idnits 2.17.1 draft-ietf-homenet-arch-12.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 date (February 14, 2014) is 3723 days in the past. Is this intentional? Checking references for intended status: Informational ---------------------------------------------------------------------------- ** Obsolete normative reference: RFC 2460 (Obsoleted by RFC 8200) ** Obsolete normative reference: RFC 3633 (Obsoleted by RFC 8415) -- Obsolete informational reference (is this intentional?): RFC 4941 (Obsoleted by RFC 8981) -- Obsolete informational reference (is this intentional?): RFC 6145 (Obsoleted by RFC 7915) -- Obsolete informational reference (is this intentional?): RFC 6204 (Obsoleted by RFC 7084) -- Obsolete informational reference (is this intentional?): RFC 6555 (Obsoleted by RFC 8305) -- Obsolete informational reference (is this intentional?): RFC 6824 (Obsoleted by RFC 8684) == Outdated reference: A later version (-15) exists of draft-ietf-ospf-ospfv3-autoconfig-06 Summary: 2 errors (**), 0 flaws (~~), 2 warnings (==), 6 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Network Working Group T. Chown, Ed. 3 Internet-Draft University of Southampton 4 Intended status: Informational J. Arkko 5 Expires: August 18, 2014 Ericsson 6 A. Brandt 7 Sigma Designs 8 O. Troan 9 Cisco Systems, Inc. 10 J. Weil 11 Time Warner Cable 12 February 14, 2014 14 IPv6 Home Networking Architecture Principles 15 draft-ietf-homenet-arch-12 17 Abstract 19 This text describes evolving networking technology within residential 20 home networks with increasing numbers of devices and a trend towards 21 increased internal routing. The goal of this document is to define a 22 general architecture for IPv6-based home networking, describing the 23 associated principles, considerations and requirements. The text 24 briefly highlights specific implications of the introduction of IPv6 25 for home networking, discusses the elements of the architecture, and 26 suggests how standard IPv6 mechanisms and addressing can be employed 27 in home networking. The architecture describes the need for specific 28 protocol extensions for certain additional functionality. It is 29 assumed that the IPv6 home network is not actively managed, and runs 30 as an IPv6-only or dual-stack network. There are no recommendations 31 in this text for the IPv4 part of the network. 33 Status of this Memo 35 This Internet-Draft is submitted in full conformance with the 36 provisions of BCP 78 and BCP 79. 38 Internet-Drafts are working documents of the Internet Engineering 39 Task Force (IETF). Note that other groups may also distribute 40 working documents as Internet-Drafts. The list of current Internet- 41 Drafts is at http://datatracker.ietf.org/drafts/current/. 43 Internet-Drafts are draft documents valid for a maximum of six months 44 and may be updated, replaced, or obsoleted by other documents at any 45 time. It is inappropriate to use Internet-Drafts as reference 46 material or to cite them other than as "work in progress." 48 This Internet-Draft will expire on August 18, 2014. 50 Copyright Notice 52 Copyright (c) 2014 IETF Trust and the persons identified as the 53 document authors. All rights reserved. 55 This document is subject to BCP 78 and the IETF Trust's Legal 56 Provisions Relating to IETF Documents 57 (http://trustee.ietf.org/license-info) in effect on the date of 58 publication of this document. Please review these documents 59 carefully, as they describe your rights and restrictions with respect 60 to this document. Code Components extracted from this document must 61 include Simplified BSD License text as described in Section 4.e of 62 the Trust Legal Provisions and are provided without warranty as 63 described in the Simplified BSD License. 65 Table of Contents 67 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 4 68 1.1. Terminology and Abbreviations . . . . . . . . . . . . . . 5 69 2. Effects of IPv6 on Home Networking . . . . . . . . . . . . . . 6 70 2.1. Multiple subnets and routers . . . . . . . . . . . . . . . 7 71 2.2. Global addressability and elimination of NAT . . . . . . . 8 72 2.3. Multi-Addressing of devices . . . . . . . . . . . . . . . 8 73 2.4. Unique Local Addresses (ULAs) . . . . . . . . . . . . . . 9 74 2.5. Avoiding manual configuration of IP addresses . . . . . . 10 75 2.6. IPv6-only operation . . . . . . . . . . . . . . . . . . . 11 76 3. Homenet Architecture Principles . . . . . . . . . . . . . . . 11 77 3.1. General Principles . . . . . . . . . . . . . . . . . . . . 12 78 3.1.1. Reuse existing protocols . . . . . . . . . . . . . . . 12 79 3.1.2. Minimise changes to hosts and routers . . . . . . . . 12 80 3.2. Homenet Topology . . . . . . . . . . . . . . . . . . . . . 13 81 3.2.1. Supporting arbitrary topologies . . . . . . . . . . . 13 82 3.2.2. Network topology models . . . . . . . . . . . . . . . 13 83 3.2.3. Dual-stack topologies . . . . . . . . . . . . . . . . 18 84 3.2.4. Multihoming . . . . . . . . . . . . . . . . . . . . . 19 85 3.3. A Self-Organising Network . . . . . . . . . . . . . . . . 20 86 3.3.1. Differentiating neighbouring homenets . . . . . . . . 21 87 3.3.2. Largest practical subnets . . . . . . . . . . . . . . 21 88 3.3.3. Handling varying link technologies . . . . . . . . . . 22 89 3.3.4. Homenet realms and borders . . . . . . . . . . . . . . 22 90 3.3.5. Configuration information from the ISP . . . . . . . . 23 91 3.4. Homenet Addressing . . . . . . . . . . . . . . . . . . . . 23 92 3.4.1. Use of ISP-delegated IPv6 prefixes . . . . . . . . . . 23 93 3.4.2. Stable internal IP addresses . . . . . . . . . . . . . 25 94 3.4.3. Internal prefix delegation . . . . . . . . . . . . . . 26 95 3.4.4. Coordination of configuration information . . . . . . 27 96 3.4.5. Privacy . . . . . . . . . . . . . . . . . . . . . . . 28 98 3.5. Routing functionality . . . . . . . . . . . . . . . . . . 28 99 3.5.1. Multicast support . . . . . . . . . . . . . . . . . . 29 100 3.5.2. Mobility support . . . . . . . . . . . . . . . . . . . 30 101 3.6. Security . . . . . . . . . . . . . . . . . . . . . . . . . 30 102 3.6.1. Addressability vs reachability . . . . . . . . . . . . 31 103 3.6.2. Filtering at borders . . . . . . . . . . . . . . . . . 31 104 3.6.3. Partial Effectiveness of NAT and Firewalls . . . . . . 32 105 3.6.4. Exfiltration concerns . . . . . . . . . . . . . . . . 32 106 3.6.5. Device capabilities . . . . . . . . . . . . . . . . . 33 107 3.6.6. ULAs as a hint of connection origin . . . . . . . . . 33 108 3.7. Naming and Service Discovery . . . . . . . . . . . . . . . 33 109 3.7.1. Discovering services . . . . . . . . . . . . . . . . . 33 110 3.7.2. Assigning names to devices . . . . . . . . . . . . . . 34 111 3.7.3. The homenet name service . . . . . . . . . . . . . . . 35 112 3.7.4. Name spaces . . . . . . . . . . . . . . . . . . . . . 36 113 3.7.5. Independent operation . . . . . . . . . . . . . . . . 38 114 3.7.6. Considerations for LLNs . . . . . . . . . . . . . . . 38 115 3.7.7. DNS resolver discovery . . . . . . . . . . . . . . . . 39 116 3.7.8. Devices roaming to/from the homenet . . . . . . . . . 39 117 3.8. Other Considerations . . . . . . . . . . . . . . . . . . . 39 118 3.8.1. Quality of Service . . . . . . . . . . . . . . . . . . 39 119 3.8.2. Operations and Management . . . . . . . . . . . . . . 40 120 3.9. Implementing the Architecture on IPv6 . . . . . . . . . . 41 121 4. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . 41 122 5. Security Considerations . . . . . . . . . . . . . . . . . . . 42 123 6. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 42 124 7. References . . . . . . . . . . . . . . . . . . . . . . . . . . 42 125 7.1. Normative References . . . . . . . . . . . . . . . . . . . 42 126 7.2. Informative References . . . . . . . . . . . . . . . . . . 42 127 Appendix A. Acknowledgments . . . . . . . . . . . . . . . . . . . 45 128 Appendix B. Changes . . . . . . . . . . . . . . . . . . . . . . . 45 129 B.1. Version 12 . . . . . . . . . . . . . . . . . . . . . . . . 46 130 B.2. Version 11 (after IESG review) . . . . . . . . . . . . . . 46 131 B.3. Version 10 (after AD review) . . . . . . . . . . . . . . . 46 132 B.4. Version 09 (after WGLC) . . . . . . . . . . . . . . . . . 46 133 B.5. Version 08 . . . . . . . . . . . . . . . . . . . . . . . . 47 134 B.6. Version 07 . . . . . . . . . . . . . . . . . . . . . . . . 47 135 B.7. Version 06 . . . . . . . . . . . . . . . . . . . . . . . . 48 136 B.8. Version 05 . . . . . . . . . . . . . . . . . . . . . . . . 48 137 B.9. Version 04 . . . . . . . . . . . . . . . . . . . . . . . . 49 138 B.10. Version 03 . . . . . . . . . . . . . . . . . . . . . . . . 49 139 B.11. Version 02 . . . . . . . . . . . . . . . . . . . . . . . . 50 140 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . . 51 142 1. Introduction 144 This document focuses on evolving networking technology within 145 residential home networks with increasing numbers of devices and a 146 trend towards increased internal routing, and the associated 147 challenges with their deployment and operation. There is a growing 148 trend in home networking for the proliferation of networking 149 technology through an increasingly broad range of devices and media. 150 This evolution in scale and diversity sets requirements on IETF 151 protocols. Some of these requirements relate to the introduction of 152 IPv6, others to the introduction of specialised networks for home 153 automation and sensors. 155 While at the time of writing some complex home network topologies 156 exist, most are relatively simple single subnet networks, and 157 ostensibly operate using just IPv4. While there may be IPv6 traffic 158 within the network, e.g., for service discovery, the homenet is 159 provisioned by the ISP as an IPv4 network. Such networks also 160 typically employ solutions that should be avoided, such as private 161 [RFC1918] addressing with (cascaded) network address translation 162 (NAT) [RFC3022], or they may require expert assistance to set up. 164 In contrast, emerging IPv6-capable home networks are very likely to 165 have multiple internal subnets, e.g., to facilitate private and guest 166 networks, heterogeneous link layers, and smart grid components, and 167 have enough address space available to allow every device to have a 168 globally unique address. This implies that internal routing 169 functionality is required, and that the homenet's ISP both provides a 170 large enough prefix to allocate a prefix to each subnet, and that a 171 method is supported for such prefixes to be delegated efficiently to 172 those subnets. 174 It is not practical to expect home users to configure their networks. 175 Thus the assumption of this document is that the homenet is as far as 176 possible self-organising and self-configuring, i.e., it should 177 function without pro-active management by the residential user. 179 The architectural constructs in this document are focused on the 180 problems to be solved when introducing IPv6, with an eye towards a 181 better result than what we have today with IPv4, as well as aiming at 182 a more consistent solution that addresses as many of the identified 183 requirements as possible. The document aims to provide the basis and 184 guiding principles for how standard IPv6 mechanisms and addressing 185 [RFC2460] [RFC4291] can be employed in home networking, while 186 coexisting with existing IPv4 mechanisms. In emerging dual-stack 187 home networks it is vital that introducing IPv6 does not adversely 188 affect IPv4 operation. We assume that the IPv4 network architecture 189 in home networks is what it is, and can not be modified by new 190 recommendations. This document does not discuss how IPv4 home 191 networks provision or deliver support for multiple subnets. It 192 should not be assumed that any future new functionality created with 193 IPv6 in mind will be backward-compatible to include IPv4 support. 194 Further, future deployments, or specific subnets within an otherwise 195 dual-stack home network, may be IPv6-only, in which case 196 considerations for IPv4 impact would not apply. 198 This document proposes a baseline homenet architecture, using 199 protocols and implementations that are as far as possible proven and 200 robust. The scope of the document is primarily the network layer 201 technologies that provide the basic functionality to enable 202 addressing, connectivity, routing, naming and service discovery. 203 While it may, for example, state that homenet components must be 204 simple to deploy and use, it does not discuss specific user 205 interfaces, nor does it discuss specific physical, wireless or data- 206 link layer considerations. Likewise, we also do not specify the 207 whole design of a homenet router from top to bottom, rather we focus 208 on the Layer 3 aspects. This means that Layer 2 is largely out of 209 scope, we're assuming a data link layer that supports IPv6 is 210 present, and that we react accordingly. Any IPv6-over-Foo 211 definitions occur elsewhere. 213 [RFC6204] defines basic requirements for customer edge routers 214 (CERs). This document has recently been updated with the definition 215 of requirements for specific transition tools on the CER in 216 [RFC7084], specifically DS-Lite [RFC6333] and 6rd [RFC5969]. Such 217 detailed specification of CER devices is considered out of scope of 218 this architecture document, and we assume that any required update of 219 the CER device specification as a result of adopting this 220 architecture will be handled as separate and specific updates to 221 these existing documents. Further, the scope of this text is the 222 internal homenet, and thus specific features on the WAN side of the 223 CER are out of scope for this text. 225 1.1. Terminology and Abbreviations 227 In this section we define terminology and abbreviations used 228 throughout the text. 230 o Border: a point, typically resident on a router, between two 231 networks, e.g., between the main internal homenet and a guest 232 network. This defines point(s) at which filtering and forwarding 233 policies for different types of traffic may be applied. 235 o CER: Customer Edge Router: A border router intended for use in a 236 homenet, which connects the homenet to a service provider network. 238 o FQDN: Fully Qualified Domain Name. A globally unique name. 240 o Guest network: A part of the home network intended for use by 241 visitors or guests to the home(net). Devices on the guest network 242 may typically not see or be able to use all services in the 243 home(net). 245 o Homenet: A home network, comprising host and router equipment, 246 with one or more CERs providing connectivity to service provider 247 network(s). 