draft-behringer-anima-autonomic-control-plane-00 . . . . 15 82 12.3. draft-behringer-anima-autonomic-control-plane-01 . . . . 15 83 12.4. draft-behringer-anima-autonomic-control-plane-02 . . . . 16 84 12.5. draft-behringer-anima-autonomic-control-plane-03 . . . . 16 85 12.6. draft-ietf-anima-autonomic-control-plane-00 . . . . . . 16 86 13. References . . . . . . . . . . . . . . . . . . . . . . . . . 16 87 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 17 89 1. Introduction 91 Autonomic Networking is a concept of self-management: Autonomic 92 functions self-configure, and negotiate parameters and settings 93 across the network. [RFC7575] defines the fundamental ideas and 94 design goals of Autonomic Networking. A gap analysis of Autonomic 95 Networking is given in [RFC7576]. The reference architecture for 96 Autonomic Networking in the IETF is currently being defined in the 97 document [I-D.behringer-anima-reference-model] 99 Autonomic functions need a stable and robust infrastructure to 100 communicate on. This infrastructure should be as robust as possible, 101 and it should be re-usable by all autonomic functions. [RFC7575] 102 calls it the "Autonomic Control Plane". This document defines the 103 requirements and implementation options of an Autonomic Control 104 Plane. 106 Today, the management and control plane of networks typically runs in 107 the global routing table, which is dependent on correct configuration 108 and routing. Misconfigurations or routing problems can therefore 109 disrupt management and control channels. Traditionally, an out of 110 band network has been used to recover from such problems, or 111 personnel is sent on site to access devices through console ports. 112 However, both options are operationally expensive. 114 In increasingly automated networks either controllers or distributed 115 autonomic service agents in the network require a control plane which 116 is independent of the network they manage, to avoid impacting their 117 own operations. 119 This document describes options for a self-forming, self-managing and 120 self-protecting "Autonomic Control Plane" (ACP) which is inband on 121 the network, yet as independent as possible of configuration, 122 addressing and routing problems (for details how this achieved, see 123 Section 5). It therefore remains operational even in the presence of 124 configuration errors, addressing or routing issues, or where policy 125 could inadvertently affect control plane connectivity. The Autonomic 126 Control Plane serves several purposes at the same time: 128 o Autonomic functions communicate over the ACP. 130 o An operator can use it to log into remote devices, even if the 131 data plane is misconfigured or unconfigured. 133 o A controller or network management system can use it to securely 134 bootstrap network devices in remote locations, even if the network 135 in between is not yet configured; no data-plane dependent 136 bootstrap configuration is required. An example of such a secure 137 bootstrap process is described in 138 [I-D.pritikin-anima-bootstrapping-keyinfra] 140 o Devices can use the ACP for direct decentralised communications, 141 such as negotiations or discovery. The ACP therefore supports 142 directly Autonomic Networking functions, as described in 144 [I-D.behringer-anima-reference-model]. For example, GDNP 145 [I-D.carpenter-anima-gdn-protocol] can run inside the ACP. 147 This document describes some use cases for the ACP in Section 2, it 148 defines the requirements in Section 3, Section 4 gives an overview 149 how an Autonomic Control Plane is constructed, and in Section 5 the 150 detailed process is explained. The document "Autonomic Network 151 Stable Connectivity" [I-D.eckert-anima-stable-connectivity] describes 152 how the ACP can be used to provide stable connectivity for OAM 153 applications. It also explains on how existing management solutions 154 can leverage the ACP in parallel with traditional management models, 155 when to use the ACP versus the data plane, how to integrate IPv4 156 based management, etc. 158 2. Use Cases for an Autonomic Control Plane 160 2.1. An Infrastructure for Autonomic Functions 162 Autonomic Functions need a stable infrastructure to run on, and all 163 autonomic functions should use the same infrastructure to minimise 164 the complexity of the network. This way, there is only need for a 165 single discovery mechanism, a single security mechanism, and other 166 process that distributed functions require. 168 2.2. Secure Bootstrap over an Unconfigured Network 170 Today, bootstrapping a new device typically requires all devices 171 between a controlling node (such as an SDN controller) and the new 172 device to be completely and correctly addressed, configured and 173 secured. Therefore, bootstrapping a network happens in layers around 174 the controller. Without console access (for example through an out 175 of band network) it is not possible today to make devices securely 176 reachable before having configured the entire network between. 