idnits 2.17.1 draft-ietf-homenet-arch-16.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 (June 10, 2014) is 3606 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) Summary: 2 errors (**), 0 flaws (~~), 1 warning (==), 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: December 12, 2014 Ericsson 6 A. Brandt 7 Sigma Designs 8 O. Troan 9 Cisco Systems, Inc. 10 J. Weil 11 Time Warner Cable 12 June 10, 2014 14 IPv6 Home Networking Architecture Principles 15 draft-ietf-homenet-arch-16 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 December 12, 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 3.2.5. Mobility support . . . . . . . . . . . . . . . . . . . 20 86 3.3. A Self-Organising Network . . . . . . . . . . . . . . . . 20 87 3.3.1. Differentiating neighbouring homenets . . . . . . . . 21 88 3.3.2. Largest practical subnets . . . . . . . . . . . . . . 21 89 3.3.3. Handling varying link technologies . . . . . . . . . . 22 90 3.3.4. Homenet realms and borders . . . . . . . . . . . . . . 22 91 3.3.5. Configuration information from the ISP . . . . . . . . 23 92 3.4. Homenet Addressing . . . . . . . . . . . . . . . . . . . . 23 93 3.4.1. Use of ISP-delegated IPv6 prefixes . . . . . . . . . . 24 94 3.4.2. Stable internal IP addresses . . . . . . . . . . . . . 26 95 3.4.3. Internal prefix delegation . . . . . . . . . . . . . . 27 96 3.4.4. Coordination of configuration information . . . . . . 28 97 3.4.5. Privacy . . . . . . . . . . . . . . . . . . . . . . . 28 98 3.5. Routing functionality . . . . . . . . . . . . . . . . . . 28 99 3.5.1. Multicast support . . . . . . . . . . . . . . . . . . 30 100 3.6. Security . . . . . . . . . . . . . . . . . . . . . . . . . 31 101 3.6.1. Addressability vs reachability . . . . . . . . . . . . 31 102 3.6.2. Filtering at borders . . . . . . . . . . . . . . . . . 32 103 3.6.3. Partial Effectiveness of NAT and Firewalls . . . . . . 32 104 3.6.4. Exfiltration concerns . . . . . . . . . . . . . . . . 33 105 3.6.5. Device capabilities . . . . . . . . . . . . . . . . . 33 106 3.6.6. ULAs as a hint of connection origin . . . . . . . . . 33 107 3.7. Naming and Service Discovery . . . . . . . . . . . . . . . 34 108 3.7.1. Discovering services . . . . . . . . . . . . . . . . . 34 109 3.7.2. Assigning names to devices . . . . . . . . . . . . . . 35 110 3.7.3. The homenet name service . . . . . . . . . . . . . . . 35 111 3.7.4. Name spaces . . . . . . . . . . . . . . . . . . . . . 36 112 3.7.5. Independent operation . . . . . . . . . . . . . . . . 39 113 3.7.6. Considerations for LLNs . . . . . . . . . . . . . . . 39 114 3.7.7. DNS resolver discovery . . . . . . . . . . . . . . . . 39 115 3.7.8. Devices roaming to/from the homenet . . . . . . . . . 40 116 3.8. Other Considerations . . . . . . . . . . . . . . . . . . . 40 117 3.8.1. Quality of Service . . . . . . . . . . . . . . . . . . 40 118 3.8.2. Operations and Management . . . . . . . . . . . . . . 40 119 3.9. Implementing the Architecture on IPv6 . . . . . . . . . . 42 120 4. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . 42 121 5. Security Considerations . . . . . . . . . . . . . . . . . . . 42 122 6. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 43 123 7. References . . . . . . . . . . . . . . . . . . . . . . . . . . 43 124 7.1. Normative References . . . . . . . . . . . . . . . . . . . 43 125 7.2. Informative References . . . . . . . . . . . . . . . . . . 43 126 Appendix A. Acknowledgments . . . . . . . . . . . . . . . . . . . 46 127 Appendix B. Changes . . . . . . . . . . . . . . . . . . . . . . . 46 128 B.1. Version 14 . . . . . . . . . . . . . . . . . . . . . . . . 46 129 B.2. Version 13 . . . . . . . . . . . . . . . . . . . . . . . . 47 130 B.3. Version 12 . . . . . . . . . . . . . . . . . . . . . . . . 47 131 B.4. Version 11 (after IESG review) . . . . . . . . . . . . . . 47 132 B.5. Version 10 (after AD review) . . . . . . . . . . . . . . . 47 133 B.6. Version 09 (after WGLC) . . . . . . . . . . . . . . . . . 47 134 B.7. Version 08 . . . . . . . . . . . . . . . . . . . . . . . . 48 135 B.8. Version 07 . . . . . . . . . . . . . . . . . . . . . . . . 48 136 B.9. Version 06 . . . . . . . . . . . . . . . . . . . . . . . . 49 137 B.10. Version 05 . . . . . . . . . . . . . . . . . . . . . . . . 49 138 B.11. Version 04 . . . . . . . . . . . . . . . . . . . . . . . . 50 139 B.12. Version 03 . . . . . . . . . . . . . . . . . . . . . . . . 50 140 B.13. Version 02 . . . . . . . . . . . . . . . . . . . . . . . . 51 141 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . . 52 143 1. Introduction 145 This document focuses on evolving networking technology within 146 residential home networks with increasing numbers of devices and a 147 trend towards increased internal routing, and the associated 148 challenges with their deployment and operation. There is a growing 149 trend in home networking for the proliferation of networking 150 technology through an increasingly broad range of devices and media. 151 This evolution in scale and diversity sets requirements on IETF 152 protocols. Some of these requirements relate to the introduction of 153 IPv6, others to the introduction of specialised networks for home 154 automation and sensors. 156 While at the time of writing some complex home network topologies 157 exist, most are relatively simple single subnet networks, and 158 ostensibly operate using just IPv4. While there may be IPv6 traffic 159 within the network, e.g., for service discovery, the homenet is 160 provisioned by the ISP as an IPv4 network. Such networks also 161 typically employ solutions that should be avoided, such as private 162 [RFC1918] addressing with (cascaded) network address translation 163 (NAT) [RFC3022], or they may require expert assistance to set up. 165 In contrast, emerging IPv6-capable home networks are very likely to 166 have multiple internal subnets, e.g., to facilitate private and guest 167 networks, heterogeneous link layers, and smart grid components, and 168 have enough address space available to allow every device to have a 169 globally unique address. This implies that internal routing 170 functionality is required, and that the homenet's ISP both provides a 171 large enough prefix to allocate a prefix to each subnet, and that a 172 method is supported for such prefixes to be delegated efficiently to 173 those subnets. 175 It is not practical to expect home users to configure their networks. 176 Thus the assumption of this document is that the homenet is as far as 177 possible self-organising and self-configuring, i.e., it should 178 function without pro-active management by the residential user. 180 The architectural constructs in this document are focused on the 181 problems to be solved when introducing IPv6, with an eye towards a 182 better result than what we have today with IPv4, as well as aiming at 183 a more consistent solution that addresses as many of the identified 184 requirements as possible. The document aims to provide the basis and 185 guiding principles for how standard IPv6 mechanisms and addressing 186 [RFC2460] [RFC4291] can be employed in home networking, while 187 coexisting with existing IPv4 mechanisms. In emerging dual-stack 188 home networks it is vital that introducing IPv6 does not adversely 189 affect IPv4 operation. We assume that the IPv4 network architecture 190 in home networks is what it is, and can not be modified by new 191 recommendations. This document does not discuss how IPv4 home 192 networks provision or deliver support for multiple subnets. It 193 should not be assumed that any future new functionality created with 194 IPv6 in mind will be backward-compatible to include IPv4 support. 195 Further, future deployments, or specific subnets within an otherwise 196 dual-stack home network, may be IPv6-only, in which case 197 considerations for IPv4 impact would not apply. 199 This document proposes a baseline homenet architecture, using 200 protocols and implementations that are as far as possible proven and 201 robust. The scope of the document is primarily the network layer 202 technologies that provide the basic functionality to enable 203 addressing, connectivity, routing, naming and service discovery. 204 While it may, for example, state that homenet components must be 205 simple to deploy and use, it does not discuss specific user 206 interfaces, nor does it discuss specific physical, wireless or data- 207 link layer considerations. Likewise, we also do not specify the 208 whole design of a homenet router from top to bottom, rather we focus 209 on the Layer 3 aspects. This means that Layer 2 is largely out of 210 scope, we're assuming a data link layer that supports IPv6 is 211 present, and that we react accordingly. Any IPv6-over-Foo 212 definitions occur elsewhere. 