249 o Internet Service Provider (ISP): an entity that provides access to 250 the Internet. In this document, a service provider specifically 251 offers Internet access using IPv6, and may also offer IPv4 252 Internet access. The service provider can provide such access 253 over a variety of different transport methods such as DSL, cable, 254 wireless, and others. 256 o LLN: Low-power and lossy network. 258 o LQDN: Locally Qualified Domain Name. A name local to the homenet. 260 o NAT: Network Address Translation. Typically referring to IPv4 261 Network Address and Port Translation (NAPT) [RFC3022]. 263 o NPTv6: Network Prefix Translation for IPv6 [RFC6296]. 265 o PCP: Port Control Protocol [RFC6887]. 267 o Realm: a network delimited by a defined border. A guest network 268 within a homenet may form one realm. 270 o 'Simple Security'. Defined in [RFC4864] and expanded further in 271 [RFC6092]; describes recommended perimeter security capabilities 272 for IPv6 networks. 274 o ULA: IPv6 Unique Local Address [RFC4193]. 276 o VM: Virtual machine. 278 2. Effects of IPv6 on Home Networking 280 While IPv6 resembles IPv4 in many ways, there are some notable 281 differences in the way it may typically be deployed. It changes 282 address allocation principles, making multi-addressing the norm, and, 283 through the vastly increased address space, allows globally unique IP 284 addresses to be used for all devices in a home network. This section 285 presents an overview of some of the key implications of the 286 introduction of IPv6 for home networking, that are simultaneously 287 both promising and problematic. 289 2.1. Multiple subnets and routers 291 While simple layer 3 topologies involving as few subnets as possible 292 are preferred in home networks, the incorporation of dedicated 293 (routed) subnets remains necessary for a variety of reasons. For 294 instance, an increasingly common feature in modern home routers is 295 the ability to support both guest and private network subnets. 296 Likewise, there may be a need to separate home automation or 297 corporate extension LANs (whereby a home worker can have their 298 corporate network extended into the home using a virtual private 299 network, commonly presented as one port on an Ethernet device) from 300 the main Internet access network, or different subnets may in general 301 be associated with parts of the homenet that have different routing 302 and security policies. Further, link layer networking technology is 303 poised to become more heterogeneous, as networks begin to employ both 304 traditional Ethernet technology and link layers designed for low- 305 power and lossy networks (LLNs), such as those used for certain types 306 of sensor devices. Constraining the flow of certain traffic from 307 Ethernet links to much lower capacity links thus becomes an important 308 topic. 310 The introduction of IPv6 for home networking makes it possible for 311 every home network to be delegated enough address space from its ISP 312 to provision globally unique prefixes for each such subnet in the 313 home. While the number of addresses in a standard /64 IPv6 prefix is 314 practically unlimited, the number of prefixes available for 315 assignment to the home network is not. As a result the growth 316 inhibitor for the home network shifts from the number of addresses to 317 the number of prefixes offered by the provider; this topic is 318 discussed in [RFC6177] (BCP 157), which recommends that "end sites 319 always be able to obtain a reasonable amount of address space for 320 their actual and planned usage". 322 The addition of routing between subnets raises a number of issues. 323 One is a method by which prefixes can be efficiently allocated to 324 each subnet, without user intervention. Another is the issue of how 325 to extend mechanisms such as zero configuration service discovery 326 which currently only operate within a single subnet using link-local 327 traffic. In a typical IPv4 home network, there is only one subnet, 328 so such mechanisms would normally operate as expected. For multi- 329 subnet IPv6 home networks there are two broad choices to enable such 330 protocols to work across the scope of the entire homenet; extend 331 existing protocols to work across that scope, or introduce proxies 332 for existing link layer protocols. This topic is discussed in 333 Section 3.7. 335 2.2. Global addressability and elimination of NAT 337 The possibility for direct end-to-end communication on the Internet 338 to be restored by the introduction of IPv6 is on the one hand an 339 incredible opportunity for innovation and simpler network operation, 340 but on the other hand it is also a concern as it potentially exposes 341 nodes in the internal networks to receipt of unwanted and possibly 342 malicious traffic from the Internet. 344 With devices and applications able to talk directly to each other 345 when they have globally unique addresses, there may be an expectation 346 of improved host security to compensate for this. It should be noted 347 that many devices may (for example) ship with default settings that 348 make them readily vulnerable to compromise by external attackers if 349 globally accessible, or may simply not have robustness designed-in 350 because it was either assumed such devices would only be used on 351 private networks or the device itself doesn't have the computing 352 power to apply the necessary security methods. In addition, the 353 upgrade cycle for devices (or their firmware) may be slow, and/or 354 lack auto-update mechanisms. 356 It is thus important to distinguish between addressability and 357 reachability. While IPv6 offers global addressability through use of 358 globally unique addresses in the home, whether devices are globally 359 reachable or not would depend on any firewall or filtering 360 configuration, and not, as is commonly the case with IPv4, the 361 presence or use of NAT. In this respect, IPv6 networks may or may 362 not have filters applied at their borders to control such traffic, 363 i.e., at the homenet CER. [RFC4864] and [RFC6092] discuss such 364 filtering, and the merits of 'default allow' against 'default deny' 365 policies for external traffic initiated into a homenet. This 366 document takes no position on which mode is the default, but assumes 367 the choice for the homenet to use either mode would be available. 369 This topic is discussed further in Section 3.6.1. 371 2.3. Multi-Addressing of devices 373 In an IPv6 network, devices will often acquire multiple addresses, 374 typically at least a link-local address and one or more globally 375 unique addresses. Where a homenet is multihomed, a device would 376 typically receive a globally unique address (GUA) from within the 377 delegated prefix from each upstream ISP. Devices may also have an 378 IPv4 address if the network is dual-stack, an IPv6 Unique Local 379 Address (ULA) [RFC4193] (see below), and one or more IPv6 Privacy 380 Addresses [RFC4941]. 382 It should thus be considered the norm for devices on IPv6 home 383 networks to be multi-addressed, and to need to make appropriate 384 address selection decisions for the candidate source and destination 385 address pairs for any given connection. Default Address Selection 386 for IPv6 [RFC6724] provides a solution for this, though it may face 387 problems in the event of multihoming where, as described above, nodes 388 will be configured with one address from each upstream ISP prefix. 389 In such cases the presence of upstream BCP 38 [RFC2827] ingress 390 filtering requires multi-addressed nodes to select the correct source 391 address to be used for the corresponding uplink. A challenge here is 392 that the node may not have the information it needs to make that 393 decision based on addresses alone. We discuss this challenge in 394 Section 3.2.4. 396 2.4. Unique Local Addresses (ULAs) 398 [RFC4193] defines Unique Local Addresses (ULAs) for IPv6 that may be 399 used to address devices within the scope of a single site. Support 400 for ULAs for IPv6 CERs is described in [RFC6204]. A home network 401 running IPv6 should deploy ULAs alongside its globally unique 402 prefix(es) to allow stable communication between devices (on 403 different subnets) within the homenet where that externally allocated 404 globally unique prefix may change over time, e.g., due to renumbering 405 within the subscriber's ISP, or where external connectivity may be 406 temporarily unavailable. A homenet using provider-assigned global 407 addresses is exposed to its ISP renumbering the network to a much 408 larger degree than before whereas, for IPv4, NAT isolated the user 409 against ISP renumbering to some extent. 411 While setting up a network there may be a period where it has no 412 external connectivity, in which case ULAs would be required for 413 inter-subnet communication. In the case where home automation 414 networks are being set up in a new home/deployment (as early as 415 during construction of the home), such networks will likely need to 416 use their own /48 ULA prefix. Depending upon circumstances beyond 417 the control of the owner of the homenet, it may be impossible to 418 renumber the ULA used by the home automation network so routing 419 between ULA /48s may be required. Also, some devices, particularly 420 constrained devices, may have only a ULA (in addition to a link- 421 local), while others may have both a GUA and a ULA. 423 Note that unlike private IPv4 RFC 1918 space, the use of ULAs does 424 not imply use of an IPv6 equivalent of a traditional IPv4 NAT 425 [RFC3022], or of NPTv6 prefix-based NAT [RFC6296]. When an IPv6 node 426 in a homenet has both a ULA and a globally unique IPv6 address, it 427 should only use its ULA address internally, and use its additional 428 globally unique IPv6 address as a source address for external 429 communications. This should be the natural behaviour given support 430 for Default Address Selection for IPv6 [RFC6724]. By using such 431 globally unique addresses between hosts and devices in remote 432 networks, the architectural cost and complexity, particularly to 433 applications, of NAT or NPTv6 translation is avoided. As such, 434 neither IPv6 NAT or NPTv6 is recommended for use in the homenet 435 architecture. Further, the homenet border router(s) should filter 436 packets with ULA source/destination addresses as discussed in 437 Section 3.4.2. 439 Devices in a homenet may be given only a ULA as a means to restrict 440 reachability from outside the homenet. ULAs can be used by default 441 for devices that, without additional configuration (e.g., via a web 442 interface), would only offer services to the internal network. For 443 example, a printer might only accept incoming connections on a ULA 444 until configured to be globally reachable, at which point it acquires 445 a global IPv6 address and may be advertised via a global name space. 447 Where both a ULA and a global prefix are in use, the ULA source 448 address is used to communicate with ULA destination addresses when 449 appropriate, i.e., when the ULA source and destination lie within the 450 /48 ULA prefix(es) known to be used within the same homenet. In 451 cases where multiple /48 ULA prefixes are in use within a single 452 homenet (perhaps because multiple homenet routers each independently 453 auto-generate a /48 ULA prefix and then share prefix/routing 454 information), utilising a ULA source address and a ULA destination 455 address from two disjoint internal ULA prefixes is preferable to 456 using GUAs. 458 While a homenet should operate correctly with two or more /48 ULAs 459 enabled, a mechanism for the creation and use of a single /48 ULA 460 prefix is desirable for addressing consistency and policy 461 enforcement. 463 A counter-argument to using ULAs is that it is undesirable to 464 aggressively deprecate global prefixes for temporary loss of 465 connectivity, so for a host to lose its global address there would 466 have to be a connection breakage longer than the lease period, and 467 even then, deprecating prefixes when there is no connectivity may not 468 be advisable. However, it is assumed in this architecture that 469 homenets should support and use ULAs. 471 2.5. Avoiding manual configuration of IP addresses 473 Some IPv4 home networking devices expose IPv4 addresses to users, 474 e.g., the IPv4 address of a home IPv4 CER that may be configured via 475 a web interface. In potentially complex future IPv6 homenets, users 476 should not be expected to enter IPv6 literal addresses in devices or 477 applications, given their much greater length and the apparent 478 randomness of such addresses to a typical home user. Thus, even for 479 the simplest of functions, simple naming and the associated (minimal, 480 and ideally zero configuration) discovery of services is imperative 481 for the easy deployment and use of homenet devices and applications. 483 2.6. IPv6-only operation 485 It is likely that IPv6-only networking will be deployed first in new 486 home network deployments, often refered to as 'greenfield' scenarios, 487 where there is no existing IPv4 capability, or perhaps as one element 488 of an otherwise dual-stack network. Running IPv6-only adds 489 additional requirements, e.g., for devices to get configuration 490 information via IPv6 transport (not relying on an IPv4 protocol such 491 as IPv4 DHCP), and for devices to be able to initiate communications 492 to external devices that are IPv4-only. 494 Some specific transition technologies which may be deployed by the 495 homenet's ISP are discussed in [RFC1918]. In addition, certain other 496 functions may be desirable on the CER, e.g., to access content in the 497 IPv4 Internet, NAT64 [RFC6144] and DNS64 [RFC6145] may be applicable. 499 The widespread availability of robust solutions to these types of 500 requirements will help accelerate the uptake of IPv6-only homenets. 501 The specifics of these are however beyond the scope of this document, 502 especially those functions that reside on the CER. 504 3. Homenet Architecture Principles 506 The aim of this text is to outline how to construct advanced IPv6- 507 based home networks involving multiple routers and subnets using 508 standard IPv6 addressing and protocols [RFC2460] [RFC4291] as the 509 basis. As described in Section 3.1, solutions should as far as 510 possible re-use existing protocols, and minimise changes to hosts and 511 routers, but some new protocols, or extensions, are likely to be 512 required. In this section, we present the elements of the proposed 513 home networking architecture, with discussion of the associated 514 design principles. 