178 With the ACP, secure bootstrap of new devices can happen without 179 requiring any configuration on the network. A new device can 180 automatically be bootstrapped in a secure fashion and be deployed 181 with a domain certificate. This does not require any configuration 182 on intermediate nodes, because they can communicate through the ACP. 184 2.3. Data Plane Independent Permanent Reachability 186 Today, most critical control plane protocols and network management 187 protocols are running in the data plane (global routing table) of the 188 network. This leads to undesirable dependencies between control and 189 management plane on one side and the data plane on the other: Only if 190 the data plane is operational, will the other planes work as 191 expected. 193 Data plane connectivity can be affected by errors and faults, for 194 example certain AAA misconfigurations can lock an administrator out 195 of a device; routing or addressing issues can make a device 196 unreachable; shutting down interfaces over which a current management 197 session is running can lock an admin irreversibly out of the device. 198 Traditionally only console access can help recover from such issues. 200 Data plane dependencies also affect NOC/SDN controller applications: 201 Certain network changes are today hard to operate, because the change 202 itself may affect reachability of the devices. Examples are address 203 or mask changes, routing changes, or security policies. Today such 204 changes require precise hop-by-hop planning. 206 The ACP provides reachability that is largely independent of the data 207 plane, which allows control plane and management plane to operate 208 more robustly: 210 o For management plane protocols, the ACP provides the functionality 211 of a "Virtual-out-of-band (VooB) channel", by providing 212 connectivity to all devices regardless of their configuration or 213 global routing table. 215 o For control plane protocols, the ACP allows their operation even 216 when the data plane is temporarily faulty, or during transitional 217 events, such as routing changes, which may affect the control 218 plane at least temporarily. This is specifically important for 219 autonomic service agents, which could affect data plane 220 connectivity. 222 The document "Autonomic Network Stable Connectivity" 223 [I-D.eckert-anima-stable-connectivity] explains the use cases for the 224 ACP in significantly more detail and explains how the ACP can be used 225 in practical network operations. 227 3. Requirements 229 The Autonomic Control Plane has the following requirements: 231 1. The ACP SHOULD provide robust connectivity: As far as possible, 232 it should be independent of configured addressing, configuration 233 and routing. (2 and 3 build on this requirement, but also have 234 value on their own) 236 2. The ACP MUST have a separate address space from the data plane. 237 Reason: traceability, debug-ability, separation from data plane, 238 security (can block at edge) 240 3. The ACP MUST use autonomically managed address space. Reason: 241 easy bootstrap and setup ("autonomic"); robustness (admin can't 242 mess things up so easily). ULA seems like a good choice for 1 243 and 2. 245 4. The ACP MUST be generic. Usable by all the functions and 246 protocols of the AN infrastructure. MUST NOT be tied to a 247 particular protocol. 249 5. The ACP MUST provide security: Messages coming through the ACP 250 MUST be authenticated to be from a trusted node, and SHOULD (very 251 strong SHOULD) be encrypted. 253 The default mode of operation of the ACP is hop-by-hop, because this 254 interaction can be built on IPv6 link local addressing, which is 255 autonomic, and has no dependency on configuration (requirement 1). 256 It may be necessary to have end-to-end connectivity in some cases, 257 for example to provide an end-to-end security association for some 258 protocols. This is possible, but then has a dependency on routable 259 address space. 261 4. Overview 263 The Autonomic Control Plane is constructed in the following way (for 264 details, see Section 5): 266 o Each autonomic node creates a virtual routing and forwarding (VRF) 267 instance, or a similar virtual context. 269 o When an autonomic node discovers another autonomic node from the 270 same domain, it authenticates that node and negotiates a secure 271 tunnel to it. These tunnels are placed into the previously set up 272 VRF. This creates an overlay network with hop-by-hop tunnels. 274 o Inside the ACP VRF, each node sets up a loopback interface with a 275 ULA IPv6 address. 