214 [RFC6204] defines basic requirements for customer edge routers 215 (CERs). This document has recently been updated with the definition 216 of requirements for specific transition tools on the CER in 217 [RFC7084], specifically DS-Lite [RFC6333] and 6rd [RFC5969]. Such 218 detailed specification of CER devices is considered out of scope of 219 this architecture document, and we assume that any required update of 220 the CER device specification as a result of adopting this 221 architecture will be handled as separate and specific updates to 222 these existing documents. Further, the scope of this text is the 223 internal homenet, and thus specific features on the WAN side of the 224 CER are out of scope for this text. 226 1.1. Terminology and Abbreviations 228 In this section we define terminology and abbreviations used 229 throughout the text. 231 o Border: a point, typically resident on a router, between two 232 networks, e.g., between the main internal homenet and a guest 233 network. This defines point(s) at which filtering and forwarding 234 policies for different types of traffic may be applied. 236 o CER: Customer Edge Router: A border router intended for use in a 237 homenet, which connects the homenet to a service provider network. 239 o FQDN: Fully Qualified Domain Name. A globally unique name. 241 o Guest network: A part of the home network intended for use by 242 visitors or guests to the home(net). Devices on the guest network 243 may typically not see or be able to use all services in the 244 home(net). 246 o Homenet: A home network, comprising host and router equipment, 247 with one or more CERs providing connectivity to service provider 248 network(s). 250 o Internet Service Provider (ISP): an entity that provides access to 251 the Internet. In this document, a service provider specifically 252 offers Internet access using IPv6, and may also offer IPv4 253 Internet access. The service provider can provide such access 254 over a variety of different transport methods such as DSL, cable, 255 wireless, and others. 257 o LLN: Low-power and lossy network. 259 o LQDN: Locally Qualified Domain Name. A name local to the homenet. 261 o NAT: Network Address Translation. Typically referring to IPv4 262 Network Address and Port Translation (NAPT) [RFC3022]. 264 o NPTv6: Network Prefix Translation for IPv6 [RFC6296]. 266 o PCP: Port Control Protocol [RFC6887]. 268 o Realm: a network delimited by a defined border. A guest network 269 within a homenet may form one realm. 271 o 'Simple Security'. Defined in [RFC4864] and expanded further in 272 [RFC6092]; describes recommended perimeter security capabilities 273 for IPv6 networks. 275 o ULA: IPv6 Unique Local Address [RFC4193]. 277 o VM: Virtual machine. 279 2. Effects of IPv6 on Home Networking 281 While IPv6 resembles IPv4 in many ways, there are some notable 282 differences in the way it may typically be deployed. It changes 283 address allocation principles, making multi-addressing the norm, and, 284 through the vastly increased address space, allows globally unique IP 285 addresses to be used for all devices in a home network. This section 286 presents an overview of some of the key implications of the 287 introduction of IPv6 for home networking, that are simultaneously 288 both promising and problematic. 290 2.1. Multiple subnets and routers 292 While simple layer 3 topologies involving as few subnets as possible 293 are preferred in home networks, the incorporation of dedicated 294 (routed) subnets remains necessary for a variety of reasons. For 295 instance, an increasingly common feature in modern home routers is 296 the ability to support both guest and private network subnets. 297 Likewise, there may be a need to separate home automation or 298 corporate extension LANs (whereby a home worker can have their 299 corporate network extended into the home using a virtual private 300 network, commonly presented as one port on an Ethernet device) from 301 the main Internet access network, or different subnets may in general 302 be associated with parts of the homenet that have different routing 303 and security policies. Further, link layer networking technology is 304 poised to become more heterogeneous, as networks begin to employ both 305 traditional Ethernet technology and link layers designed for low- 306 power and lossy networks (LLNs), such as those used for certain types 307 of sensor devices. Constraining the flow of certain traffic from 308 Ethernet links to much lower capacity links thus becomes an important 309 topic. 311 The introduction of IPv6 for home networking makes it possible for 312 every home network to be delegated enough address space from its ISP 313 to provision globally unique prefixes for each such subnet in the 314 home. While the number of addresses in a standard /64 IPv6 prefix is 315 practically unlimited, the number of prefixes available for 316 assignment to the home network is not. As a result the growth 317 inhibitor for the home network shifts from the number of addresses to 318 the number of prefixes offered by the provider; this topic is 319 discussed in [RFC6177] (BCP 157), which recommends that "end sites 320 always be able to obtain a reasonable amount of address space for 321 their actual and planned usage". 323 The addition of routing between subnets raises a number of issues. 324 One is a method by which prefixes can be efficiently allocated to 325 each subnet, without user intervention. Another is the issue of how 326 to extend mechanisms such as zero configuration service discovery 327 which currently only operate within a single subnet using link-local 328 traffic. In a typical IPv4 home network, there is only one subnet, 329 so such mechanisms would normally operate as expected. For multi- 330 subnet IPv6 home networks there are two broad choices to enable such 331 protocols to work across the scope of the entire homenet; extend 332 existing protocols to work across that scope, or introduce proxies 333 for existing link layer protocols. This topic is discussed in 334 Section 3.7. 336 2.2. Global addressability and elimination of NAT 338 The possibility for direct end-to-end communication on the Internet 339 to be restored by the introduction of IPv6 is on the one hand an 340 incredible opportunity for innovation and simpler network operation, 341 but on the other hand it is also a concern as it potentially exposes 342 nodes in the internal networks to receipt of unwanted and possibly 343 malicious traffic from the Internet. 345 With devices and applications able to talk directly to each other 346 when they have globally unique addresses, there may be an expectation 347 of improved host security to compensate for this. It should be noted 348 that many devices may (for example) ship with default settings that 349 make them readily vulnerable to compromise by external attackers if 350 globally accessible, or may simply not have robustness designed-in 351 because it was either assumed such devices would only be used on 352 private networks or the device itself doesn't have the computing 353 power to apply the necessary security methods. In addition, the 354 upgrade cycle for devices (or their firmware) may be slow, and/or 355 lack auto-update mechanisms. 357 It is thus important to distinguish between addressability and 358 reachability. While IPv6 offers global addressability through use of 359 globally unique addresses in the home, whether devices are globally 360 reachable or not would depend on any firewall or filtering 361 configuration, and not, as is commonly the case with IPv4, the 362 presence or use of NAT. In this respect, IPv6 networks may or may 363 not have filters applied at their borders to control such traffic, 364 i.e., at the homenet CER. [RFC4864] and [RFC6092] discuss such 365 filtering, and the merits of 'default allow' against 'default deny' 366 policies for external traffic initiated into a homenet. This topic 367 is discussed further in Section 3.6.1. 369 2.3. Multi-Addressing of devices 371 In an IPv6 network, devices will often acquire multiple addresses, 372 typically at least a link-local address and one or more globally 373 unique addresses. Where a homenet is multihomed, a device would 374 typically receive a globally unique address (GUA) from within the 375 delegated prefix from each upstream ISP. Devices may also have an 376 IPv4 address if the network is dual-stack, an IPv6 Unique Local 377 Address (ULA) [RFC4193] (see below), and one or more IPv6 Privacy 378 Addresses [RFC4941]. 380 It should thus be considered the norm for devices on IPv6 home 381 networks to be multi-addressed, and to need to make appropriate 382 address selection decisions for the candidate source and destination 383 address pairs for any given connection. In multihoming scenarios 384 nodes will be configured with one address from each upstream ISP 385 prefix. In such cases the presence of upstream BCP 38 [RFC2827] 386 ingress filtering requires such multi-addressed nodes to select the 387 correct source address to be used for the corresponding uplink. 388 Default Address Selection for IPv6 [RFC6724] provides a solution for 389 this, but a challenge here is that the node may not have the 390 information it needs to make that decision based on addresses alone. 391 We discuss this challenge in Section 3.2.4. 