516 In general, home network equipment needs to be able to operate in 517 networks with a range of different properties and topologies, where 518 home users may plug components together in arbitrary ways and expect 519 the resulting network to operate. Significant manual configuration 520 is rarely, if at all, possible, or even desirable given the knowledge 521 level of typical home users. Thus the network should, as far as 522 possible, be self-configuring, though configuration by advanced users 523 should not be precluded. 525 The homenet needs to be able to handle or provision at least 527 o Routing 529 o Prefix configuration for routers 531 o Name resolution 533 o Service discovery 535 o Network security 537 The remainder of this document describes the principles by which the 538 homenet architecture may deliver these properties. 540 3.1. General Principles 542 There is little that the Internet standards community can do about 543 the physical topologies or the need for some networks to be separated 544 at the network layer for policy or link layer compatibility reasons. 545 However, there is a lot of flexibility in using IP addressing and 546 inter-networking mechanisms. This text discusses how such 547 flexibility should be used to provide the best user experience and 548 ensure that the network can evolve with new applications in the 549 future. The principles described in this text should be followed 550 when designing homenet protocol solutions. 552 3.1.1. Reuse existing protocols 554 It is desirable to reuse existing protocols where possible, but at 555 the same time to avoid consciously precluding the introduction of new 556 or emerging protocols. A generally conservative approach, giving 557 weight to running (and available) code, is preferable. Where new 558 protocols are required, evidence of commitment to implementation by 559 appropriate vendors or development communities is highly desirable. 560 Protocols used should be backwardly compatible, and forward 561 compatible where changes are made. 563 3.1.2. Minimise changes to hosts and routers 565 In order to maximise deployability of new homenets, where possible 566 any requirement for changes to hosts and routers should be minimised, 567 though solutions which, for example, incrementally improve capability 568 with host or router changes may be acceptable. There may be cases 569 where changes are unavoidable, e.g., to allow a given homenet routing 570 protocol to be self-configuring. 572 3.2. Homenet Topology 574 This section considers homenet topologies, and the principles that 575 may be applied in designing an architecture to support as wide a 576 range of such topologies as possible. 578 3.2.1. Supporting arbitrary topologies 580 There should ideally be no built-in assumptions about the topology in 581 home networks, as users are capable of connecting their devices in 582 'ingenious' ways. Thus arbitrary topologies and arbitrary routing 583 will need to be supported, or at least the failure mode for when the 584 user makes a mistake should be as robust as possible, e.g., de- 585 activating a certain part of the infrastructure to allow the rest to 586 operate. In such cases, the user should ideally have some useful 587 indication of the failure mode encountered. 589 There should be no topology scenarios which cause loss of 590 connectivity, except when the user creates a physical island within 591 the topology. Some potentially pathological cases that can be 592 created include bridging ports of a router together, however this 593 case can be detected and dealt with by the router. Loops within a 594 routed topology are in a sense good in that they offer redundancy. 595 Bridging loops can be dangerous but are also detectable when a switch 596 learns the MAC of one of its interfaces on another or runs a spanning 597 tree or link state protocol. It is only loops using simple repeaters 598 that are truly pathological. 600 The topology of the homenet may change over time, due to the addition 601 or removal of equipment, but also due to temporary failures or 602 connectivity problems. In some cases this may lead to, for example, 603 a multihomed homenet being split into two isolated homenets, or, 604 after such a fault is remedied, two isolated parts reconfiguring back 605 to a single network. 607 3.2.2. Network topology models 609 As hinted above, while the architecture may focus on likely common 610 topologies, it should not preclude any arbitrary topology from being 611 constructed. 613 Most IPv4 home network models at the time of writing tend to be 614 relatively simple, typically a single NAT router to the ISP and a 615 single internal subnet but, as discussed earlier, evolution in 616 network architectures is driving more complex topologies, such as the 617 separation of guest and private networks. There may also be some 618 cascaded IPv4 NAT scenarios, which we mention in the next section. 619 For IPv6 homenets, the Network Architectures described in [RFC6204] 620 and its successor [RFC7084] should, as a minimum, be supported. 622 There are a number of properties or attributes of a home network that 623 we can use to describe its topology and operation. The following 624 properties apply to any IPv6 home network: 626 o Presence of internal routers. The homenet may have one or more 627 internal routers, or may only provide subnetting from interfaces 628 on the CER. 630 o Presence of isolated internal subnets. There may be isolated 631 internal subnets, with no direct connectivity between them within 632 the homenet (with each having its own external connectivity). 633 Isolation may be physical, or implemented via IEEE 802.1q VLANs. 634 The latter is however not something a typical user would be 635 expected to configure. 637 o Demarcation of the CER. The CER(s) may or may not be managed by 638 the ISP. If the demarcation point is such that the customer can 639 provide or manage the CER, its configuration must be simple. Both 640 models must be supported. 642 Various forms of multihoming are likely to become more prevalent with 643 IPv6 home networks, where the homenet may have two or more external 644 ISP connections, as discussed further below. Thus the following 645 properties should also be considered for such networks: 647 o Number of upstream providers. The majority of home networks today 648 consist of a single upstream ISP, but it may become more common in 649 the future for there to be multiple ISPs, whether for resilience 650 or provision of additional services. Each would offer its own 651 prefix. Some may or may not provide a default route to the public 652 Internet. 654 o Number of CERs. The homenet may have a single CER, which might be 655 used for one or more providers, or multiple CERs. The presence of 656 multiple CERs adds additional complexity for multihoming 657 scenarios, and protocols like PCP that may need to manage 658 connection-oriented state mappings on the same CER as used for 659 subsequent traffic flows. 661 In the following sections we give some examples of the types of 662 homenet topologies we may see in the future. This is not intended to 663 be an exhaustive or complete list, rather an indicative one to 664 facilitate the discussion in this text. 666 3.2.2.1. A: Single ISP, Single CER, Internal routers 668 Figure 1 shows a home network with multiple local area networks. 669 These may be needed for reasons relating to different link layer 670 technologies in use or for policy reasons, e.g., classic Ethernet in 671 one subnet and a LLN link layer technology in another. In this 672 example there is no single router that a priori understands the 673 entire topology. The topology itself may also be complex, and it may 674 not be possible to assume a pure tree form, for instance (because 675 home users may plug routers together to form arbitrary topologies 676 including loops). 678 +-------+-------+ \ 679 | Service | \ 680 | Provider | | Service 681 | Router | | Provider 682 +-------+-------+ | network 683 | / 684 | Customer / 685 | Internet connection 686 | 687 +------+--------+ \ 688 | IPv6 | \ 689 | Customer Edge | \ 690 | Router | | 691 +----+-+---+----+ | 692 Network A | | | Network B(E) | 693 ----+-------------+----+ | +---+-------------+------+ | 694 | | | | | | | 695 +----+-----+ +-----+----+ | +----+-----+ +-----+----+ | | 696 |IPv6 Host | |IPv6 Host | | | IPv6 Host| |IPv6 Host | | | 697 | H1 | | H2 | | | H3 | | H4 | | | 698 +----------+ +----------+ | +----------+ +----------+ | | 699 | | | | | 700 Link F | ---+------+------+-----+ | 701 | | Network E(B) | 702 +------+--------+ | | End-User 703 | IPv6 | | | networks 704 | Interior +------+ | 705 | Router | | 706 +---+-------+-+-+ | 707 Network C | | Network D | 708 ----+-------------+---+ +---+-------------+--- | 709 | | | | | 710 +----+-----+ +-----+----+ +----+-----+ +-----+----+ | 711 |IPv6 Host | |IPv6 Host | | IPv6 Host| |IPv6 Host | | 712 | H5 | | H6 | | H7 | | H8 | / 713 +----------+ +----------+ +----------+ +----------+ / 715 Figure 1 717 In this diagram there is one CER. It has a single uplink interface. 718 It has three additional interfaces connected to Network A, Link F, 719 and Network B. IPv6 Internal Router (IR) has four interfaces 720 connected to Link F, Network C, Network D and Network E. Network B 721 and Network E have been bridged, likely inadvertently. This could be 722 as a result of connecting a wire between a switch for Network B and a 723 switch for Network E. 725 Any of logical Networks A through F might be wired or wireless. 727 Where multiple hosts are shown, this might be through one or more 728 physical ports on the CER or IPv6 (IR), wireless networks, or through 729 one or more layer-2 only Ethernet switches. 731 3.2.2.2. B: Two ISPs, Two CERs, Shared subnet 733 +-------+-------+ +-------+-------+ \ 734 | Service | | Service | \ 735 | Provider A | | Provider B | | Service 736 | Router | | Router | | Provider 737 +------+--------+ +-------+-------+ | network 738 | | / 739 | Customer | / 740 | Internet connections | / 741 | | 742 +------+--------+ +-------+-------+ \ 743 | IPv6 | | IPv6 | \ 744 | Customer Edge | | Customer Edge | \ 745 | Router 1 | | Router 2 | / 746 +------+--------+ +-------+-------+ / 747 | | / 748 | | | End-User 749 ---+---------+---+---------------+--+----------+--- | network(s) 750 | | | | \ 751 +----+-----+ +-----+----+ +----+-----+ +-----+----+ \ 752 |IPv6 Host | |IPv6 Host | | IPv6 Host| |IPv6 Host | / 753 | H1 | | H2 | | H3 | | H4 | / 754 +----------+ +----------+ +----------+ +----------+ 756 Figure 2 758 Figure 2 illustrates a multihomed homenet model, where the customer 759 has connectivity via CER1 to ISP A and via CER2 to ISP B. This 760 example shows one shared subnet where IPv6 nodes would potentially be 761 multihomed and receive multiple IPv6 global prefixes, one per ISP. 762 This model may also be combined with that shown in Figure 1 to create 763 a more complex scenario with multiple internal routers. Or the above 764 shared subnet may be split in two, such that each CER serves a 765 separate isolated subnet, which is a scenario seen with some IPv4 766 networks today. 768 3.2.2.3. C: Two ISPs, One CER, Shared subnet 770 +-------+-------+ +-------+-------+ \ 771 | Service | | Service | \ 772 | Provider A | | Provider B | | Service 773 | Router | | Router | | Provider 774 +-------+-------+ +-------+-------+ | network 775 | | / 776 | Customer | / 777 | Internet | / 778 | connections | 779 +---------+---------+ \ 780 | IPv6 | \ 781 | Customer Edge | \ 782 | Router | / 783 +---------+---------+ / 784 | / 785 | | End-User 786 ---+------------+-------+--------+-------------+--- | network(s) 787 | | | | \ 788 +----+-----+ +----+-----+ +----+-----+ +-----+----+ \ 789 |IPv6 Host | |IPv6 Host | | IPv6 Host| |IPv6 Host | / 790 | H1 | | H2 | | H3 | | H4 | / 791 +----------+ +----------+ +----------+ +----------+ 793 Figure 3 795 Figure 3 illustrates a model where a home network may have multiple 796 connections to multiple providers or multiple logical connections to 797 the same provider, with shared internal subnets. 799 3.2.3. Dual-stack topologies 801 It is expected that most homenet deployments will for the immediate 802 future be dual-stack IPv4/IPv6. In such networks it is important not 803 to introduce new IPv6 capabilities that would cause a failure if used 804 alongside IPv4+NAT, given that such dual-stack homenets will be 805 commonplace for some time. That said, it is desirable that IPv6 806 works better than IPv4 in as many scenarios as possible. Further, 807 the homenet architecture must operate in the absence of IPv4. 809 A general recommendation is to follow the same topology for IPv6 as 810 is used for IPv4, but not to use NAT. Thus there should be routed 811 IPv6 where an IPv4 NAT is used and, where there is no NAT, routing or 812 bridging may be used. Routing may have advantages when compared to 813 bridging together high speed and lower speed shared media, and in 814 addition bridging may not be suitable for some networks, such as ad- 815 hoc mobile networks. 817 In some cases IPv4 home networks may feature cascaded NATs. End 818 users are frequently unaware that they have created such networks as 819 'home routers' and 'home switches' are frequently confused. In 820 addition, there are cases where NAT routers are included within 821 Virtual Machine Hypervisors, or where Internet connection sharing 822 services have been enabled. This document applies equally to such 823 hidden NAT 'routers'. IPv6 routed versions of such cases will be 824 required. We should thus also note that routers in the homenet may 825 not be separate physical devices; they may be embedded within other 826 devices. 828 3.2.4. Multihoming 830 A homenet may be multihomed to multiple providers, as the network 831 models above illustrate. This may either take a form where there are 832 multiple isolated networks within the home or a more integrated 833 network where the connectivity selection needs to be dynamic. 834 Current practice is typically of the former kind, but the latter is 835 expected to become more commonplace. 837 In the general homenet architecture, multihomed hosts should be 838 multi-addressed with a global IPv6 address from the global prefix 839 delegated from each ISP they communicate with or through. When such 840 multi-addressing is in use, hosts need some way to pick source and 841 destination address pairs for connections. A host may choose a 842 source address to use by various methods, most commonly [RFC6724]. 843 Applications may of course do different things, and this should not 844 be precluded. 