277 o Each node runs a lightweight routing protocol, to announce 278 reachability of the loopback addresses inside the ACP. 280 o NMS systems or controllers have to be manually connected into the 281 ACP. 283 o None of the above operations is reflected in the configuration of 284 the device. 286 The following figure illustrates the ACP. 288 autonomic node 1 autonomic node 2 289 ................... ................... 290 secure . . secure . . secure 291 tunnel : +-----------+ : tunnel : +-----------+ : tunnel 292 ..--------| ACP VRF |---------------------| ACP VRF |---------.. 293 : / \ / \ <--routing--> / \ / \ : 294 : \ / \ / \ / \ / : 295 ..--------| loopback |---------------------| loopback |---------.. 296 : +-----------+ : : +-----------+ : 297 : : : : 298 : data plane :...............: data plane : 299 : : link : : 300 :.................: :.................: 302 Figure 1 304 The resulting overlay network is normally based exclusively on hop- 305 by-hop tunnels. This is because addressing used on links is IPv6 306 link local addressing, which does not require any prior set-up. This 307 way the ACP can be built even if there is no configuration on the 308 devices, or if the data plane has issues such as addressing or 309 routing problems. 311 An alternative ACP design can be achieved without the VRFs. In this 312 case, the autonomic virtual addresses are part of the data plane, and 313 subject to routing, filtering, QoS, etc on the data plane. The 314 secure tunnels are in this case used by traffic to and from the 315 autonomic address space. They are still required to provide the 316 authentication function for all autonomic packets. 318 5. Self-Creation of an Autonomic Control Plane 320 This section describes the steps to set up an Autonomic Control 321 Plane, and highlights the key properties which make it 322 "indestructible" against many inadvert changes to the data plane, for 323 example caused by misconfigurations. 325 5.1. Preconditions 327 Each autonomic device has a globally unique domain certificate, with 328 which it can cryptographically assert its membership of the domain. 329 The document [I-D.pritikin-anima-bootstrapping-keyinfra] describes 330 how a domain certificate can be automatically and securely derived 331 from a vendor specific Unique Device Identifier (UDI) or IDevID 332 certificate. (Note the UDI used in this document is NOT the UUID 333 specified in [RFC4122].) 335 5.2. Adjacency Discovery 337 Adjacency discovery exchanges identity information about neighbors, 338 either the UDI or, if present, the domain certificate (see 339 Section 5.1. This document assumes the existence of a domain 340 certificate. 342 Adjacency discovery provides a table of information of adjacent 343 neighbors. Each neighbor is identified by a globally unique device 344 identifier (UDI). 346 The adjacency table contains the following information about the 347 adjacent neighbors. 349 o Globally valid Unique device identifier (UDI). 351 o Link Local IPv6 address with its scope. 353 o Trust information: The certificate chain, if available. 355 o Validity of the trust (once validated, see next section). 357 Adjacency discovery can populate this table by several means. One 358 such mechanism is to discover using link local multicast probes, 359 which has no dependency on configured addressing and is preferable in 360 an autonomic network. 362 The "Generic Discovery and Negotiation Protocol" GDNP described in 363 [I-D.carpenter-anima-gdn-protocol] is a possible candidate protocol 364 to meet the requirements for Adjacency Discovery described here. 366 5.3. Authenticating Neighbors 368 Each neighbor in the adjacency table is authenticated. The result of 369 the authentication of the neighbor information is stored in the 370 adjacency table. We distinguish the following cases: 372 o Inside the domain: If the domain certificate presented is 373 validated (including proof of possession of the corresponding 374 private key) to be in the same domain as that of the autonomic 375 entity then the neighbor is deemed to be inside the autonomic 376 domain. Only entities inside the autonomic domain will by default 377 be able to establish the autonomic control plane. Alternatively, 378 policy can define whether to simply trust devices with the same 379 trust anchor. An ACP channel will be established. 381 o Outside the domain: If there is no domain certificate presented by 382 the neighbor, or if the domain certificate presented is invalid or 383 expired, then the neighbor is deemed to be outside the autonomic 384 domain. No ACP channel will be established. 386 Certificate management questions such as enrolment, revocation, 387 renewal, etc, are not discussed in this draft. Please refer to 388 [I-D.pritikin-anima-bootstrapping-keyinfra] for more details. 