393 2.4. Unique Local Addresses (ULAs) 395 [RFC4193] defines Unique Local Addresses (ULAs) for IPv6 that may be 396 used to address devices within the scope of a single site. Support 397 for ULAs for IPv6 CERs is described in [RFC6204]. A home network 398 running IPv6 should deploy ULAs alongside its globally unique 399 prefix(es) to allow stable communication between devices (on 400 different subnets) within the homenet where that externally allocated 401 globally unique prefix may change over time, e.g., due to renumbering 402 within the subscriber's ISP, or where external connectivity may be 403 temporarily unavailable. A homenet using provider-assigned global 404 addresses is exposed to its ISP renumbering the network to a much 405 larger degree than before whereas, for IPv4, NAT isolated the user 406 against ISP renumbering to some extent. 408 While setting up a network there may be a period where it has no 409 external connectivity, in which case ULAs would be required for 410 inter-subnet communication. In the case where home automation 411 networks are being set up in a new home/deployment (as early as 412 during construction of the home), such networks will likely need to 413 use their own /48 ULA prefix. Depending upon circumstances beyond 414 the control of the owner of the homenet, it may be impossible to 415 renumber the ULA used by the home automation network so routing 416 between ULA /48s may be required. Also, some devices, particularly 417 constrained devices, may have only a ULA (in addition to a link- 418 local), while others may have both a GUA and a ULA. 420 Note that unlike private IPv4 RFC 1918 space, the use of ULAs does 421 not imply use of an IPv6 equivalent of a traditional IPv4 NAT 422 [RFC3022], or of NPTv6 prefix-based NAT [RFC6296]. When an IPv6 node 423 in a homenet has both a ULA and a globally unique IPv6 address, it 424 should only use its ULA address internally, and use its additional 425 globally unique IPv6 address as a source address for external 426 communications. This should be the natural behaviour given support 427 for Default Address Selection for IPv6 [RFC6724]. By using such 428 globally unique addresses between hosts and devices in remote 429 networks, the architectural cost and complexity, particularly to 430 applications, of NAT or NPTv6 translation is avoided. As such, 431 neither IPv6 NAT or NPTv6 is recommended for use in the homenet 432 architecture. Further, the homenet border router(s) should filter 433 packets with ULA source/destination addresses as discussed in 434 Section 3.4.2. 436 Devices in a homenet may be given only a ULA as a means to restrict 437 reachability from outside the homenet. ULAs can be used by default 438 for devices that, without additional configuration (e.g., via a web 439 interface), would only offer services to the internal network. For 440 example, a printer might only accept incoming connections on a ULA 441 until configured to be globally reachable, at which point it acquires 442 a global IPv6 address and may be advertised via a global name space. 444 Where both a ULA and a global prefix are in use, the ULA source 445 address is used to communicate with ULA destination addresses when 446 appropriate, i.e., when the ULA source and destination lie within the 447 /48 ULA prefix(es) known to be used within the same homenet. In 448 cases where multiple /48 ULA prefixes are in use within a single 449 homenet (perhaps because multiple homenet routers each independently 450 auto-generate a /48 ULA prefix and then share prefix/routing 451 information), utilising a ULA source address and a ULA destination 452 address from two disjoint internal ULA prefixes is preferable to 453 using GUAs. 455 While a homenet should operate correctly with two or more /48 ULAs 456 enabled, a mechanism for the creation and use of a single /48 ULA 457 prefix is desirable for addressing consistency and policy 458 enforcement. 460 A counter-argument to using ULAs is that it is undesirable to 461 aggressively deprecate global prefixes for temporary loss of 462 connectivity, so for a host to lose its global address there would 463 have to be a connection breakage longer than the lease period, and 464 even then, deprecating prefixes when there is no connectivity may not 465 be advisable. However, it is assumed in this architecture that 466 homenets should support and use ULAs. 468 2.5. Avoiding manual configuration of IP addresses 470 Some IPv4 home networking devices expose IPv4 addresses to users, 471 e.g., the IPv4 address of a home IPv4 CER that may be configured via 472 a web interface. In potentially complex future IPv6 homenets, users 473 should not be expected to enter IPv6 literal addresses in devices or 474 applications, given their much greater length and the apparent 475 randomness of such addresses to a typical home user. Thus, even for 476 the simplest of functions, simple naming and the associated (minimal, 477 and ideally zero configuration) discovery of services is imperative 478 for the easy deployment and use of homenet devices and applications. 480 2.6. IPv6-only operation 482 It is likely that IPv6-only networking will be deployed first in new 483 home network deployments, often referred to as 'greenfield' 484 scenarios, where there is no existing IPv4 capability, or perhaps as 485 one element of an otherwise dual-stack network. Running IPv6-only 486 adds additional requirements, e.g., for devices to get configuration 487 information via IPv6 transport (not relying on an IPv4 protocol such 488 as IPv4 DHCP), and for devices to be able to initiate communications 489 to external devices that are IPv4-only. 491 Some specific transition technologies which may be deployed by the 492 homenet's ISP are discussed in [RFC7084]. In addition, certain other 493 functions may be desirable on the CER, e.g., to access content in the 494 IPv4 Internet, NAT64 [RFC6144] and DNS64 [RFC6145] may be applicable. 496 The widespread availability of robust solutions to these types of 497 requirements will help accelerate the uptake of IPv6-only homenets. 498 The specifics of these are however beyond the scope of this document, 499 especially those functions that reside on the CER. 501 3. Homenet Architecture Principles 503 The aim of this text is to outline how to construct advanced IPv6- 504 based home networks involving multiple routers and subnets using 505 standard IPv6 addressing and protocols [RFC2460] [RFC4291] as the 506 basis. As described in Section 3.1, solutions should as far as 507 possible re-use existing protocols, and minimise changes to hosts and 508 routers, but some new protocols, or extensions, are likely to be 509 required. In this section, we present the elements of the proposed 510 home networking architecture, with discussion of the associated 511 design principles. 513 In general, home network equipment needs to be able to operate in 514 networks with a range of different properties and topologies, where 515 home users may plug components together in arbitrary ways and expect 516 the resulting network to operate. Significant manual configuration 517 is rarely, if at all, possible, or even desirable given the knowledge 518 level of typical home users. Thus the network should, as far as 519 possible, be self-configuring, though configuration by advanced users 520 should not be precluded. 522 The homenet needs to be able to handle or provision at least 523 o Routing 525 o Prefix configuration for routers 527 o Name resolution 529 o Service discovery 531 o Network security 533 The remainder of this document describes the principles by which the 534 homenet architecture may deliver these properties. 536 3.1. General Principles 538 There is little that the Internet standards community can do about 539 the physical topologies or the need for some networks to be separated 540 at the network layer for policy or link layer compatibility reasons. 541 However, there is a lot of flexibility in using IP addressing and 542 inter-networking mechanisms. This text discusses how such 543 flexibility should be used to provide the best user experience and 544 ensure that the network can evolve with new applications in the 545 future. The principles described in this text should be followed 546 when designing homenet protocol solutions. 548 3.1.1. Reuse existing protocols 550 Existing protocols will be used to meet the requirements of home 551 networks. Where necessary, extensions will be made to those 552 protocols. When no existing protocol is found to be suitable, a new 553 or emerging protocol may be used. Therefore, it is important that no 554 design or architectural decisions are made that would preclude the 555 use of new or emerging protocols. 557 A generally conservative approach, giving weight to running (and 558 available) code, is preferable. Where new protocols are required, 559 evidence of commitment to implementation by appropriate vendors or 560 development communities is highly desirable. Protocols used should 561 be backwardly compatible, and forward compatible where changes are 562 made. 564 3.1.2. Minimise changes to hosts and routers 566 In order to maximise deployability of new homenets, where possible 567 any requirement for changes to hosts and routers should be minimised, 568 though solutions which, for example, incrementally improve capability 569 with host or router changes may be acceptable. There may be cases 570 where changes are unavoidable, e.g., to allow a given homenet routing 571 protocol to be self-configuring, or to support routing based on 572 sources addresses in addition to destination addresses (to improve 573 multihoming support, as discussed in Section 3.2.4). 