846 For the single CER Network Model C illustrated above, multihoming may 847 be offered by source-based routing at the CER. With multiple exit 848 routers, as in CER Network Model B, the complexity rises. Given a 849 packet with a source address on the home network, the packet must be 850 routed to the proper egress to avoid BCP 38 ingress filtering if 851 exiting through the wrong ISP. It is highly desirable that the 852 packet is routed in the most efficient manner to the correct exit, 853 though as a minimum requirement the packet should not be dropped. 855 The homenet architecture should support both the above models, i.e., 856 one or more CERs. However, the general multihoming problem is broad, 857 and solutions suggested to date within the IETF have included complex 858 architectures for monitoring connectivity, traffic engineering, 859 identifier-locator separation, connection survivability across 860 multihoming events, and so on. It is thus important that the homenet 861 architecture should as far as possible minimise the complexity of any 862 multihoming support. 864 An example of such a 'simpler' approach has been documented in 865 [I-D.ietf-v6ops-ipv6-multihoming-without-ipv6nat]. Alternatively a 866 flooding/routing protocol could potentially be used to pass 867 information through the homenet, such that internal routers and 868 ultimately end hosts could learn per-prefix configuration 869 information, allowing better address selection decisions to be made. 870 However, this would imply router and, most likely, host changes. 871 Another avenue is to introduce support throughout the homenet for 872 routing which is based on the source as well as the destination 873 address of each packet. While greatly improving the 'intelligence' 874 of routing decisions within the homenet, such an approach would 875 require relatively significant router changes but avoid host changes. 877 As explained previously, while NPTv6 has been proposed for providing 878 multi-homing support in networks, its use is not recommended in the 879 homenet architecture. 881 It should be noted that some multihoming scenarios may see one 882 upstream being a "walled garden", and thus only appropriate for 883 connectivity to the services of that provider; an example may be a 884 VPN service that only routes back to the enterprise business network 885 of a user in the homenet. As per [RFC3002] (section 4.2.1) we do not 886 specifically target walled garden multihoming as a goal of this 887 document. 889 The homenet architecture should also not preclude use of host or 890 application-oriented tools, e.g., Shim6 [RFC5533], MPTCP [RFC6824] or 891 Happy Eyeballs [RFC6555]. In general, any incremental improvements 892 obtained by host changes should give benefit for the hosts 893 introducing them, but not be required. 895 3.3. A Self-Organising Network 897 The home network infrastructure should be naturally self-organising 898 and self-configuring under different circumstances relating to the 899 connectivity status to the Internet, number of devices, and physical 900 topology. At the same time, it should be possible for advanced users 901 to manually adjust (override) the current configuration. 903 While a goal of the homenet architecture is for the network to be as 904 self-organising as possible, there may be instances where some manual 905 configuration is required, e.g., the entry of a cryptographic key to 906 apply wireless security, or to configure a shared routing secret. 907 The latter may be relevant when considering how to bootstrap a 908 routing configuration. It is highly desirable that the number of 909 such configurations is minimised. 911 3.3.1. Differentiating neighbouring homenets 913 It is important that self-configuration with 'unintended' devices is 914 avoided. There should be a way for a user to administratively assert 915 in a simple way whether or not a device belongs to a given homenet. 916 The goal is to allow the establishment of borders, particularly 917 between two adjacent homenets, and to avoid unauthorised devices from 918 participating in the homenet. Such an authorisation capability may 919 need to operate through multiple hops in the homenet. 921 The homenet should thus support a way for a homenet owner to claim 922 ownership of their devices in a reasonably secure way. This could be 923 achieved by a pairing mechanism, by for example pressing buttons 924 simultaneously on an authenticated and a new homenet device, or by an 925 enrolment process as part of an autonomic networking environment. 927 While there may be scenarios where one homenet may wish to 928 intentionally gain access through another, e.g. to share external 929 connectivity costs, such scenarios are not discussed in this 930 document. 932 3.3.2. Largest practical subnets 934 Today's IPv4 home networks generally have a single subnet, and early 935 dual-stack deployments have a single congruent IPv6 subnet, possibly 936 with some bridging functionality. More recently, some vendors have 937 started to introduce 'home' and 'guest' functions, which in IPv6 938 would be implemented as two subnets. 940 Future home networks are highly likely to have one or more internal 941 routers and thus need multiple subnets, for the reasons described 942 earlier. As part of the self-organisation of the network, the 943 homenet should subdivide itself into the largest practical subnets 944 that can be constructed within the constraints of link layer 945 mechanisms, bridging, physical connectivity, and policy, and where 946 applicable performance or other criteria. In such subdivisions the 947 logical topology may not necessarily match the physical topology. 948 This text does not, however, make recommendations on how such 949 subdivision should occur. It is expected that subsequent documents 950 will address this problem. 952 While it may be desirable to maximise the chance of link-local 953 protocols operating across a homenet by maximising the size of a 954 subnet, multi-subnet home networks are inevitable, so their support 955 must be included. 957 3.3.3. Handling varying link technologies 959 Homenets tend to grow organically over many years, and a homenet will 960 typically be built over link-layer technologies from different 961 generations. Current homenets typically use links ranging from 962 1Mbit/s up to 1Gbit/s, which is a three orders of magnitude 963 throughput discrepancy. We expect this discrepancy to widen further 964 as both high-speed and low-power technologies are deployed. 966 Homenet protocols should be designed to deal well with 967 interconnecting links of very different throughputs. In particular, 968 flows local to a link should not be flooded throughout the homenet, 969 even when sent over multicast, and, whenever possible, the homenet 970 protocols should be able to choose the faster links and avoid the 971 slower ones. 973 Links (particularly wireless links) may also have limited numbers of 974 transmit opportunities (txops), and there is a clear trend driven by 975 both power and downward compatibility constraints toward aggregation 976 of packets into these limited txops while increasing throughput. 977 Transmit opportunities may be a system's scarcest resource and 978 therefore also strongly limit actual throughput available. 980 3.3.4. Homenet realms and borders 982 The homenet will need to be aware of the extent of its own 'site', 983 which will, for example, define the borders for ULA and site scope 984 multicast traffic, and may require specific security policies to be 985 applied. The homenet will have one or more such borders with 986 external connectivity providers. 988 A homenet will most likely also have internal borders between 989 internal realms, e.g., a guest realm or a corporate network extension 990 realm. It is desirable that appropriate borders can be configured to 991 determine, for example, the scope of where network prefixes, routing 992 information, network traffic, service discovery and naming may be 993 shared. The default mode internally should be to share everything. 995 It is expected that a realm would span at least an entire subnet, and 996 thus the borders lie at routers which receive delegated prefixes 997 within the homenet. It is also desirable, for a richer security 998 model, that hosts are able to make communication decisions based on 999 available realm and associated prefix information in the same way 1000 that routers at realm borders can. 1002 A simple homenet model may just consider three types of realm and the 1003 borders between them, namely the internal homenet, the ISP and a 1004 guest network. In this case the borders will include that from the 1005 homenet to the ISP, that from the guest network to the ISP, and that 1006 from the homenet to the guest network. Regardless, it should be 1007 possible for additional types of realms and borders to be defined, 1008 e.g., for some specific LLN-based network, such as Smart Grid, and 1009 for these to be detected automatically, and for an appropriate 1010 default policy to be applied as to what type of traffic/data can flow 1011 across such borders. 1013 It is desirable to classify the external border of the home network 1014 as a unique logical interface separating the home network from 1015 service provider network/s. This border interface may be a single 1016 physical interface to a single service provider, multiple layer 2 1017 sub-interfaces to a single service provider, or multiple connections 1018 to a single or multiple providers. This border makes it possible to 1019 describe edge operations and interface requirements across multiple 1020 functional areas including security, routing, service discovery, and 1021 router discovery. 1023 It should be possible for the homenet user to override any 1024 automatically determined borders and the default policies applied 1025 between them, the exception being that it may not be possible to 1026 override policies defined by the ISP at the external border. 1028 3.3.5. Configuration information from the ISP 1030 In certain cases, it may be useful for the homenet to get certain 1031 configuration information from its ISP. For example, the homenet 1032 DHCP server may request and forward some options that it gets from 1033 its upstream DHCP server, though the specifics of the options may 1034 vary across deployments. There is potential complexity here of 1035 course should the homenet be multihomed. 1037 3.4. Homenet Addressing 1039 The IPv6 addressing scheme used within a homenet must conform to the 1040 IPv6 addressing architecture [RFC4291]. In this section we discuss 1041 how the homenet needs to adapt to the prefixes made available to it 1042 by its upstream ISP, such that internal subnets, hosts and devices 1043 can obtain the and configure the necessary addressing information to 1044 operate. 1046 3.4.1. Use of ISP-delegated IPv6 prefixes 1048 Discussion of IPv6 prefix allocation policies is included in 1049 [RFC6177]. In practice, a homenet may receive an arbitrary length 1050 IPv6 prefix from its provider, e.g., /60, /56 or /48. The offered 1051 prefix may be stable or change from time to time; it is generally 1052 expected that ISPs will offer relatively stable prefixes to their 1053 residential customers. Regardless, the home network needs to be 1054 adaptable as far as possible to ISP prefix allocation policies, and 1055 thus make no assumptions about the stability of the prefix received 1056 from an ISP, or the length of the prefix that may be offered. 1058 However, if, for example, only a /64 is offered by the ISP, the 1059 homenet may be severely constrained or even unable to function. 1060 [RFC6177] (BCP 157) states that "a key principle for address 1061 management is that end sites always be able to obtain a reasonable 1062 amount of address space for their actual and planned usage, and over 1063 time ranges specified in years rather than just months. In practice, 1064 that means at least one /64, and in most cases significantly more. 1065 One particular situation that must be avoided is having an end site 1066 feel compelled to use IPv6-to-IPv6 Network Address Translation or 1067 other burdensome address conservation techniques because it could not 1068 get sufficient address space." This architecture document assumes 1069 that the guidance in the quoted text is being followed by ISPs. 1071 There are many problems that would arise from a homenet not being 1072 offered a sufficient prefix size for its needs. Rather than attempt 1073 to contrive a method for a homenet to operate in a constrained manner 1074 when faced with insufficient prefixes, such as the use of subnet 1075 prefixes longer than /64 (which would break stateless address 1076 autoconfiguration [RFC4862]), use of NPTv6, or falling back to 1077 bridging across potentially very different media, it is recommended 1078 that the receiving router instead enters an error state and issues 1079 appropriate warnings. Some consideration may need to be given to how 1080 such a warning or error state should best be presented to a typical 1081 home user. 1083 Thus a homenet CER should request, for example via DHCP Prefix 1084 Delegation (DHCP PD) [RFC3633], that it would like a /48 prefix from 1085 its ISP, i.e., it asks the ISP for the maximum size prefix it might 1086 expect to be offered, even if in practice it may only be offered a 1087 /56 or /60. For a typical IPv6 homenet, it is not recommended that 1088 an ISP offer less than a /60 prefix, and it is highly preferable that 1089 the ISP offers at least a /56. It is expected that the allocated 1090 prefix to the homenet from any single ISP is a contiguous, aggregated 1091 one. While it may be possible for a homenet CER to issue multiple 1092 prefix requests to attempt to obtain multiple delegations, such 1093 behaviour is out of scope of this document. 1095 The norm for residential customers of large ISPs may be similar to 1096 their single IPv4 address provision; by default it is likely to 1097 remain persistent for some time, but changes in the ISP's own 1098 provisioning systems may lead to the customer's IP (and in the IPv6 1099 case their prefix pool) changing. It is not expected that ISPs will 1100 generally support Provider Independent (PI) addressing for 1101 residential homenets. 