390 5.4. Capability Negotiation 392 Autonomic devices have different capabilities based on the type of 393 device and where it is deployed. To establish a trusted secure 394 communication channel, devices must be able to negotiate with each 395 neighbor a set of parameters for establishing the communication 396 channel, most notably channel type and security type. the 397 communication channel, most notably channel type and security type. 398 The channel type could be any tunnel mechanism that is feasible 399 between two adjacent neighbors, for example a GRE tunnel. The 400 security type could be any of the channel protection mechanism that 401 is available between two adjacent neighbors on a given channel type, 402 for example TLS, DTLS or IPsec. The establishment of the autonomic 403 control plane can happen after the channel type and security type is 404 negotiated. 406 The "Generic Discovery and Negotiation Protocol GDNP described in 407 [I-D.carpenter-anima-gdn-protocol] is a possible candidate protocol 408 to meet the requirements for capability negotiation described here. 410 5.5. Channel Establishment 412 After authentication and capability negotiation autonomic nodes 413 establish a secure channel towards their direct AN neighbors with the 414 above negotiated parameters. In order to be independent of 415 configured link addresses, these channels can be implemented in 416 several ways: 418 o As a secure IP tunnel (e.g., IPsec, DTLS, TLS, etc.), using IPv6 419 link local addresses between two adjacent neighbors. This way, 420 the ACP tunnels are independent of correct network wide routing. 421 They also do not require larger than link local scope addresses, 422 which would normally need to be configured or maintained. Each AN 423 node MUST support this function. 425 o L2 separation, for example via a separate 802.1q tag for ACP 426 traffic. This even further reduces dependency against the data 427 plane (not even IPv6 link-local there required), but may be harder 428 to implement. 430 Since channels are established between adjacent neighbors, the 431 resulting overlay network does hop by hop encryption. Each node 432 decrypts incoming traffic from the ACP, and encrypts outgoing traffic 433 to its neighbors in the ACP. Routing is discussed in Section 5.8. 435 If two nodes are connected via several links, the ACP SHOULD be 436 established on every link, but it is possible to establish the ACP 437 only on a sub-set of links. Having an ACP channel on every link has 438 a number of advantages, for example it allows for a faster failover 439 in case of link failure, and it reflects the physical topology more 440 closely. Using a subset of links (for example, a single link), 441 reduces resource consumption on the devices, because state needs to 442 be kept per ACP channel. 444 5.6. Context Separation 446 The ACP is in a separate context from the normal data plane of the 447 device. This context includes the ACP channels IPv6 forwarding and 448 routing as well as any required higher layer ACP functions. 450 In classical network device platforms, a dedicated so called "Virtual 451 routing and forwarding instance" (VRF) is one logical implementation 452 option for the ACP. If possible by the platform SW architecture, 453 separation options that minimize shared components are preferred. 454 The context for the ACP needs to be established automatically during 455 bootstrap of a device and - as necessitated by the implementation 456 option be protected from being modified unintential from data plane 457 configuration. 459 In addition this provides for security, because the ACP is not 460 reachable from the global routing table. Also, configuration errors 461 from the data plane setup do not affect the ACP. 463 5.7. Addressing inside the ACP 465 The channels explained above only establish communication between two 466 adjacent neighbors. In order for the communication to happen across 467 multiple hops, the autonomic control plane requires internal network 468 wide valid addresses and routing. Each autonomic node must create a 469 loopback interface with a network wide unique address inside the ACP 470 context mentioned in Section 5.6. 472 We suggest to create network wide Unique Local Addresses (ULA) in 473 accordance with [RFC4193] with the following algorithm: 475 o Prefix FC01::/8 476 o Global ID: a hash of the domain ID; this way all devices in the 477 same domain have the same /48 prefix. Conversely, global ID from 478 different domains are unlikely to clash, such that two networks 479 can be merged, as long as the policy allows that merge. See also 480 Section 6 for a discussion on merging domains. 482 o Subnet ID and interface ID: These can be either derived 483 deterministically from the name of the device, or assigned at 484 registration time of the device. 