575 3.2. Homenet Topology 577 This section considers homenet topologies, and the principles that 578 may be applied in designing an architecture to support as wide a 579 range of such topologies as possible. 581 3.2.1. Supporting arbitrary topologies 583 There should ideally be no built-in assumptions about the topology in 584 home networks, as users are capable of connecting their devices in 585 'ingenious' ways. Thus arbitrary topologies and arbitrary routing 586 will need to be supported, or at least the failure mode for when the 587 user makes a mistake should be as robust as possible, e.g., de- 588 activating a certain part of the infrastructure to allow the rest to 589 operate. In such cases, the user should ideally have some useful 590 indication of the failure mode encountered. 592 There should be no topology scenarios which cause loss of 593 connectivity, except when the user creates a physical island within 594 the topology. Some potentially pathological cases that can be 595 created include bridging ports of a router together, however this 596 case can be detected and dealt with by the router. Loops within a 597 routed topology are in a sense good in that they offer redundancy. 598 Topologies that include potential bridging loops can be dangerous but 599 are also detectable when a switch learns the MAC of one of its 600 interfaces on another or runs a spanning tree or link state protocol. 601 It is only topologies with such potential loops using simple 602 repeaters that are truly pathological. 604 The topology of the homenet may change over time, due to the addition 605 or removal of equipment, but also due to temporary failures or 606 connectivity problems. In some cases this may lead to, for example, 607 a multihomed homenet being split into two isolated homenets, or, 608 after such a fault is remedied, two isolated parts reconfiguring back 609 to a single network. 611 3.2.2. Network topology models 613 As hinted above, while the architecture may focus on likely common 614 topologies, it should not preclude any arbitrary topology from being 615 constructed. 617 Most IPv4 home network models at the time of writing tend to be 618 relatively simple, typically a single NAT router to the ISP and a 619 single internal subnet but, as discussed earlier, evolution in 620 network architectures is driving more complex topologies, such as the 621 separation of guest and private networks. There may also be some 622 cascaded IPv4 NAT scenarios, which we mention in the next section. 623 For IPv6 homenets, the Network Architectures described in [RFC6204] 624 and its successor [RFC7084] should, as a minimum, be supported. 626 There are a number of properties or attributes of a home network that 627 we can use to describe its topology and operation. The following 628 properties apply to any IPv6 home network: 630 o Presence of internal routers. The homenet may have one or more 631 internal routers, or may only provide subnetting from interfaces 632 on the CER. 634 o Presence of isolated internal subnets. There may be isolated 635 internal subnets, with no direct connectivity between them within 636 the homenet (with each having its own external connectivity). 637 Isolation may be physical, or implemented via IEEE 802.1q VLANs. 638 The latter is however not something a typical user would be 639 expected to configure. 641 o Demarcation of the CER. The CER(s) may or may not be managed by 642 the ISP. If the demarcation point is such that the customer can 643 provide or manage the CER, its configuration must be simple. Both 644 models must be supported. 646 Various forms of multihoming are likely to become more prevalent with 647 IPv6 home networks, where the homenet may have two or more external 648 ISP connections, as discussed further below. Thus the following 649 properties should also be considered for such networks: 651 o Number of upstream providers. The majority of home networks today 652 consist of a single upstream ISP, but it may become more common in 653 the future for there to be multiple ISPs, whether for resilience 654 or provision of additional services. Each would offer its own 655 prefix. Some may or may not provide a default route to the public 656 Internet. 658 o Number of CERs. The homenet may have a single CER, which might be 659 used for one or more providers, or multiple CERs. The presence of 660 multiple CERs adds additional complexity for multihoming 661 scenarios, and protocols like PCP that may need to manage 662 connection-oriented state mappings on the same CER as used for 663 subsequent traffic flows. 665 In the following sections we give some examples of the types of 666 homenet topologies we may see in the future. This is not intended to 667 be an exhaustive or complete list, rather an indicative one to 668 facilitate the discussion in this text. 670 3.2.2.1. A: Single ISP, Single CER, Internal routers 672 Figure 1 shows a home network with multiple local area networks. 673 These may be needed for reasons relating to different link layer 674 technologies in use or for policy reasons, e.g., classic Ethernet in 675 one subnet and a LLN link layer technology in another. In this 676 example there is no single router that a priori understands the 677 entire topology. The topology itself may also be complex, and it may 678 not be possible to assume a pure tree form, for instance (because 679 home users may plug routers together to form arbitrary topologies 680 including those with potential loops in them). 682 +-------+-------+ \ 683 | Service | \ 684 | Provider | | Service 685 | Router | | Provider 686 +-------+-------+ | network 687 | / 688 | Customer / 689 | Internet connection 690 | 691 +------+--------+ \ 692 | IPv6 | \ 693 | Customer Edge | \ 694 | Router | | 695 +----+-+---+----+ | 696 Network A | | | Network B(E) | 697 ----+-------------+----+ | +---+-------------+------+ | 698 | | | | | | | 699 +----+-----+ +-----+----+ | +----+-----+ +-----+----+ | | 700 |IPv6 Host | |IPv6 Host | | | IPv6 Host| |IPv6 Host | | | 701 | H1 | | H2 | | | H3 | | H4 | | | 702 +----------+ +----------+ | +----------+ +----------+ | | 703 | | | | | 704 Link F | ---+------+------+-----+ | 705 | | Network E(B) | 706 +------+--------+ | | End-User 707 | IPv6 | | | networks 708 | Interior +------+ | 709 | Router | | 710 +---+-------+-+-+ | 711 Network C | | Network D | 712 ----+-------------+---+ +---+-------------+--- | 713 | | | | | 714 +----+-----+ +-----+----+ +----+-----+ +-----+----+ | 715 |IPv6 Host | |IPv6 Host | | IPv6 Host| |IPv6 Host | | 716 | H5 | | H6 | | H7 | | H8 | / 717 +----------+ +----------+ +----------+ +----------+ / 719 Figure 1 721 In this diagram there is one CER. It has a single uplink interface. 722 It has three additional interfaces connected to Network A, Link F, 723 and Network B. IPv6 Internal Router (IR) has four interfaces 724 connected to Link F, Network C, Network D and Network E. Network B 725 and Network E have been bridged, likely inadvertently. This could be 726 as a result of connecting a wire between a switch for Network B and a 727 switch for Network E. 729 Any of logical Networks A through F might be wired or wireless. 731 Where multiple hosts are shown, this might be through one or more 732 physical ports on the CER or IPv6 (IR), wireless networks, or through 733 one or more layer-2 only Ethernet switches. 735 3.2.2.2. B: Two ISPs, Two CERs, Shared subnet 737 +-------+-------+ +-------+-------+ \ 738 | Service | | Service | \ 739 | Provider A | | Provider B | | Service 740 | Router | | Router | | Provider 741 +------+--------+ +-------+-------+ | network 742 | | / 743 | Customer | / 744 | Internet connections | / 745 | | 746 +------+--------+ +-------+-------+ \ 747 | IPv6 | | IPv6 | \ 748 | Customer Edge | | Customer Edge | \ 749 | Router 1 | | Router 2 | / 750 +------+--------+ +-------+-------+ / 751 | | / 752 | | | End-User 753 ---+---------+---+---------------+--+----------+--- | network(s) 754 | | | | \ 755 +----+-----+ +-----+----+ +----+-----+ +-----+----+ \ 756 |IPv6 Host | |IPv6 Host | | IPv6 Host| |IPv6 Host | / 757 | H1 | | H2 | | H3 | | H4 | / 758 +----------+ +----------+ +----------+ +----------+ 760 Figure 2 762 Figure 2 illustrates a multihomed homenet model, where the customer 763 has connectivity via CER1 to ISP A and via CER2 to ISP B. This 764 example shows one shared subnet where IPv6 nodes would potentially be 765 multihomed and receive multiple IPv6 global prefixes, one per ISP. 766 This model may also be combined with that shown in Figure 1 to create 767 a more complex scenario with multiple internal routers. Or the above 768 shared subnet may be split in two, such that each CER serves a 769 separate isolated subnet, which is a scenario seen with some IPv4 770 networks today. 