1103 When an ISP does need to restructure, and in doing so renumber its 1104 customer homenets, 'flash' renumbering is likely to be imposed. This 1105 implies a need for the homenet to be able to handle a sudden 1106 renumbering event which, unlike the process described in [RFC4192], 1107 would be a 'flag day" event, which means that a graceful renumbering 1108 process moving through a state with two active prefixes in use would 1109 not be possible. While renumbering can be viewed as an extended 1110 version of an initial numbering process, the difference between flash 1111 renumbering and an initial 'cold start' is the need to provide 1112 service continuity. 1114 There may be cases where local law means some ISPs are required to 1115 change IPv6 prefixes (current IPv4 addresses) for privacy reasons for 1116 their customers. In such cases it may be possible to avoid an 1117 instant 'flash' renumbering and plan a non-flag day renumbering as 1118 per RFC 4192. Similarly, if an ISP has a planned renumbering 1119 process, it may be able to adjust lease timers, etc appropriately. 1121 The customer may of course also choose to move to a new ISP, and thus 1122 begin using a new prefix. In such cases the customer should expect a 1123 discontinuity, and not only may the prefix change, but potentially 1124 also the prefix length if the new ISP offers a different default size 1125 prefix. The homenet may also be forced to renumber itself if 1126 significant internal 'replumbing' is undertaken by the user. 1127 Regardless, it's desirable that homenet protocols support rapid 1128 renumbering and that operational processes don't add unnecessary 1129 complexity for the renumbering process. Further, the introduction of 1130 any new homenet protocols should not make any form of renumbering any 1131 more complex than it already is. 1133 Finally, the internal operation of the home network should also not 1134 depend on the availability of the ISP network at any given time, 1135 other than of course for connectivity to services or systems off the 1136 home network. This reinforces the use of ULAs for stable internal 1137 communication, and the need for a naming and service discovery 1138 mechanism that can operate independently within the homenet. 1140 3.4.2. Stable internal IP addresses 1142 The network should by default attempt to provide IP-layer 1143 connectivity between all internal parts of the homenet as well as to 1144 and from the external Internet, subject to the filtering policies or 1145 other policy constraints discussed later in the security section. 1147 ULAs should be used within the scope of a homenet to support stable 1148 routing and connectivity between subnets and hosts regardless of 1149 whether a globally unique ISP-provided prefix is available. In the 1150 case of a prolonged external connectivity outage, ULAs allow internal 1151 operations across routed subnets to continue. ULA addresses also 1152 allow constrained devices to create permanent relationships between 1153 IPv6 addresses, e.g., from a wall controller to a lamp, where 1154 symbolic host names would require additional non-volatile memory and 1155 updating global prefixes in sleeping devices might also be 1156 problematic. 1158 As discussed previously, it would be expected that ULAs would 1159 normally be used alongside one or more global prefixes in a homenet, 1160 such that hosts become multi-addressed with both globally unique and 1161 ULA prefixes. ULAs should be used for all devices, not just those 1162 intended to only have internal connectivity. Default address 1163 selection would then enable ULAs to be preferred for internal 1164 communications between devices that are using ULA prefixes generated 1165 within the same homenet. 1167 In cases where ULA prefixes are in use within a homenet but there is 1168 no external IPv6 connectivity (and thus no GUAs in use), 1169 recommendations ULA-5, L-3 and L-4 in RFC 6204 should be followed to 1170 ensure correct operation, in particular where the homenet may be 1171 dual-stack with IPv4 external connectivity. The use of the Route 1172 Information Option described in [RFC4191] provides a mechanism to 1173 advertise such more-specific ULA routes. 1175 The use of ULAs should be restricted to the homenet scope through 1176 filtering at the border(s) of the homenet, as mandated by RFC 6204 1177 requirement S-2. 1179 Note that it is possible that in some cases multiple /48 ULA prefixes 1180 may be in use within the same homenet, e.g., when the network is 1181 being deployed, perhaps also without external connectivity. In cases 1182 where multiple ULA /48's are in use, hosts need to know that each /48 1183 is local to the homenet, e.g., by inclusion in their local address 1184 selection policy table. 1186 3.4.3. Internal prefix delegation 1188 As mentioned above, there are various sources of prefixes. From the 1189 homenet perspective, a single global prefix from each ISP should be 1190 received on the border CER [RFC3633]. Where multiple CERs exist with 1191 multiple ISP prefix pools, it is expected that routers within the 1192 homenet would assign themselves prefixes from each ISP they 1193 communicate with/through. As discussed above, a ULA prefix should be 1194 provisioned for stable internal communications or for use on 1195 constrained/LLN networks. 1197 The delegation or availability of a prefix pool to the homenet should 1198 allow subsequent internal autonomous delegation of prefixes for use 1199 within the homenet. Such internal delegation should not assume a 1200 flat or hierarchical model, nor should it make an assumption about 1201 whether the delegation of internal prefixes is distributed or 1202 centralised. The assignment mechanism should provide reasonable 1203 efficiency, so that typical home network prefix allocation sizes can 1204 accommodate all the necessary /64 allocations in most cases, and not 1205 waste prefixes. Further, duplicate assignment of multiple /64s to 1206 the same network should be avoided, and the network should behave as 1207 gracefully as possible in the event of prefix exhaustion (though the 1208 options in such cases may be limited). 1210 Where the home network has multiple CERs and these are delegated 1211 prefix pools from their attached ISPs, the internal prefix delegation 1212 would be expected to be served by each CER for each prefix associated 1213 with it. Where ULAs are used, it is preferable that only one /48 ULA 1214 covers the whole homenet, from which /64's can be delegated to the 1215 subnets. In cases where two /48 ULAs are generated within a homenet, 1216 the network should still continue to function, meaning that hosts 1217 will need to determine that each ULA is local to the homenet. 1219 Delegation within the homenet should result in each link being 1220 assigned a stable prefix that is persistent across reboots, power 1221 outages and similar short-term outages. The availability of 1222 persistent prefixes should not depend on the router boot order. The 1223 addition of a new routing device should not affect existing 1224 persistent prefixes, but persistence may not be expected in the face 1225 of significant 'replumbing' of the homenet. However, delegated ULA 1226 prefixes within the homenet should remain persistent through an ISP- 1227 driven renumbering event. 1229 Provisioning such persistent prefixes may imply the need for stable 1230 storage on routing devices, and also a method for a home user to 1231 'reset' the stored prefix should a significant reconfiguration be 1232 required (though ideally the home user should not be involved at 1233 all). 1235 This document makes no specific recommendation towards solutions, but 1236 notes that it is very likely that all routing devices participating 1237 in a homenet must use the same internal prefix delegation method. 1238 This implies that only one delegation method should be in use. 1240 3.4.4. Coordination of configuration information 1242 The network elements will need to be integrated in a way that takes 1243 account of the various lifetimes on timers that are used on different 1244 elements, e.g., DHCPv6 PD, router, valid prefix and preferred prefix 1245 timers. 1247 3.4.5. Privacy 1249 If ISPs offer relatively stable IPv6 prefixes to customers, the 1250 network prefix part of addresses associated with the homenet may not 1251 change over a reasonably long period of time. 1253 The exposure of which traffic is sourced from the same homenet is 1254 thus similar to IPv4; the single IPv4 global address seen through use 1255 of IPv4 NAT gives the same hint as the global IPv6 prefix seen for 1256 IPv6 traffic. 1258 While IPv4 NAT may obfuscate to an external observer which internal 1259 devices traffic is sourced from, IPv6, even with use of Privacy 1260 Addresses [RFC4941], adds additional exposure of which traffic is 1261 sourced from the same internal device, through use of the same IPv6 1262 source address for a period of time. 1264 3.5. Routing functionality 1266 Routing functionality is required when there are multiple routers 1267 deployed within the internal home network. This functionality could 1268 be as simple as the current 'default route is up' model of IPv4 NAT, 1269 or, more likely, it would involve running an appropriate routing 1270 protocol. Regardless of the solution method, the functionality 1271 discussed below should be met. 1273 The homenet unicast routing protocol should be based on a previously 1274 deployed protocol that has been shown to be reliable and robust, and 1275 that allows lightweight implementations. The availability of open 1276 source implementations is an important consideration. It is 1277 desirable, but not absolutely required, that the routing protocol be 1278 able to give a complete view of the network, and that it be able to 1279 pass around more than just routing information. 1281 Multiple types of physical interfaces must be accounted for in the 1282 homenet routed topology. Technologies such as Ethernet, WiFi, 1283 Multimedia over Coax Alliance (MoCA), etc. must be capable of 1284 coexisting in the same environment and should be treated as part of 1285 any routed deployment. The inclusion of physical layer 1286 characteristics including bandwidth, loss, and latency in path 1287 computation should be considered for optimising communication in the 1288 homenet. 1290 The routing protocol should support the generic use of multiple 1291 customer Internet connections, and the concurrent use of multiple 1292 delegated prefixes. A routing protocol that can make routing 1293 decisions based on source and destination addresses is thus 1294 desirable, to avoid upstream ISP BCP38 ingress filtering problems. 1295 Multihoming support should also include load-balancing to multiple 1296 providers, and failover from a primary to a backup link when 1297 available. The protocol however should not require upstream ISP 1298 connectivity to be established to continue routing within the 1299 homenet. 1301 The routing environment should be self-configuring, as discussed 1302 previously. An example of how OSPFv3 can be self-configuring in a 1303 homenet is described in [I-D.ietf-ospf-ospfv3-autoconfig]. 1304 Minimising convergence time should be a goal in any routed 1305 environment, but as a guideline a maximum convergence time at most 30 1306 seconds should be the target (this target is somewhat arbitrary, and 1307 was chosen based on how long a typical home user might wait before 1308 attempting another reset; ideally the routers might have some status 1309 light indicating they are converging, similar to an ADSL router light 1310 indicating it is establishing a connection to its ISP). 1312 As per prefix delegation, it is assumed that a single routing 1313 solution is in use in the homenet architecture. If there is an 1314 identified need to support multiple solutions, these must be 1315 interoperable. 1317 An appropriate mechanism is required to discover which router(s) in 1318 the homenet are providing the CER function. Borders may include but 1319 are not limited to the interface to the upstream ISP, a gateway 1320 device to a separate home network such as a LLN network, or a gateway 1321 to a guest or private corporate extension network. In some cases 1322 there may be no border present, which may for example occur before an 1323 upstream connection has been established. The border discovery 1324 functionality may be integrated into the routing protocol itself, but 1325 may also be imported via a separate discovery mechanism. 1327 In general, LLN or other networks should be able to attach and 1328 participate in the same way as the main homenet, or alternatively 1329 map/be gatewayed to the main homenet. Current home deployments use 1330 largely different mechanisms in sensor and basic Internet 1331 connectivity networks. IPv6 virtual machine (VM) solutions may also 1332 add additional routing requirements. 1334 3.5.1. Multicast support 1336 It is desirable that, subject to the capacities of devices on certain 1337 media types, multicast routing is supported across the homenet. 1339 [RFC4291] requires that any boundary of scope 4 or higher (i.e., 1340 admin-local or higher) be administratively configured. Thus the 1341 boundary at the homenet-ISP border must be administratively 1342 configured, though that may be triggered by an administrative 1343 function such as DHCP-PD. Other multicast forwarding policy borders 1344 may also exist within the homenet, e.g., to/from a guest subnet, 1345 whilst the use of certain media types may also affect where specific 1346 multicast traffic is forwarded or routed. 1348 There may be different drivers for multicast to be supported across 1349 the homenet, e.g., for homenet-wide service discovery should a 1350 multicast service discovery protocol of scope greater than link-local 1351 be defined, or potentially for multicast-based streaming or 1352 filesharing applications. Where multicast is routed across a homenet 1353 an appropriate multicast routing protocol is required, one that as 1354 per the unicast routing protocol should be self-configuring. As 1355 hinted above, it must be possible to scope or filter multicast 1356 traffic to avoid it being flooded to network media where devices 1357 cannot reasonably support it. 1359 A homenet may not only use multicast internally, it may also be a 1360 consumer or provider of external multicast traffic, where the 1361 homenet's ISP supports such multicast operation. This may be 1362 valuable for example where live video applications are being sourced 1363 to/from the homenet. 