486 Links inside the ACP only use link-local IPv6 addressing, such that 487 each node only requires one routable loopback address. 489 5.8. Routing in the ACP 491 Once ULA address are set up all autonomic entities should run a 492 routing protocol within the autonomic control plane context. This 493 routing protocol distributes the ULA created in the previous section 494 for reachability. The use of the autonomic control plane specific 495 context eliminates the probable clash with the global routing table 496 and also secures the ACP from interference from the configuration 497 mismatch or incorrect routing updates. 499 The establishment of the routing plane and its parameters are 500 automatic and strictly within the confines of the autonomic control 501 plane. Therefore, no manual configuration is required. 503 All routing updates are automatically secured in transit as the 504 channels of the autonomic control plane are by default secured. 506 The routing protocol inside the ACP should be light weight and highly 507 scalable to ensure that the ACP does not become a limiting factor in 508 network scalability. We suggest the use of RPL as one such protocol 509 which is light weight and scales well for the control plane traffic. 511 5.9. Connecting a Controller / NMS system 513 The Autonomic Control Plane can be used by management systems, such 514 as controllers or network management system (NMS) hosts (henceforth 515 called simply "NMS hosts"), to connect to devices through it. For 516 this, an NMS host must have access to the ACP. By default, the ACP 517 is a self-protecting overlay network, which only allows access to 518 trusted systems. Therefore, a traditional NMS system does not have 519 access to the ACP by default, just like any other external device. 521 The preferred way for an NMS host to connect to the ACP of a network 522 is to enrol that NMS host as a domain device, such that it shares a 523 domain certificate with the same trust anchor as the network devices. 525 Then, the NMS host can automatically discover an adjacent network 526 element, and join the ACP automatically, just like a network device 527 would connect to a neighboring device. Alternatively, if there is no 528 directly connected autonomic network element, a secure connection to 529 a single remote network element can be established by configuration, 530 authenticated using the domain certificates. There, the NMS host 531 "enters" the ACP, from which point it can use the ACP to reach 532 further nodes. 534 If the NMS host does not support autonomic negotiation of the ACP, 535 then it can be brought into the ACP by configuration. On an adjacent 536 autonomic node with ACP, the interface with the NMS host can be 537 configured to be part of the ACP. In this case, the NMS host is with 538 this interface entirely and exclusively inside the ACP. It would 539 likely require a second interface for connections between the NMS 540 host and administrators, or Internet based services. This mode of 541 connecting an NMS host has security consequences: All systems and 542 processes connected to this implicitly trusted interface have access 543 to all autonomic nodes on the entire ACP, without further 544 authentication. Thus, this connection must be physically controlled. 546 In both options, the NMS host must be routed in the ACP. This 547 involves two parts: 1) the NMS host must point default to the AN 548 device for all IPv6, or for the ULA prefix used inside the ACP, and 549 2) the prefix used between AN node and NMS host must be announced 550 into the ACP, and distributed there. 552 The document "Autonomic Network Stable Connectivity" 553 [I-D.eckert-anima-stable-connectivity] explains in more detail how 554 the ACP can be integrated in a mixed NOC environment. 556 6. Self-Healing Properties 558 The ACP is self-healing: 560 o New neighbors will automatically join the ACP after successful 561 validation and will become reachable using their unique ULA 562 address across the ACP. 564 o When any changes happen in the topology, the routing protocol used 565 in the ACP will automatically adapt to the changes and will 566 continue to provide reachability to all devices. 