772 3.2.2.3. C: Two ISPs, One CER, Shared subnet 774 +-------+-------+ +-------+-------+ \ 775 | Service | | Service | \ 776 | Provider A | | Provider B | | Service 777 | Router | | Router | | Provider 778 +-------+-------+ +-------+-------+ | network 779 | | / 780 | Customer | / 781 | Internet | / 782 | connections | 783 +---------+---------+ \ 784 | IPv6 | \ 785 | Customer Edge | \ 786 | Router | / 787 +---------+---------+ / 788 | / 789 | | End-User 790 ---+------------+-------+--------+-------------+--- | network(s) 791 | | | | \ 792 +----+-----+ +----+-----+ +----+-----+ +-----+----+ \ 793 |IPv6 Host | |IPv6 Host | | IPv6 Host| |IPv6 Host | / 794 | H1 | | H2 | | H3 | | H4 | / 795 +----------+ +----------+ +----------+ +----------+ 797 Figure 3 799 Figure 3 illustrates a model where a home network may have multiple 800 connections to multiple providers or multiple logical connections to 801 the same provider, with shared internal subnets. 803 3.2.3. Dual-stack topologies 805 It is expected that most homenet deployments will for the immediate 806 future be dual-stack IPv4/IPv6. In such networks it is important not 807 to introduce new IPv6 capabilities that would cause a failure if used 808 alongside IPv4+NAT, given that such dual-stack homenets will be 809 commonplace for some time. That said, it is desirable that IPv6 810 works better than IPv4 in as many scenarios as possible. Further, 811 the homenet architecture must operate in the absence of IPv4. 813 A general recommendation is to follow the same topology for IPv6 as 814 is used for IPv4, but not to use NAT. Thus there should be routed 815 IPv6 where an IPv4 NAT is used and, where there is no NAT, routing or 816 bridging may be used. Routing may have advantages when compared to 817 bridging together high speed and lower speed shared media, and in 818 addition bridging may not be suitable for some networks, such as ad- 819 hoc mobile networks. 821 In some cases IPv4 home networks may feature cascaded NATs. End 822 users are frequently unaware that they have created such networks as 823 'home routers' and 'home switches' are frequently confused. In 824 addition, there are cases where NAT routers are included within 825 Virtual Machine Hypervisors, or where Internet connection sharing 826 services have been enabled. This document applies equally to such 827 hidden NAT 'routers'. IPv6 routed versions of such cases will be 828 required. We should thus also note that routers in the homenet may 829 not be separate physical devices; they may be embedded within other 830 devices. 832 3.2.4. Multihoming 834 A homenet may be multihomed to multiple providers, as the network 835 models above illustrate. This may either take a form where there are 836 multiple isolated networks within the home or a more integrated 837 network where the connectivity selection needs to be dynamic. 838 Current practice is typically of the former kind, but the latter is 839 expected to become more commonplace. 841 In the general homenet architecture, multihomed hosts should be 842 multi-addressed with a global IPv6 address from the global prefix 843 delegated from each ISP they communicate with or through. When such 844 multi-addressing is in use, hosts need some way to pick source and 845 destination address pairs for connections. A host may choose a 846 source address to use by various methods, most commonly [RFC6724]. 847 Applications may of course do different things, and this should not 848 be precluded. 850 For the single CER Network Model C illustrated above, multihoming may 851 be offered by source-based routing at the CER. With multiple exit 852 routers, as in CER Network Model B, the complexity rises. Given a 853 packet with a source address on the home network, the packet must be 854 routed to the proper egress to avoid BCP 38 ingress filtering if 855 exiting through the wrong ISP. It is highly desirable that the 856 packet is routed in the most efficient manner to the correct exit, 857 though as a minimum requirement the packet should not be dropped. 859 The homenet architecture should support both the above models, i.e., 860 one or more CERs. However, the general multihoming problem is broad, 861 and solutions suggested to date within the IETF have included complex 862 architectures for monitoring connectivity, traffic engineering, 863 identifier-locator separation, connection survivability across 864 multihoming events, and so on. It is thus important that the homenet 865 architecture should as far as possible minimise the complexity of any 866 multihoming support. 868 An example of such a 'simpler' approach has been documented in 869 [RFC7157]. Alternatively a flooding/routing protocol could 870 potentially be used to pass information through the homenet, such 871 that internal routers and ultimately end hosts could learn per-prefix 872 configuration information, allowing better address selection 873 decisions to be made. However, this would imply router and, most 874 likely, host changes. Another avenue is to introduce support 875 throughout the homenet for routing which is based on the source as 876 well as the destination address of each packet. While greatly 877 improving the 'intelligence' of routing decisions within the homenet, 878 such an approach would require relatively significant router changes 879 but avoid host changes. 881 As explained previously, while NPTv6 has been proposed for providing 882 multi-homing support in networks, its use is not recommended in the 883 homenet architecture. 885 It should be noted that some multihoming scenarios may see one 886 upstream being a "walled garden", and thus only appropriate for 887 connectivity to the services of that provider; an example may be a 888 VPN service that only routes back to the enterprise business network 889 of a user in the homenet. As per Section 4.2.1 of [RFC3002] we do 890 not specifically target walled garden multihoming as a goal of this 891 document. 893 The homenet architecture should also not preclude use of host or 894 application-oriented tools, e.g., Shim6 [RFC5533], MPTCP [RFC6824] or 895 Happy Eyeballs [RFC6555]. In general, any incremental improvements 896 obtained by host changes should give benefit for the hosts 897 introducing them, but not be required. 899 3.2.5. Mobility support 901 Devices may be mobile within the homenet. While resident on the same 902 subnet, their address will remain persistent, but should devices move 903 to a different (wireless) subnet, they will acquire a new address in 904 that subnet. It is desirable that the homenet supports internal 905 device mobility. To do so, the homenet may either extend the reach 906 of specific wireless subnets to enable wireless roaming across the 907 home (availability of a specific subnet across the home), or it may 908 support mobility protocols to facilitate such roaming where multiple 909 subnets are used. 911 3.3. A Self-Organising Network 913 The home network infrastructure should be naturally self-organising 914 and self-configuring under different circumstances relating to the 915 connectivity status to the Internet, number of devices, and physical 916 topology. At the same time, it should be possible for advanced users 917 to manually adjust (override) the current configuration. 919 While a goal of the homenet architecture is for the network to be as 920 self-organising as possible, there may be instances where some manual 921 configuration is required, e.g., the entry of a cryptographic key to 922 apply wireless security, or to configure a shared routing secret. 923 The latter may be relevant when considering how to bootstrap a 924 routing configuration. It is highly desirable that the number of 925 such configurations is minimised. 927 3.3.1. Differentiating neighbouring homenets 929 It is important that self-configuration with 'unintended' devices is 930 avoided. There should be a way for a user to administratively assert 931 in a simple way whether or not a device belongs to a given homenet. 932 The goal is to allow the establishment of borders, particularly 933 between two adjacent homenets, and to avoid unauthorised devices from 934 participating in the homenet. Such an authorisation capability may 935 need to operate through multiple hops in the homenet. 937 The homenet should thus support a way for a homenet owner to claim 938 ownership of their devices in a reasonably secure way. This could be 939 achieved by a pairing mechanism, by for example pressing buttons 940 simultaneously on an authenticated and a new homenet device, or by an 941 enrolment process as part of an autonomic networking environment. 943 While there may be scenarios where one homenet may wish to 944 intentionally gain access through another, e.g. to share external 945 connectivity costs, such scenarios are not discussed in this 946 document. 948 3.3.2. Largest practical subnets 950 Today's IPv4 home networks generally have a single subnet, and early 951 dual-stack deployments have a single congruent IPv6 subnet, possibly 952 with some bridging functionality. More recently, some vendors have 953 started to introduce 'home' and 'guest' functions, which in IPv6 954 would be implemented as two subnets. 956 Future home networks are highly likely to have one or more internal 957 routers and thus need multiple subnets, for the reasons described 958 earlier. As part of the self-organisation of the network, the 959 homenet should subdivide itself into the largest practical subnets 960 that can be constructed within the constraints of link layer 961 mechanisms, bridging, physical connectivity, and policy, and where 962 applicable performance or other criteria. In such subdivisions the 963 logical topology may not necessarily match the physical topology. 