1365 The multicast environment should support the ability for applications 1366 to pick a unique multicast group to use. 1368 3.5.2. Mobility support 1370 Devices may be mobile within the homenet. While resident on the same 1371 subnet, their address will remain persistent, but should devices move 1372 to a different (wireless) subnet, they will acquire a new address in 1373 that subnet. It is desirable that the homenet supports internal 1374 device mobility. To do so, the homenet may either extend the reach 1375 of specific wireless subnets to enable wireless roaming across the 1376 home (availability of a specific subnet across the home), or it may 1377 support mobility protocols to facilitate such roaming where multiple 1378 subnets are used. 1380 3.6. Security 1382 The security of an IPv6 homenet is an important consideration. The 1383 most notable difference to the IPv4 operational model is the removal 1384 of NAT, the introduction of global addressability of devices, and 1385 thus a need to consider whether devices should have global 1386 reachability. Regardless, hosts need to be able to operate securely, 1387 end-to-end where required, and also be robust against malicious 1388 traffic directed towards them. However, there are other challenges 1389 introduced, e.g., default filtering policies at the borders between 1390 various homenet realms. 1392 3.6.1. Addressability vs reachability 1394 An IPv6-based home network architecture should embrace the 1395 transparent end-to-end communications model as described in 1396 [RFC2775]. Each device should be globally addressable, and those 1397 addresses must not be altered in transit. However, security 1398 perimeters can be applied to restrict end-to-end communications, and 1399 thus while a host may be globally addressable it may not be globally 1400 reachable. 1402 [RFC4864] describes a 'Simple Security' model for IPv6 networks, 1403 whereby stateful perimeter filtering can be applied to control the 1404 reachability of devices in a homenet. RFC 4864 states in Section 4.2 1405 that "the use of firewalls ... is recommended for those that want 1406 boundary protection in addition to host defences". It should be 1407 noted that a 'default deny' filtering approach would effectively 1408 replace the need for IPv4 NAT traversal protocols with a need to use 1409 a signalling protocol to request a firewall hole be opened, e.g., a 1410 protocol such as PCP [RFC6887]. In networks with multiple CERs, the 1411 signalling would need to handle the cases of flows that may use one 1412 or more exit routers. CERs would need to be able to advertise their 1413 existence for such protocols. 1415 [RFC6092] expands on RFC 4864, giving a more detailed discussion of 1416 IPv6 perimeter security recommendations, without mandating a 'default 1417 deny' approach. Indeed, RFC 6092 does not enforce a particular mode 1418 of operation, instead stating that CERs must provide an easily 1419 selected configuration option that permits a 'transparent' mode, thus 1420 ensuring a 'default allow' model is available. The homenet 1421 architecture text makes no recommendation on the default setting, and 1422 refers the reader to RFC 6092. 1424 3.6.2. Filtering at borders 1426 It is desirable that there are mechanisms to detect different types 1427 of borders within the homenet, as discussed previously, and further 1428 mechanisms to then apply different types of filtering policies at 1429 those borders, e.g., whether naming and service discovery should pass 1430 a given border. Any such policies should be able to be easily 1431 applied by typical home users, e.g., to give a user in a guest 1432 network access to media services in the home, or access to a printer. 1433 Simple mechanisms to apply policy changes, or associations between 1434 devices, will be required. 1436 There are cases where full internal connectivity may not be 1437 desirable, e.g., in certain utility networking scenarios, or where 1438 filtering is required for policy reasons against guest network 1439 subnet(s). Some scenarios/models may as a result involve running 1440 isolated subnet(s) with their own CERs. In such cases connectivity 1441 would only be expected within each isolated network (though traffic 1442 may potentially pass between them via external providers). 1444 LLNs provide an another example of where there may be secure 1445 perimeters inside the homenet. Constrained LLN nodes may implement 1446 network key security but may depend on access policies enforced by 1447 the LLN border router. 1449 Considerations for differentiating neighbouring homenets are 1450 discussed in Section 3.3.1. 1452 3.6.3. Partial Effectiveness of NAT and Firewalls 1454 Security by way of obscurity (address translation) or through 1455 firewalls (filtering) is at best only partially effective. The very 1456 poor security track record of home computer, home networking and 1457 business PC computers and networking is testimony to this. A 1458 security compromise behind the firewall of any device exposes all 1459 others, making an entire network that relies on obscurity or a 1460 firewall as vulnerable as the most insecure device on the private 1461 side of the network. 1463 However, given current evidence of home network products with very 1464 poor default device security, putting a firewall in place does 1465 provide some level of protection. The use of firewalls today, 1466 whether a good practice or not, is common practice and whatever 1467 protection afforded, even if marginally effective, should not be 1468 lost. Thus, while it is highly desirable that all hosts in a homenet 1469 be adequately protected by built-in security functions, it should 1470 also be assumed that all CERs will continue to support appropriate 1471 perimeter defence functions, as per [RFC7084]. 1473 3.6.4. Exfiltration concerns 1475 As homenets become more complex, with more devices, and with service 1476 discovery potentially enabled across the whole home, there are 1477 potential concerns over the leakage of information should devices use 1478 discovery protocols to gather information and report it to equipment 1479 vendors or application service providers. 1481 While it is not clear how such exfiltration could be easily avoided, 1482 the threat should be recognised, be it from a new piece of hardware 1483 or some 'app' installed on a personal device. 1485 3.6.5. Device capabilities 1487 In terms of the devices, homenet hosts should implement their own 1488 security policies in accordance to their computing capabilities. 1489 They should have the means to request transparent communications to 1490 be able to be initiated to them through security filters in the 1491 homenet, either for all ports or for specific services. Users should 1492 have simple methods to associate devices to services that they wish 1493 to operate transparently through (CER) borders. 1495 3.6.6. ULAs as a hint of connection origin 1497 As noted in Section 3.6, if appropriate filtering is in place on the 1498 CER(s), as mandated by RFC 6204 requirement S-2, a ULA source address 1499 may be taken as an indication of locally sourced traffic. This 1500 indication could then be used with security settings to designate 1501 between which nodes a particular application is allowed to 1502 communicate, provided ULA address space is filtered appropriately at 1503 the boundary of the realm. 1505 3.7. Naming and Service Discovery 1507 The homenet requires devices to be able to determine and use unique 1508 names by which they can be accessed on the network. Users and 1509 devices will need to be able to discover devices and services 1510 available on the network, e.g., media servers, printers, displays or 1511 specific home automation devices. Thus naming and service discovery 1512 must be supported in the homenet, and, given the nature of typical 1513 home network users, the service(s) providing this function must as 1514 far as possible support unmanaged operation. 1516 The naming system will be required to work internally or externally, 1517 be the user within the homenet or outside it, i.e., the user should 1518 be able to refer to devices by name, and potentially connect to them, 1519 wherever they may be. The most natural way to think about such 1520 naming and service discovery is to enable it to work across the 1521 entire homenet residence (site), disregarding technical borders such 1522 as subnets but respecting policy borders such as those between guest 1523 and other internal network realms. Remote access may be desired by 1524 the homenet residents while travelling, but also potentially by 1525 manufacturers or other 'benevolent' third parties. 1527 3.7.1. Discovering services 1529 Users will typically perform service discovery through graphical user 1530 interfaces (GUIs) that allow them to browse services on their network 1531 in an appropriate and intuitive way. Devices may also need to 1532 discover other devices, without any user intervention or choice. 1534 Either way, such interfaces are beyond the scope of this document, 1535 but the interface should have an appropriate application programming 1536 interface (API) for the discovery to be performed. 1538 Such interfaces may also typically hide the local domain name element 1539 from users, especially where only one name space is available. 1540 However, as we discuss below, in some cases the ability to discover 1541 available domains may be useful. 1543 We note that current zero-configuration service discovery protocols 1544 are generally aimed at single subnets. There is thus a choice to 1545 make for multi-subnet homenets as to whether such protocols should be 1546 proxied or extended to operate across a whole homenet. In this 1547 context, that may mean bridging a link-local method, taking care to 1548 avoid loops, or extending the scope of multicast traffic used for the 1549 purpose. It may mean that some proxy or hybrid service is utilised, 1550 perhaps co-resident on the CER. Or it may be that a new approach is 1551 preferable, e.g., flooding information around the homenet as 1552 attributes within the routing protocol (which could allow per-prefix 1553 configuration). However, we should prefer approaches that are 1554 backwardly compatible, and allow current implementations to continue 1555 to be used. Note that this document does not mandate a particular 1556 solution, rather it expresses the principles that should be used for 1557 a homenet naming and service discovery environment. 1559 One of the primary challenges facing service discovery today is lack 1560 of interoperability due to the ever increasing number of service 1561 discovery protocols available. While it is conceivable for consumer 1562 devices to support multiple discovery protocols, this is clearly not 1563 the most efficient use of network and computational resources. One 1564 goal of the homenet architecture should be a path to service 1565 discovery protocol interoperability either through a standards based 1566 translation scheme, hooks into current protocols to allow some for of 1567 communication among discovery protocols, extensions to support a 1568 central service repository in the homenet, or simply convergence 1569 towards a unified protocol suite. 1571 3.7.2. Assigning names to devices 1573 Given the large number of devices that may be networked in the 1574 future, devices should have a means to generate their own unique 1575 names within a homenet, and to detect clashes should they arise, 1576 e.g., where a second device of the same type/vendor as an existing 1577 device with the same default name is deployed, or where a new subnet 1578 is added to the homenet which already has a device of the same name. 1579 It is expected that a device should have a fixed name while within 1580 the scope of the homenet. 1582 Users will also want simple ways to (re)name devices, again most 1583 likely through an appropriate and intuitive interface that is beyond 1584 the scope of this document. Note the name a user assigns to a device 1585 may be a label that is stored on the device as an attribute of the 1586 device, and may be distinct from the name used in a name service, 1587 e.g., 'Study Laser Printer' as opposed to printer2.. 1589 3.7.3. The homenet name service 1591 The homenet name service should support both lookups and discovery. 1592 A lookup would operate via a direct query to a known service, while 1593 discovery may use multicast messages or a service where applications 1594 register in order to be found. 1596 It is highly desirable that the homenet name service must at the very 1597 least co-exist with the Internet name service. There should also be 1598 a bias towards proven, existing solutions. The strong implication is 1599 thus that the homenet service is DNS-based, or DNS-compatible. There 1600 are naming protocols that are designed to be configured and operate 1601 Internet-wide, like unicast-based DNS, but also protocols that are 1602 designed for zero-configuration local environments, like mDNS 1603 [RFC6762]. 1605 When DNS is used as the homenet name service, it typically includes 1606 both a resolving service and an authoritative service. The 1607 authoritative service hosts the homenet related zone. One approach 1608 when provisioning such a name service, which is designed to 1609 facilitate name resolution from the global Internet, is to run an 1610 authoritative name service on the CER and a secondary authoritative 1611 name service provided by the ISP or perhaps an external third party. 1613 Where zero configuration name services are used, it is desirable that 1614 these can also coexist with the Internet name service. In 1615 particular, where the homenet is using a global name space, it is 1616 desirable that devices have the ability, where desired, to add 1617 entries to that name space. There should also be a mechanism for 1618 such entries to be removed or expired from the global name space. 1620 To protect against attacks such as cache poisoning, where an attacker 1621 is able to insert a bogus DNS entry in the local cache, it is 1622 desirable to support appropriate name service security methods, 1623 including DNS Security Extensions (DNSSEC) [RFC4033], on both the 1624 authoritative server and the resolver sides. Where DNS is used, the 1625 homenet router or naming service must not prevent DNSSEC from 1626 operating. 1628 While this document does not specify hardware requirements, it is 1629 worth noting briefly here that e.g., in support of DNSSEC, 1630 appropriate homenet devices should have good random number generation 1631 capability, and future homenet specifications should indicate where 1632 high quality random number generators, i.e., with decent entropy, are 1633 needed. 1635 Finally, the impact of a change in CER must be considered. It would 1636 be desirable to retain any relevant state (configuration) that was 1637 held in the old CER. This might imply that state information should 1638 be distributed in the homenet, to be recoverable by/to the new CER, 1639 or to the homenet's ISP or a third party externally provided service 1640 by some means. 