568 o If an existing device gets revoked, it will automatically be 569 denied access to the ACP as its domain certificate will be 570 validated against a Certificate Revocation List during 571 authentication. Since the revocation check is only done at the 572 establishment of a new security association, existing ones are not 573 automatically torn down. If an immediate disconnect is required, 574 existing sessions to a freshly revoked device can be re-set. 576 The ACP can also sustain network partitions and mergers. Practically 577 all ACP operations are link local, where a network partition has no 578 impact. Devices authenticate each other using the domain 579 certificates to establish the ACP locally. Addressing inside the ACP 580 remains unchanged, and the routing protocol inside both parts of the 581 ACP will lead to two working (although partitioned) ACPs. 583 There are few central dependencies: A certificate revocation list 584 (CRL) may not be available during a network partition; a suitable 585 policy to not immediately disconnect neighbors when no CRL is 586 available can address this issue. Also, a registrar or Certificate 587 Authority might not be available during a partition. This may delay 588 renewal of certificates that are to expire in the future, and it may 589 prevent the enrolment of new devices during the partition. 591 After a network partition, a re-merge will just establish the 592 previous status, certificates can be renewed, the CRL is available, 593 and new devices can be enrolled everywhere. Since all devices use 594 the same trust anchor, a re-merge will be smooth. 596 Merging two networks with different trust anchors requires the trust 597 anchors to mutually trust each other (for example, by cross-signing). 598 As long as the domain names are different, the addressing will not 599 overlap (see Section 5.7). 601 7. Self-Protection Properties 603 As explained in Section 5, the ACP is based on channels being built 604 between devices which have been previously authenticated based on 605 their domain certificates. The channels themselves are protected 606 using standard encryption technologies like DTLS or IPsec which 607 provide additional authentication during channel establishment, data 608 integrity and data confidentiality protection of data inside the ACP 609 and in addition, provide replay protection. 611 An attacker will therefore not be able to join the ACP unless having 612 a valid domain certificate, also packet injection and sniffing 613 traffic will not be possible due to the security provided by the 614 encryption protocol. 616 The remaining attack vector would be to attack the underlying AN 617 protocols themselves, either via directed attacks or by denial-of- 618 service attacks. However, as the ACP is built using link-local IPv6 619 address, remote attacks are impossible. The ULA addresses are only 620 reachable inside the ACP context, therefore unreachable from the data 621 plane. Also, the ACP protocols should be implemented to be attack 622 resistant and not consume unnecessary resources even while under 623 attack. 625 8. The Administrator View 627 An ACP is self-forming, self-managing and self-protecting, therefore 628 has minimal dependencies on the administrator of the network. 629 Specifically, it cannot be configured, there is therefore no scope 630 for configuration errors on the ACP itself. The administrator may 631 have the option to enable or disable the entire approach, but 632 detailed configuration is not possible. This means that the ACP must 633 not be reflected in the running configuration of devices, except a 634 possible on/off switch. 636 While configuration is not possible, an administrator must have full 637 visibility of the ACP and all its parameters, to be able to do 638 trouble-shooting. Therefore, an ACP must support all show and debug 639 options, as for any other network function. Specifically, a network 640 management system or controller must be able to discover the ACP, and 641 monitor its health. This visibility of ACP operations must clearly 642 be separated from visibility of data plane so automated systems will 643 never have to deal with ACP aspect unless they explicitly desire to 644 do so. 646 Since an ACP is self-protecting, a device not supporting the ACP, or 647 without a valid domain certificate cannot connect to it. This means 648 that by default a traditional controller or network management system 649 cannot connect to an ACP. See Section 5.9 for more details on how to 650 connect an NMS host into the ACP. 652 9. Security Considerations 654 An ACP is self-protecting and there is no need to apply configuration 655 to make it secure. Its security therefore does not depend on 656 configuration. 658 However, the security of the ACP depends on a number of other 659 factors: 661 o The usage of domain certificates depends on a valid supporting PKI 662 infrastructure. If the chain of trust of this PKI infrastructure 663 is compromised, the security of the ACP is also compromised. This 664 is typically under the control of the network administrator. 666 o Security can be compromised by implementation errors (bugs), as in 667 all products. 