965 This text does not, however, make recommendations on how such 966 subdivision should occur. It is expected that subsequent documents 967 will address this problem. 969 While it may be desirable to maximise the chance of link-local 970 protocols operating across a homenet by maximising the size of a 971 subnet, multi-subnet home networks are inevitable, so their support 972 must be included. 974 3.3.3. Handling varying link technologies 976 Homenets tend to grow organically over many years, and a homenet will 977 typically be built over link-layer technologies from different 978 generations. Current homenets typically use links ranging from 979 1Mbit/s up to 1Gbit/s, which is a three orders of magnitude 980 throughput discrepancy. We expect this discrepancy to widen further 981 as both high-speed and low-power technologies are deployed. 983 Homenet protocols should be designed to deal well with 984 interconnecting links of very different throughputs. In particular, 985 flows local to a link should not be flooded throughout the homenet, 986 even when sent over multicast, and, whenever possible, the homenet 987 protocols should be able to choose the faster links and avoid the 988 slower ones. 990 Links (particularly wireless links) may also have limited numbers of 991 transmit opportunities (txops), and there is a clear trend driven by 992 both power and downward compatibility constraints toward aggregation 993 of packets into these limited txops while increasing throughput. 994 Transmit opportunities may be a system's scarcest resource and 995 therefore also strongly limit actual throughput available. 997 3.3.4. Homenet realms and borders 999 The homenet will need to be aware of the extent of its own 'site', 1000 which will, for example, define the borders for ULA and site scope 1001 multicast traffic, and may require specific security policies to be 1002 applied. The homenet will have one or more such borders with 1003 external connectivity providers. 1005 A homenet will most likely also have internal borders between 1006 internal realms, e.g., a guest realm or a corporate network extension 1007 realm. It is desirable that appropriate borders can be configured to 1008 determine, for example, the scope of where network prefixes, routing 1009 information, network traffic, service discovery and naming may be 1010 shared. The default mode internally should be to share everything. 1012 It is expected that a realm would span at least an entire subnet, and 1013 thus the borders lie at routers which receive delegated prefixes 1014 within the homenet. It is also desirable, for a richer security 1015 model, that hosts are able to make communication decisions based on 1016 available realm and associated prefix information in the same way 1017 that routers at realm borders can. 1019 A simple homenet model may just consider three types of realm and the 1020 borders between them, namely the internal homenet, the ISP and a 1021 guest network. In this case the borders will include that from the 1022 homenet to the ISP, that from the guest network to the ISP, and that 1023 from the homenet to the guest network. Regardless, it should be 1024 possible for additional types of realms and borders to be defined, 1025 e.g., for some specific LLN-based network, such as Smart Grid, and 1026 for these to be detected automatically, and for an appropriate 1027 default policy to be applied as to what type of traffic/data can flow 1028 across such borders. 1030 It is desirable to classify the external border of the home network 1031 as a unique logical interface separating the home network from 1032 service provider network/s. This border interface may be a single 1033 physical interface to a single service provider, multiple layer 2 1034 sub-interfaces to a single service provider, or multiple connections 1035 to a single or multiple providers. This border makes it possible to 1036 describe edge operations and interface requirements across multiple 1037 functional areas including security, routing, service discovery, and 1038 router discovery. 1040 It should be possible for the homenet user to override any 1041 automatically determined borders and the default policies applied 1042 between them, the exception being that it may not be possible to 1043 override policies defined by the ISP at the external border. 1045 3.3.5. Configuration information from the ISP 1047 In certain cases, it may be useful for the homenet to get certain 1048 configuration information from its ISP. For example, the homenet 1049 DHCP server may request and forward some options that it gets from 1050 its upstream DHCP server, though the specifics of the options may 1051 vary across deployments. There is potential complexity here of 1052 course should the homenet be multihomed. 1054 3.4. Homenet Addressing 1056 The IPv6 addressing scheme used within a homenet must conform to the 1057 IPv6 addressing architecture [RFC4291]. In this section we discuss 1058 how the homenet needs to adapt to the prefixes made available to it 1059 by its upstream ISP, such that internal subnets, hosts and devices 1060 can obtain the and configure the necessary addressing information to 1061 operate. 1063 3.4.1. Use of ISP-delegated IPv6 prefixes 1065 Discussion of IPv6 prefix allocation policies is included in 1066 [RFC6177]. In practice, a homenet may receive an arbitrary length 1067 IPv6 prefix from its provider, e.g., /60, /56 or /48. The offered 1068 prefix may be stable or change from time to time; it is generally 1069 expected that ISPs will offer relatively stable prefixes to their 1070 residential customers. Regardless, the home network needs to be 1071 adaptable as far as possible to ISP prefix allocation policies, and 1072 thus make no assumptions about the stability of the prefix received 1073 from an ISP, or the length of the prefix that may be offered. 1075 However, if, for example, only a /64 is offered by the ISP, the 1076 homenet may be severely constrained or even unable to function. 1077 [RFC6177] (BCP 157) states that "a key principle for address 1078 management is that end sites always be able to obtain a reasonable 1079 amount of address space for their actual and planned usage, and over 1080 time ranges specified in years rather than just months. In practice, 1081 that means at least one /64, and in most cases significantly more. 1082 One particular situation that must be avoided is having an end site 1083 feel compelled to use IPv6-to-IPv6 Network Address Translation or 1084 other burdensome address conservation techniques because it could not 1085 get sufficient address space." This architecture document assumes 1086 that the guidance in the quoted text is being followed by ISPs. 1088 There are many problems that would arise from a homenet not being 1089 offered a sufficient prefix size for its needs. Rather than attempt 1090 to contrive a method for a homenet to operate in a constrained manner 1091 when faced with insufficient prefixes, such as the use of subnet 1092 prefixes longer than /64 (which would break stateless address 1093 autoconfiguration [RFC4862]), use of NPTv6, or falling back to 1094 bridging across potentially very different media, it is recommended 1095 that the receiving router instead enters an error state and issues 1096 appropriate warnings. Some consideration may need to be given to how 1097 such a warning or error state should best be presented to a typical 1098 home user. 1100 Thus a homenet CER should request, for example via DHCP Prefix 1101 Delegation (DHCP PD) [RFC3633], that it would like a /48 prefix from 1102 its ISP, i.e., it asks the ISP for the maximum size prefix it might 1103 expect to be offered, even if in practice it may only be offered a 1104 /56 or /60. For a typical IPv6 homenet, it is not recommended that 1105 an ISP offer less than a /60 prefix, and it is highly preferable that 1106 the ISP offers at least a /56. It is expected that the allocated 1107 prefix to the homenet from any single ISP is a contiguous, aggregated 1108 one. While it may be possible for a homenet CER to issue multiple 1109 prefix requests to attempt to obtain multiple delegations, such 1110 behaviour is out of scope of this document. 1112 The norm for residential customers of large ISPs may be similar to 1113 their single IPv4 address provision; by default it is likely to 1114 remain persistent for some time, but changes in the ISP's own 1115 provisioning systems may lead to the customer's IP (and in the IPv6 1116 case their prefix pool) changing. It is not expected that ISPs will 1117 generally support Provider Independent (PI) addressing for 1118 residential homenets. 1120 When an ISP does need to restructure, and in doing so renumber its 1121 customer homenets, 'flash' renumbering is likely to be imposed. This 1122 implies a need for the homenet to be able to handle a sudden 1123 renumbering event which, unlike the process described in [RFC4192], 1124 would be a 'flag day" event, which means that a graceful renumbering 1125 process moving through a state with two active prefixes in use would 1126 not be possible. While renumbering can be viewed as an extended 1127 version of an initial numbering process, the difference between flash 1128 renumbering and an initial 'cold start' is the need to provide 1129 service continuity. 