1642 3.7.4. Name spaces 1644 If access to homenet devices is required remotely from anywhere on 1645 the Internet, then at least one globally unique name space is 1646 required, though the use of multiple name spaces should not be 1647 precluded. One approach is that the name space(s) used for the 1648 homenet would be served authoritatively by the homenet, most likely 1649 by a server resident on the CER. Such name spaces may be acquired by 1650 the user or provided/generated by their ISP or an alternative 1651 externally provided service. It is likely that the default case is 1652 that a homenet will use a global domain provided by the ISP, but 1653 advanced users wishing to use a name space that is independent of 1654 their provider in the longer term should be able to acquire and use 1655 their own domain name. For users wanting to use their own 1656 independent domain names, such services are already available. 1658 Devices may also be assigned different names in different name 1659 spaces, e.g., by third parties who may manage systems or devices in 1660 the homenet on behalf of the resident(s). Remote management of the 1661 homenet is out of scope of this document. 1663 If however a global name space is not available, the homenet will 1664 need to pick and use a local name space which would only have meaning 1665 within the local homenet (i.e., it would not be used for remote 1666 access to the homenet). The .local name space currently has a 1667 special meaning for certain existing protocols which have link-local 1668 scope, and is thus not appropriate for multi-subnet home networks. A 1669 different name space is thus required for the homenet. 1671 One approach for picking a local name space is to use an Ambiguous 1672 Local Qualified Domain Name (ALQDN) space, such as .sitelocal (or an 1673 appropriate name reserved for the purpose). While this is a simple 1674 approach, there is the potential in principle for devices that are 1675 bookmarked somehow by name by an application in one homenet to be 1676 confused with a device with the same name in another homenet. In 1677 practice however the underlying service discovery protocols should be 1678 capable of handling moving to a network where a new device is using 1679 the same name as a device used previously in another homenet. 1681 An alternative approach for a local name space would be to use a 1682 Unique Locally Qualified Domain Name (ULQDN) space such as 1683 ..sitelocal. The could be generated in 1684 a variety of ways, one potentially being based on the local /48 ULA 1685 prefix being used across the homenet. Such a should 1686 survive a cold restart, i.e., be consistent after a network power- 1687 down, or, if a value is not set on startup, the CER or device running 1688 the name service should generate a default value. It would be 1689 desirable for the homenet user to be able to override the 1690 with a value of their choice, but that would increase 1691 the likelihood of a name conflict. Any generated 1692 should not be predictable; thus adding a salt/hash function would be 1693 desirable. 1695 In the (likely) event that the homenet is accessible from outside the 1696 homenet (using the global name space), it is vital that the homenet 1697 name space follow the rules and conventions of the global name space. 1698 In this mode of operation, names in the homenet (including those 1699 automatically generated by devices) must be usable as labels in the 1700 global name space. [RFC5890] describes considerations for 1701 Internationalizing Domain Names in Applications (IDNA). 1703 Also, with the introduction of new 'dotless' top level domains, there 1704 is also potential for ambiguity between, for example, a local host 1705 called 'computer' and (if it is registered) a .computer gTLD. Thus 1706 qualified names should always be used, whether these are exposed to 1707 the user or not. The IAB has issued a statement which explains why 1708 dotless domains should be considered harmful [IABdotless]. 1710 There may be use cases where either different name spaces may be 1711 desired for different realms in the homenet, or for segmentation of a 1712 single name space within the homenet. Thus hierarchical name space 1713 management is likely to be required. There should also be nothing to 1714 prevent individual device(s) being independently registered in 1715 external name spaces. 1717 It may be the case that if there are two or more CERs serving the 1718 home network, that if each has name space delegated from a different 1719 ISP there is the potential for devices in the home to have multiple 1720 fully qualified names under multiple domains. 1722 Where a user is in a remote network wishing to access devices in 1723 their home network, there may be a requirement to consider the domain 1724 search order presented where multiple associated name spaces exist. 1725 This also implies that a domain discovery function is desirable. 1727 It may be the case that not all devices in the homenet are made 1728 available by name via an Internet name space, and that a 'split view' 1729 (as described in [RFC6950] Section 4) is preferred for certain 1730 devices, whereby devices inside the homenet see different DNS 1731 responses to those outside. 1733 Finally, this document makes no assumption about the presence or 1734 omission of a reverse lookup service. There is an argument that it 1735 may be useful for presenting logging information to users with 1736 meaningful device names rather than literal addresses. There are 1737 also some services, most notably email mail exchangers, where some 1738 operators have chosen to require a valid reverse lookup before 1739 accepting connections. 1741 3.7.5. Independent operation 1743 Name resolution and service discovery for reachable devices must 1744 continue to function if the local network is disconnected from the 1745 global Internet, e.g., a local media server should still be available 1746 even if the Internet link is down for an extended period. This 1747 implies the local network should also be able to perform a complete 1748 restart in the absence of external connectivity, and have local 1749 naming and service discovery operate correctly. 1751 The approach described above of a local authoritative name service 1752 with a cache would allow local operation for sustained ISP outages. 1754 Having an independent local trust anchor is desirable, to support 1755 secure exchanges should external connectivity be unavailable. 1757 A change in ISP should not affect local naming and service discovery. 1758 However, if the homenet uses a global name space provided by the ISP, 1759 then this will obviously have an impact if the user changes their 1760 network provider. 1762 3.7.6. Considerations for LLNs 1764 In some parts of the homenet, in particular LLNs or any devices where 1765 battery power is used, devices may be sleeping, in which case a proxy 1766 for such nodes may be required, that could respond (for example) to 1767 multicast service discovery requests. Those same devices or parts of 1768 the network may have less capacity for multicast traffic that may be 1769 flooded from other parts of the network. In general, message 1770 utilisation should be efficient considering the network technologies 1771 and constrained devices that the service may need to operate over. 1773 There are efforts underway to determine naming and discovery 1774 solutions for use by the Constrained Application Protocol (CoAP) 1776 [I-D.ietf-core-coap] in LLN networks. These are outside the scope of 1777 this document. 1779 3.7.7. DNS resolver discovery 1781 Automatic discovery of a name service to allow client devices in the 1782 homenet to resolve external domains on the Internet is required, and 1783 such discovery must support clients that may be a number of router 1784 hops away from the name service. Similarly it may be desirable to 1785 convey any DNS domain search list that may be in effect for the 1786 homenet. 1788 3.7.8. Devices roaming to/from the homenet 1790 It is likely that some devices which have registered names within the 1791 homenet Internet name space and that are mobile will attach to the 1792 Internet at other locations and acquire an IP address at those 1793 locations. Devices may move between different homenets. In such 1794 cases it is desirable that devices may be accessed by the same name 1795 as is used in their home network. 1797 Solutions to this problem are not discussed in this document. They 1798 may include use of Mobile IPv6 or Dynamic DNS, either of which would 1799 put additional requirements on to the homenet, or establishment of a 1800 (VPN) tunnel to a server in the home network. 1802 3.8. Other Considerations 1804 This section discusses two other considerations for home networking 1805 that the architecture should not preclude, but that this text is 1806 neutral towards. 1808 3.8.1. Quality of Service 1810 Support for Quality of Service in a multi-service homenet may be a 1811 requirement, e.g., for a critical system (perhaps healthcare 1812 related), or for differentiation between different types of traffic 1813 (file sharing, cloud storage, live streaming, VoIP, etc). Different 1814 media types may have different such properties or capabilities. 1816 However, homenet scenarios should require no new Quality of Service 1817 protocols. A DiffServ [RFC2475] approach with a small number of 1818 predefined traffic classes may generally be sufficient, though at 1819 present there is little experience of Quality of Service deployment 1820 in home networks. It is likely that QoS, or traffic prioritisation, 1821 methods will be required at the CER, and potentially around 1822 boundaries between different media types (where for example some 1823 traffic may simply not be appropriate for some media, and need to be 1824 dropped to avoid overloading the constrained media). 1826 There may also be complementary mechanisms that could be beneficial 1827 to application performance and behaviour in the homenet domain, such 1828 as ensuring proper buffering algorithms are used as described in 1829 [Gettys11]. 1831 3.8.2. Operations and Management 1833 The homenet should have the general goal of being self-organising and 1834 configuring, and thus the network elements should not need to be pro- 1835 actively managed by the home user. Thus specific protocols that may 1836 be available to manage the network are not discussed in this 1837 document. 1839 The network protocols used should, as far as possible, be able to 1840 self-configure, e.g. for prefixes to be assigned to router 1841 interfaces, or for devices to use zero-configuration protocols to 1842 discover services in the home. A home user would not be expected to, 1843 e.g., assign prefixes to links, or manage the DNS entries for the 1844 home network. Such expert operation should not be precluded, but it 1845 is not the norm. 1847 There may still be some configuration parameters which are exposed to 1848 users, e.g., SSID name(s), or wireless security key(s). Users may 1849 also be expected to be aware of the functions of certain devices they 1850 connect, e.g., which are providing a server function, and they may be 1851 able to assign 'friendly names' to those devices, though service 1852 discovery protocols should make their selection as intuitive as 1853 possible. 1855 As discussed in Section 3.6.1 the default setting on the homenet-ISP 1856 border for inbound traffic may be default deny, default allow, or 1857 some position inbetween. Whatever the default position, it should be 1858 possible for the user to change the setting. 1860 Users may also be interested in the status of their networks and 1861 devices on the network, in which case simple self-monitoring 1862 mechanisms would be desirable. This may be particularly important 1863 when some fault arises in the network or with a device. It may also 1864 be the case that an ISP, or a third party, might offer management of 1865 the homenet on behalf of a user, in which case management protocols 1866 would be required. How either model of management and monitoring is 1867 performed is out of scope of this document. It is expected that a 1868 separate document will follow to describe the operations and 1869 management model(s) for the types of home networks presented in this 1870 document. 1872 A final consideration is that all network management and monitoring 1873 functions should be available over IPv6 transport, even where the 1874 homenet is dual-stack. 1876 3.9. Implementing the Architecture on IPv6 1878 This architecture text encourages re-use of existing protocols. Thus 1879 the necessary mechanisms are largely already part of the IPv6 1880 protocol set and common implementations, though there are some 1881 exceptions. 1883 For automatic routing, it is expected that solutions can be found 1884 based on existing protocols. Some relatively smaller updates are 1885 likely to be required, e.g., a new mechanism may be needed in order 1886 to turn a selected protocol on by default, a mechanism may be 1887 required to automatically assign prefixes to links within the 1888 homenet. 1890 Some functionality, if required by the architecture, may need more 1891 significant changes or require development of new protocols, e.g., 1892 support for multihoming with multiple exit routers would likely 1893 require extensions to support source and destination address based 1894 routing within the homenet. 1896 Some protocol changes are however required in the architecture, e.g., 1897 for name resolution and service discovery, extensions to existing 1898 zero configuration link-local name resolution protocols are needed to 1899 enable them to work across subnets, within the scope of the home 1900 network site. 1902 Some of the hardest problems in developing solutions for home 1903 networking IPv6 architectures include discovering the right borders 1904 where the 'home' domain ends and the service provider domain begins, 1905 deciding whether some of the necessary discovery mechanism extensions 1906 should affect only the network infrastructure or also hosts, and the 1907 ability to turn on routing, prefix delegation and other functions in 1908 a backwards compatible manner. 1910 4. Conclusions 1912 This text defines principles and requirements for a homenet 1913 architecture. The principles and requirements documented here should 1914 be observed by any future texts describing homenet protocols for 1915 routing, prefix management, security, naming or service discovery. 