669 Fundamentally, security depends on correct operation, implementation 670 and architecture. Autonomic approaches such as the ACP largely 671 eliminate the dependency on correct operation; implementation and 672 architectural mistakes are still possible, as in all networking 673 technologies. 675 10. IANA Considerations 677 This document requests no action by IANA. 679 11. Acknowledgements 681 This work originated from an Autonomic Networking project at Cisco 682 Systems, which started in early 2010. Many people contributed to 683 this project and the idea of the Autonomic Control Plane, amongst 684 which (in alphabetical order): Ignas Bagdonas, Parag Bhide, Alex 685 Clemm, Toerless Eckert, Yves Hertoghs, Bruno Klauser, Max Pritikin, 686 Ravi Kumar Vadapalli. 688 Further input and suggestions were received from: Rene Struik, Brian 689 Carpenter, Benoit Claise. 691 12. Change log [RFC Editor: Please remove] 693 12.1. Initial version 695 First version of this document: 696 [I-D.behringer-autonomic-control-plane] 698 12.2. draft-behringer-anima-autonomic-control-plane-00 700 Initial version of the anima document; only minor edits. 702 12.3. draft-behringer-anima-autonomic-control-plane-01 704 o Clarified that the ACP should be based on, and support only IPv6. 706 o Clarified in intro that ACP is for both, between devices, as well 707 as for access from a central entity, such as an NMS. 709 o Added a section on how to connect an NMS system. 711 o Clarified the hop-by-hop crypto nature of the ACP. 713 o Added several references to GDNP as a candidate protocol. 715 o Added a discussion on network split and merge. Although, this 716 should probably go into the certificate management story longer 717 term. 719 12.4. draft-behringer-anima-autonomic-control-plane-02 721 Addresses (numerous) comments from Brian Carpenter. See mailing list 722 for details. The most important changes are: 724 Introduced a new section "overview", to ease the understanding of 725 the approach. 727 Merged the previous "problem statement" and "use case" sections 728 into a mostly re-written "use cases" section, since they were 729 overlapping. 731 Clarified the relationship with draft-eckert-anima-stable- 732 connectivity 734 12.5. draft-behringer-anima-autonomic-control-plane-03 736 o Took out requirement for IPv6 --> that's in the reference doc. 738 o Added requirement section. 740 o Changed focus: more focus on autonomic functions, not only virtual 741 out of band. This goes a bit throughout the document, starting 742 with a changed abstract and intro. 744 12.6. draft-ietf-anima-autonomic-control-plane-00 746 No changes; re-submitted as WG document. 748 13. References 750 [I-D.behringer-anima-reference-model] 751 Behringer, M., Carpenter, B., Eckert, T., Ciavaglia, L., 752 Liu, B., Jeff, J., and J. Strassner, "A Reference Model 753 for Autonomic Networking", draft-behringer-anima- 754 reference-model-03 (work in progress), June 2015. 756 [I-D.behringer-autonomic-control-plane] 757 Behringer, M., Bjarnason, S., BL, B., and T. Eckert, "An 758 Autonomic Control Plane", draft-behringer-autonomic- 759 control-plane-00 (work in progress), June 2014. 761 [I-D.carpenter-anima-gdn-protocol] 762 Carpenter, B. and B. Liu, "A Generic Discovery and 763 Negotiation Protocol for Autonomic Networking", draft- 764 carpenter-anima-gdn-protocol-04 (work in progress), June 765 2015. 767 [I-D.eckert-anima-stable-connectivity] 768 Eckert, T. and M. Behringer, "Using Autonomic Control 769 Plane for Stable Connectivity of Network OAM", draft- 770 eckert-anima-stable-connectivity-01 (work in progress), 771 March 2015. 773 [I-D.pritikin-anima-bootstrapping-keyinfra] 774 Pritikin, M., Richardson, M., Behringer, M., and S. 775 Bjarnason, "Bootstrapping Key Infrastructures", draft- 776 pritikin-anima-bootstrapping-keyinfra-02 (work in 777 progress), July 2015. 779 [RFC4122] Leach, P., Mealling, M., and R. Salz, "A Universally 780 Unique IDentifier (UUID) URN Namespace", RFC 4122, 781 DOI 10.17487/RFC4122, July 2005, 782 . 784 [RFC4193] Hinden, R. and B. Haberman, "Unique Local IPv6 Unicast 785 Addresses", RFC 4193, DOI 10.17487/RFC4193, October 2005, 786 . 788 [RFC7575] Behringer, M., Pritikin, M., Bjarnason, S., Clemm, A., 789 Carpenter, B., Jiang, S., and L. Ciavaglia, "Autonomic 790 Networking: Definitions and Design Goals", RFC 7575, 791 DOI 10.17487/RFC7575, June 2015, 792 . 794 [RFC7576] Jiang, S., Carpenter, B., and M. Behringer, "General Gap 795 Analysis for Autonomic Networking", RFC 7576, 796 DOI 10.17487/RFC7576, June 2015, 797 . 799 Authors' Addresses 801 Michael H. Behringer (editor) 802 Cisco Systems 803 Building D, 45 Allee des Ormes 804 Mougins 06250 805 France 807 Email: mbehring@cisco.com 808 Steinthor Bjarnason 809 Cisco Systems 811 Email: sbjarnas@cisco.com 813 Balaji BL 814 Cisco Systems 816 Email: blbalaji@cisco.com 818 Toerless Eckert 819 Cisco Systems 821 Email: eckert@cisco.com