1131 There may be cases where local law means some ISPs are required to 1132 change IPv6 prefixes (current IPv4 addresses) for privacy reasons for 1133 their customers. In such cases it may be possible to avoid an 1134 instant 'flash' renumbering and plan a non-flag day renumbering as 1135 per RFC 4192. Similarly, if an ISP has a planned renumbering 1136 process, it may be able to adjust lease timers, etc appropriately. 1138 The customer may of course also choose to move to a new ISP, and thus 1139 begin using a new prefix. In such cases the customer should expect a 1140 discontinuity, and not only may the prefix change, but potentially 1141 also the prefix length if the new ISP offers a different default size 1142 prefix. The homenet may also be forced to renumber itself if 1143 significant internal 'replumbing' is undertaken by the user. 1144 Regardless, it's desirable that homenet protocols support rapid 1145 renumbering and that operational processes don't add unnecessary 1146 complexity for the renumbering process. Further, the introduction of 1147 any new homenet protocols should not make any form of renumbering any 1148 more complex than it already is. 1150 Finally, the internal operation of the home network should also not 1151 depend on the availability of the ISP network at any given time, 1152 other than of course for connectivity to services or systems off the 1153 home network. This reinforces the use of ULAs for stable internal 1154 communication, and the need for a naming and service discovery 1155 mechanism that can operate independently within the homenet. 1157 3.4.2. Stable internal IP addresses 1159 The network should by default attempt to provide IP-layer 1160 connectivity between all internal parts of the homenet as well as to 1161 and from the external Internet, subject to the filtering policies or 1162 other policy constraints discussed later in the security section. 1164 ULAs should be used within the scope of a homenet to support stable 1165 routing and connectivity between subnets and hosts regardless of 1166 whether a globally unique ISP-provided prefix is available. In the 1167 case of a prolonged external connectivity outage, ULAs allow internal 1168 operations across routed subnets to continue. ULA addresses also 1169 allow constrained devices to create permanent relationships between 1170 IPv6 addresses, e.g., from a wall controller to a lamp, where 1171 symbolic host names would require additional non-volatile memory and 1172 updating global prefixes in sleeping devices might also be 1173 problematic. 1175 As discussed previously, it would be expected that ULAs would 1176 normally be used alongside one or more global prefixes in a homenet, 1177 such that hosts become multi-addressed with both globally unique and 1178 ULA prefixes. ULAs should be used for all devices, not just those 1179 intended to only have internal connectivity. Default address 1180 selection would then enable ULAs to be preferred for internal 1181 communications between devices that are using ULA prefixes generated 1182 within the same homenet. 1184 In cases where ULA prefixes are in use within a homenet but there is 1185 no external IPv6 connectivity (and thus no GUAs in use), 1186 recommendations ULA-5, L-3 and L-4 in RFC 6204 should be followed to 1187 ensure correct operation, in particular where the homenet may be 1188 dual-stack with IPv4 external connectivity. The use of the Route 1189 Information Option described in [RFC4191] provides a mechanism to 1190 advertise such more-specific ULA routes. 1192 The use of ULAs should be restricted to the homenet scope through 1193 filtering at the border(s) of the homenet, as mandated by RFC 6204 1194 requirement S-2. 1196 Note that it is possible that in some cases multiple /48 ULA prefixes 1197 may be in use within the same homenet, e.g., when the network is 1198 being deployed, perhaps also without external connectivity. In cases 1199 where multiple ULA /48's are in use, hosts need to know that each /48 1200 is local to the homenet, e.g., by inclusion in their local address 1201 selection policy table. 1203 3.4.3. Internal prefix delegation 1205 As mentioned above, there are various sources of prefixes. From the 1206 homenet perspective, a single global prefix from each ISP should be 1207 received on the border CER [RFC3633]. Where multiple CERs exist with 1208 multiple ISP prefix pools, it is expected that routers within the 1209 homenet would assign themselves prefixes from each ISP they 1210 communicate with/through. As discussed above, a ULA prefix should be 1211 provisioned for stable internal communications or for use on 1212 constrained/LLN networks. 1214 The delegation or availability of a prefix pool to the homenet should 1215 allow subsequent internal autonomous delegation of prefixes for use 1216 within the homenet. Such internal delegation should not assume a 1217 flat or hierarchical model, nor should it make an assumption about 1218 whether the delegation of internal prefixes is distributed or 1219 centralised. The assignment mechanism should provide reasonable 1220 efficiency, so that typical home network prefix allocation sizes can 1221 accommodate all the necessary /64 allocations in most cases, and not 1222 waste prefixes. Further, duplicate assignment of multiple /64s to 1223 the same network should be avoided, and the network should behave as 1224 gracefully as possible in the event of prefix exhaustion (though the 1225 options in such cases may be limited). 1227 Where the home network has multiple CERs and these are delegated 1228 prefix pools from their attached ISPs, the internal prefix delegation 1229 would be expected to be served by each CER for each prefix associated 1230 with it. Where ULAs are used, it is preferable that only one /48 ULA 1231 covers the whole homenet, from which /64's can be delegated to the 1232 subnets. In cases where two /48 ULAs are generated within a homenet, 1233 the network should still continue to function, meaning that hosts 1234 will need to determine that each ULA is local to the homenet. 1236 Delegation within the homenet should result in each link being 1237 assigned a stable prefix that is persistent across reboots, power 1238 outages and similar short-term outages. The availability of 1239 persistent prefixes should not depend on the router boot order. The 1240 addition of a new routing device should not affect existing 1241 persistent prefixes, but persistence may not be expected in the face 1242 of significant 'replumbing' of the homenet. However, delegated ULA 1243 prefixes within the homenet should remain persistent through an ISP- 1244 driven renumbering event. 1246 Provisioning such persistent prefixes may imply the need for stable 1247 storage on routing devices, and also a method for a home user to 1248 'reset' the stored prefix should a significant reconfiguration be 1249 required (though ideally the home user should not be involved at 1250 all). 1252 This document makes no specific recommendation towards solutions, but 1253 notes that it is very likely that all routing devices participating 1254 in a homenet must use the same internal prefix delegation method. 1255 This implies that only one delegation method should be in use. 1257 3.4.4. Coordination of configuration information 1259 The network elements will need to be integrated in a way that takes 1260 account of the various lifetimes on timers that are used on different 1261 elements, e.g., DHCPv6 PD, router, valid prefix and preferred prefix 1262 timers. 1264 3.4.5. Privacy 1266 If ISPs offer relatively stable IPv6 prefixes to customers, the 1267 network prefix part of addresses associated with the homenet may not 1268 change over a reasonably long period of time. 1270 The exposure of which traffic is sourced from the same homenet is 1271 thus similar to IPv4; the single IPv4 global address seen through use 1272 of IPv4 NAT gives the same hint as the global IPv6 prefix seen for 1273 IPv6 traffic. 1275 While IPv4 NAT may obfuscate to an external observer which internal 1276 devices traffic is sourced from, IPv6, even with use of Privacy 1277 Addresses [RFC4941], adds additional exposure of which traffic is 1278 sourced from the same internal device, through use of the same IPv6 1279 source address for a period of time. 1281 3.5. Routing functionality 1283 Routing functionality is required when there are multiple routers 1284 deployed within the internal home network. This functionality could 1285 be as simple as the current 'default route is up' model of IPv4 NAT, 1286 or, more likely, it would involve running an appropriate routing 1287 protocol. Regardless of the solution method, the functionality 1288 discussed below should be met. 1290 The homenet unicast routing protocol should be based on a previously 1291 deployed protocol that has been shown to be reliable and robust, and 1292 that allows lightweight implementations, but that does not preclude 1293 the selection of a newer protocol for which a high quality open 1294 source implementation becomes available. Using information 1295 distributed through the routing protocol, each node in the homenet 1296 should be able to build a graph of the topology of the whole homenet 1297 including attributes such as links, nodes, connectivity, and (if 1298 supported by the protocol in use) link metrics. 