1917 5. Security Considerations 1919 Security considerations for the homenet architecture are discussed in 1920 Section 3.6 above. 1922 6. IANA Considerations 1924 This document has no actions for IANA. 1926 7. References 1928 7.1. Normative References 1930 [RFC2460] Deering, S. and R. Hinden, "Internet Protocol, Version 6 1931 (IPv6) Specification", RFC 2460, December 1998. 1933 [RFC3633] Troan, O. and R. Droms, "IPv6 Prefix Options for Dynamic 1934 Host Configuration Protocol (DHCP) version 6", RFC 3633, 1935 December 2003. 1937 [RFC4193] Hinden, R. and B. Haberman, "Unique Local IPv6 Unicast 1938 Addresses", RFC 4193, October 2005. 1940 [RFC4291] Hinden, R. and S. Deering, "IP Version 6 Addressing 1941 Architecture", RFC 4291, February 2006. 1943 7.2. Informative References 1945 [RFC1918] Rekhter, Y., Moskowitz, R., Karrenberg, D., Groot, G., and 1946 E. Lear, "Address Allocation for Private Internets", 1947 BCP 5, RFC 1918, February 1996. 1949 [RFC2475] Blake, S., Black, D., Carlson, M., Davies, E., Wang, Z., 1950 and W. Weiss, "An Architecture for Differentiated 1951 Services", RFC 2475, December 1998. 1953 [RFC2775] Carpenter, B., "Internet Transparency", RFC 2775, 1954 February 2000. 1956 [RFC2827] Ferguson, P. and D. Senie, "Network Ingress Filtering: 1957 Defeating Denial of Service Attacks which employ IP Source 1958 Address Spoofing", BCP 38, RFC 2827, May 2000. 1960 [RFC3002] Mitzel, D., "Overview of 2000 IAB Wireless Internetworking 1961 Workshop", RFC 3002, December 2000. 1963 [RFC3022] Srisuresh, P. and K. Egevang, "Traditional IP Network 1964 Address Translator (Traditional NAT)", RFC 3022, 1965 January 2001. 1967 [RFC4033] Arends, R., Austein, R., Larson, M., Massey, D., and S. 1968 Rose, "DNS Security Introduction and Requirements", 1969 RFC 4033, March 2005. 1971 [RFC4191] Draves, R. and D. Thaler, "Default Router Preferences and 1972 More-Specific Routes", RFC 4191, November 2005. 1974 [RFC4192] Baker, F., Lear, E., and R. Droms, "Procedures for 1975 Renumbering an IPv6 Network without a Flag Day", RFC 4192, 1976 September 2005. 1978 [RFC4862] Thomson, S., Narten, T., and T. Jinmei, "IPv6 Stateless 1979 Address Autoconfiguration", RFC 4862, September 2007. 1981 [RFC4864] Van de Velde, G., Hain, T., Droms, R., Carpenter, B., and 1982 E. Klein, "Local Network Protection for IPv6", RFC 4864, 1983 May 2007. 1985 [RFC4941] Narten, T., Draves, R., and S. Krishnan, "Privacy 1986 Extensions for Stateless Address Autoconfiguration in 1987 IPv6", RFC 4941, September 2007. 1989 [RFC5533] Nordmark, E. and M. Bagnulo, "Shim6: Level 3 Multihoming 1990 Shim Protocol for IPv6", RFC 5533, June 2009. 1992 [RFC5890] Klensin, J., "Internationalized Domain Names for 1993 Applications (IDNA): Definitions and Document Framework", 1994 RFC 5890, August 2010. 1996 [RFC5969] Townsley, W. and O. Troan, "IPv6 Rapid Deployment on IPv4 1997 Infrastructures (6rd) -- Protocol Specification", 1998 RFC 5969, August 2010. 2000 [RFC6092] Woodyatt, J., "Recommended Simple Security Capabilities in 2001 Customer Premises Equipment (CPE) for Providing 2002 Residential IPv6 Internet Service", RFC 6092, 2003 January 2011. 2005 [RFC6144] Baker, F., Li, X., Bao, C., and K. Yin, "Framework for 2006 IPv4/IPv6 Translation", RFC 6144, April 2011. 2008 [RFC6145] Li, X., Bao, C., and F. Baker, "IP/ICMP Translation 2009 Algorithm", RFC 6145, April 2011. 2011 [RFC6177] Narten, T., Huston, G., and L. Roberts, "IPv6 Address 2012 Assignment to End Sites", BCP 157, RFC 6177, March 2011. 2014 [RFC6204] Singh, H., Beebee, W., Donley, C., Stark, B., and O. 2015 Troan, "Basic Requirements for IPv6 Customer Edge 2016 Routers", RFC 6204, April 2011. 2018 [RFC6296] Wasserman, M. and F. Baker, "IPv6-to-IPv6 Network Prefix 2019 Translation", RFC 6296, June 2011. 2021 [RFC6333] Durand, A., Droms, R., Woodyatt, J., and Y. Lee, "Dual- 2022 Stack Lite Broadband Deployments Following IPv4 2023 Exhaustion", RFC 6333, August 2011. 2025 [RFC6555] Wing, D. and A. Yourtchenko, "Happy Eyeballs: Success with 2026 Dual-Stack Hosts", RFC 6555, April 2012. 2028 [RFC6724] Thaler, D., Draves, R., Matsumoto, A., and T. Chown, 2029 "Default Address Selection for Internet Protocol Version 6 2030 (IPv6)", RFC 6724, September 2012. 2032 [RFC6762] Cheshire, S. and M. Krochmal, "Multicast DNS", RFC 6762, 2033 February 2013. 2035 [RFC6824] Ford, A., Raiciu, C., Handley, M., and O. Bonaventure, 2036 "TCP Extensions for Multipath Operation with Multiple 2037 Addresses", RFC 6824, January 2013. 2039 [RFC6887] Wing, D., Cheshire, S., Boucadair, M., Penno, R., and P. 2040 Selkirk, "Port Control Protocol (PCP)", RFC 6887, 2041 April 2013. 2043 [RFC6950] Peterson, J., Kolkman, O., Tschofenig, H., and B. Aboba, 2044 "Architectural Considerations on Application Features in 2045 the DNS", RFC 6950, October 2013. 2047 [RFC7084] Singh, H., Beebee, W., Donley, C., and B. Stark, "Basic 2048 Requirements for IPv6 Customer Edge Routers", RFC 7084, 2049 November 2013. 2051 [I-D.ietf-v6ops-ipv6-multihoming-without-ipv6nat] 2052 Troan, O., Miles, D., Matsushima, S., Okimoto, T., and D. 2053 Wing, "IPv6 Multihoming without Network Address 2054 Translation", 2055 draft-ietf-v6ops-ipv6-multihoming-without-ipv6nat-06 (work 2056 in progress), February 2014. 2058 [I-D.ietf-ospf-ospfv3-autoconfig] 2059 Lindem, A. and J. Arkko, "OSPFv3 Auto-Configuration", 2060 draft-ietf-ospf-ospfv3-autoconfig-06 (work in progress), 2061 February 2014. 2063 [I-D.ietf-core-coap] 2064 Shelby, Z., Hartke, K., and C. Bormann, "Constrained 2065 Application Protocol (CoAP)", draft-ietf-core-coap-18 2066 (work in progress), June 2013. 2068 [IABdotless] 2069 "IAB Statement: Dotless Domains Considered Harmful", 2070 February 2013, . 2074 [Gettys11] 2075 Gettys, J., "Bufferbloat: Dark Buffers in the Internet", 2076 March 2011, 2077 . 2079 Appendix A. Acknowledgments 2081 The authors would like to thank Aamer Akhter, Mikael Abrahamsson, 2082 Mark Andrews, Dmitry Anipko, Ran Atkinson, Fred Baker, Ray Bellis, 2083 Teco Boot, John Brzozowski, Cameron Byrne, Brian Carpenter, Stuart 2084 Cheshire, Julius Chroboczek, Lorenzo Colitti, Robert Cragie, Elwyn 2085 Davies, Ralph Droms, Lars Eggert, Jim Gettys, Olafur Gudmundsson, 2086 Wassim Haddad, Joel M. Halpern, David Harrington, Lee Howard, Ray 2087 Hunter, Joel Jaeggli, Heather Kirksey, Ted Lemon, Acee Lindem, Kerry 2088 Lynn, Daniel Migault, Erik Nordmark, Michael Richardson, Mattia 2089 Rossi, Barbara Stark, Markus Stenberg, Sander Steffann, Don Sturek, 2090 Andrew Sullivan, Dave Taht, Dave Thaler, Michael Thomas, Mark 2091 Townsley, JP Vasseur, Curtis Villamizar, Dan Wing, Russ White, and 2092 James Woodyatt for their comments and contributions within homenet WG 2093 meetings and on the WG mailing list. An acknowledgement generally 2094 means that person's text made it in to the document, or was helpful 2095 in clarifying or reinforcing an aspect of the document. It does not 2096 imply that each controbutor agrees with every point in the document. 2098 Appendix B. Changes 2100 This section will be removed in the final version of the text. 2102 B.1. Version 12 2104 Changes made include: 2106 o Fixed minor typo nits introduced in -11. 2108 o Elwyn Davies' gen-art review comments addressed. 2110 o Some further IESG DISCUSS comments addressed. 2112 B.2. Version 11 (after IESG review) 2114 Changes made include: 2116 o Jouni Korhonen's OPSDIR review comments addressed. 2118 o Elwyn Davies' gen-art review comments addressed. 2120 o Considered secdir review by Samiel Weiler; many points addressed. 2122 o Considered APPSDIR review. 2124 o Addressed a large number of IESG comments and discusses. 2126 B.3. Version 10 (after AD review) 2128 Changes made include: 2130 o Minor changes/clarifications resulting from AD review 2132 B.4. Version 09 (after WGLC) 2134 Changes made include: 2136 o Added note about multicast into or out of site 2138 o Removed further personal draft references, replaced with covering 2139 text 2141 o Routing functionality text updated to avoid ambiguity 2143 o Added note that devices away from homenet may tunnel home (via 2144 VPN) 2146 o Added note that homenets more exposed to provider renumbering than 2147 with IPv4 and NAT 2149 o Added note about devices that may be ULA-only until configured to 2150 be globally addressable 2152 o Removed paragraph about broken CERs that do not work with prefixes 2153 other than /64 2155 o Noted no recommendation on methods to convey prefix information is 2156 made in this text 2158 o Stated that this text does not recommend how to form largest 2159 possible subnets 2161 o Added text about homenet evolution and handling disparate media 2162 types 2164 o Rephrased NAT/firewall text on marginal effectiveness 2166 o Emphasised that multihoming may be to any number of ISPs 2168 B.5. Version 08 2170 Changes made include: 2172 o Various clarifications made in response to list comments 2174 o Added note on ULAs with IPv4, where no GUAs in use 2176 o Added note on naming and internationalisation (IDNA) 2178 o Added note on trust relationships when adding devices 2180 o Added note for MPTCP 2182 o Added various naming and SD notes 2184 o Added various notes on delegated ISP prefixes 2186 B.6. Version 07 2188 Changes made include: 2190 o Removed reference to NPTv6 in section 3.2.4. Instead now say it 2191 has an architectural cost to use in the earlier section, and thus 2192 it is not recommended for use in the homenet architecture. 2194 o Removed 'proxy or extend?' section. Included shorter text in main 2195 body, without mandating either approach for service discovery. 2197 o Made it clearer that ULAs are expected to be used alongside 2198 globals. 2200 o Removed reference to 'advanced security' as described in 2201 draft-vyncke-advanced-ipv6-security. 2203 o Balanced the text between ULQDN and ALQDN. 2205 o Clarify text does not assume default deny or allow on CER, but 2206 that either mode may be enabled. 2208 o Removed ULA-C reference for 'simple' addresses. Instead only 2209 suggested service discovery to find such devices. 2211 o Reiterated that single/multiple CER models to be supported for 2212 multihoming. 2214 o Reordered section 3.3 to improve flow. 2216 o Added recommendation that homenet is not allocated less than /60, 2217 and a /56 is preferable. 2219 o Tidied up first few intro sections. 2221 o Other minor edits from list feedback. 2223 B.7. Version 06 2225 Changes made include: 2227 o Stated that unmanaged goal is 'as far as possible'. 2229 o Added note about multiple /48 ULAs potentially being in use. 2231 o Minor edits from list feedback. 2233 B.8. Version 05 2235 Changes made include: 2237 o Some significant changes to naming and SD section. 2239 o Removed some expired drafts. 2241 o Added notes about issues caused by ISP only delegating a /64. 2243 o Recommended against using prefixes longer than /64. 2245 o Suggested CER asks for /48 by DHCP PD, even if it only receives 2246 less. 2248 o Added note about DS-Lite but emphasised transition is out of 2249 scope. 2251 o Added text about multicast routing. 2253 B.9. Version 04 2255 Changes made include: 2257 o Moved border section from IPv6 differences to principles section. 2259 o Restructured principles into areas. 2261 o Added summary of naming and service discovery discussion from WG 2262 list. 2264 B.10. Version 03 2266 Changes made include: 2268 o Various improvements to the readability. 2270 o Removed bullet lists of requirements, as requested by chair. 2272 o Noted 6204bis has replaced advanced-cpe draft. 2274 o Clarified the topology examples are just that. 2276 o Emphasised we are not targetting walled gardens, but they should 2277 not be precluded. 2279 o Also changed text about requiring support for walled gardens. 2281 o Noted that avoiding falling foul of ingress filtering when 2282 multihomed is desirable. 2284 o Improved text about realms, detecting borders and policies at 2285 borders. 2287 o Stated this text makes no recommendation about default security 2288 model. 2290 o Added some text about failure modes for users plugging things 2291 arbitrarily. 2293 o Expanded naming and service discovery text. 2295 o Added more text about ULAs. 2297 o Removed reference to version 1 on chair feedback. 2299 o Stated that NPTv6 adds architectural cost but is not a homenet 2300 matter if deployed at the CER. This text only considers the 2301 internal homenet. 2303 o Noted multihoming is supported. 2305 o Noted routers may not by separate devices, they may be embedded in 2306 devices. 2308 o Clarified simple and advanced security some more, and RFC 4864 and 2309 6092. 2311 o Stated that there should be just one secret key, if any are used 2312 at all. 2314 o For multihoming, support multiple CERs but note that routing to 2315 the correct CER to avoid ISP filtering may not be optimal within 2316 the homenet. 2318 o Added some ISPs renumber due to privacy laws. 2320 o Removed extra repeated references to Simple Security. 2322 o Removed some solution creep on RIOs/RAs. 2324 o Load-balancing scenario added as to be supported. 2326 B.11. Version 02 2328 Changes made include: 2330 o Made the IPv6 implications section briefer. 2332 o Changed Network Models section to describe properties of the 2333 homenet with illustrative examples, rather than implying the 2334 number of models was fixed to the six shown in 01. 2336 o Text to state multihoming support focused on single CER model. 2337 Multiple CER support is desirable, but not required. 2339 o Stated that NPTv6 not supported. 2341 o Added considerations section for operations and management. 2343 o Added bullet point principles/requirements to Section 3.4. 2345 o Changed IPv6 solutions must not adversely affect IPv4 to should 2346 not. 2348 o End-to-end section expanded to talk about "Simple Security" and 2349 borders. 2351 o Extended text on naming and service discovery. 2353 o Added reference to RFC 2775, RFC 6177. 2355 o Added reference to the new xmDNS draft. 2357 o Added naming/SD requirements from Ralph Droms. 2359 Authors' Addresses 2361 Tim Chown (editor) 2362 University of Southampton 2363 Highfield 2364 Southampton, Hampshire SO17 1BJ 2365 United Kingdom 2367 Email: tjc@ecs.soton.ac.uk 2369 Jari Arkko 2370 Ericsson 2371 Jorvas 02420 2372 Finland 2374 Email: jari.arkko@piuha.net 2376 Anders Brandt 2377 Sigma Designs 2378 Emdrupvej 26A, 1 2379 Copenhagen DK-2100 2380 Denmark 2382 Email: abr@sdesigns.dk 2383 Ole Troan 2384 Cisco Systems, Inc. 2385 Drammensveien 145A 2386 Oslo N-0212 2387 Norway 2389 Email: ot@cisco.com 2391 Jason Weil 2392 Time Warner Cable 2393 13820 Sunrise Valley Drive 2394 Herndon, VA 20171 2395 USA 2397 Email: jason.weil@twcable.com