1300 The routing protocol should support the generic use of multiple 1301 customer Internet connections, and the concurrent use of multiple 1302 delegated prefixes. A routing protocol that can make routing 1303 decisions based on source and destination addresses is thus 1304 desirable, to avoid upstream ISP BCP 38 ingress filtering problems. 1305 Multihoming support should also include load-balancing to multiple 1306 providers, and failover from a primary to a backup link when 1307 available. The protocol however should not require upstream ISP 1308 connectivity to be established to continue routing within the 1309 homenet. 1311 Routing within the homenet will determine the least cost path across 1312 the homenet towards the destination address given the source address. 1313 The path cost will be computed as a linear sum of the metric assigned 1314 to each link. The metric may be configured or automatically derived 1315 per link based on consideration of factors such as worst-case queue 1316 depth and router processing capabilities. 1318 Multiple types of physical interfaces must be accounted for in the 1319 homenet routed topology. Technologies such as Ethernet, WiFi, 1320 Multimedia over Coax Alliance (MoCA), etc. must be capable of 1321 coexisting in the same environment and should be treated as part of 1322 any routed deployment. The inclusion of physical layer 1323 characteristics including bandwidth, loss, and latency in path 1324 computation should be considered for optimising communication in the 1325 homenet. 1327 The routing environment should be self-configuring, as discussed 1328 previously. Minimising convergence time should be a goal in any 1329 routed environment, but as a guideline a maximum convergence time at 1330 most 30 seconds should be the target (this target is somewhat 1331 arbitrary, and was chosen based on how long a typical home user might 1332 wait before attempting another reset; ideally the routers might have 1333 some status light indicating they are converging, similar to an ADSL 1334 router light indicating it is establishing a connection to its ISP). 1336 Homenets may use a variety of underlying link layer technologies, and 1337 may therefore benefit from being able to use link metrics if 1338 available. It may be beneficial for traffic to use multiple paths to 1339 a given destination within the homenet where available, rather than a 1340 single best path. 1342 At most one routing protocol should be in use at a given time in a 1343 given homenet. In some simple topologies, no routing protocol may be 1344 needed. If more than one routing protocol is supported by routers in 1345 a given homenet, then a mechanism is required to ensure that all 1346 routers in that homenet use the same protocol. 1348 An appropriate mechanism is required to discover which router(s) in 1349 the homenet are providing the CER function. Borders may include but 1350 are not limited to the interface to the upstream ISP, a gateway 1351 device to a separate home network such as a LLN network, or a gateway 1352 to a guest or private corporate extension network. In some cases 1353 there may be no border present, which may for example occur before an 1354 upstream connection has been established. The border discovery 1355 functionality may be integrated into the routing protocol itself, but 1356 may also be imported via a separate discovery mechanism. 1358 Ideally, LLN or other logically separate networks should be able 1359 exchange routes such that IP traffic may be forwarded among the 1360 networks via gateway routers which interoperate with both the homenet 1361 and the LLN. Current home deployments use largely different 1362 mechanisms in sensor and basic Internet connectivity networks. IPv6 1363 virtual machine (VM) solutions may also add additional routing 1364 requirements. 1366 3.5.1. Multicast support 1368 It is desirable that, subject to the capacities of devices on certain 1369 media types, multicast routing is supported across the homenet, 1370 including source-specific multicast (SSM) [RFC4607]. 1372 [RFC4291] requires that any boundary of scope 4 or higher (i.e., 1373 admin-local or higher) be administratively configured. Thus the 1374 boundary at the homenet-ISP border must be administratively 1375 configured, though that may be triggered by an administrative 1376 function such as DHCP-PD. Other multicast forwarding policy borders 1377 may also exist within the homenet, e.g., to/from a guest subnet, 1378 whilst the use of certain link media types may also affect where 1379 specific multicast traffic is forwarded or routed. 1381 There may be different drivers for multicast to be supported across 1382 the homenet, e.g., for homenet-wide service discovery should a 1383 multicast service discovery protocol of scope greater than link-local 1384 be defined, or potentially for multicast-based streaming or 1385 filesharing applications. Where multicast is routed across a homenet 1386 an appropriate multicast routing protocol is required, one that as 1387 per the unicast routing protocol should be self-configuring. As 1388 hinted above, it must be possible to scope or filter multicast 1389 traffic to avoid it being flooded to network media where devices 1390 cannot reasonably support it. 1392 A homenet may not only use multicast internally, it may also be a 1393 consumer or provider of external multicast traffic, where the 1394 homenet's ISP supports such multicast operation. This may be 1395 valuable for example where live video applications are being sourced 1396 to/from the homenet. 1398 The multicast environment should support the ability for applications 1399 to pick a unique multicast group to use. 1401 3.6. Security 1403 The security of an IPv6 homenet is an important consideration. The 1404 most notable difference to the IPv4 operational model is the removal 1405 of NAT, the introduction of global addressability of devices, and 1406 thus a need to consider whether devices should have global 1407 reachability. Regardless, hosts need to be able to operate securely, 1408 end-to-end where required, and also be robust against malicious 1409 traffic directed towards them. However, there are other challenges 1410 introduced, e.g., default filtering policies at the borders between 1411 various homenet realms. 1413 3.6.1. Addressability vs reachability 1415 An IPv6-based home network architecture should embrace the 1416 transparent end-to-end communications model as described in 1417 [RFC2775]. Each device should be globally addressable, and those 1418 addresses must not be altered in transit. However, security 1419 perimeters can be applied to restrict end-to-end communications, and 1420 thus while a host may be globally addressable it may not be globally 1421 reachable. 1423 [RFC4864] describes a 'Simple Security' model for IPv6 networks, 1424 whereby stateful perimeter filtering can be applied to control the 1425 reachability of devices in a homenet. RFC 4864 states in Section 4.2 1426 that "the use of firewalls ... is recommended for those that want 1427 boundary protection in addition to host defences". It should be 1428 noted that a 'default deny' filtering approach would effectively 1429 replace the need for IPv4 NAT traversal protocols with a need to use 1430 a signalling protocol to request a firewall hole be opened, e.g., a 1431 protocol such as PCP [RFC6887]. In networks with multiple CERs, the 1432 signalling would need to handle the cases of flows that may use one 1433 or more exit routers. CERs would need to be able to advertise their 1434 existence for such protocols. 1436 [RFC6092] expands on RFC 4864, giving a more detailed discussion of 1437 IPv6 perimeter security recommendations, without mandating a 'default 1438 deny' approach. Indeed, RFC 6092 does not enforce a particular mode 1439 of operation, instead stating that CERs must provide an easily 1440 selected configuration option that permits a 'transparent' mode, thus 1441 ensuring a 'default allow' model is available. 1443 The topic of whether future home networks as described in this 1444 document should have have a 'default deny' or 'default allow' 1445 position has been discussed at length in various IETF meetings 1446 without any consensus being reached on which approach is more 1447 appropriate. Further, the choice of which default to apply may be 1448 situational, and thus this text makes no recommendation on the 1449 default setting beyond what is written on this topic in RFC 6092. We 1450 note in Section 3.6.3 below that the implicit firewall function of an 1451 IPv4 NAT is commonplace today, and thus future CERs targeted at home 1452 networks should continue to support the option of running in 'default 1453 deny mode', whether or not that is the default setting 1455 3.6.2. Filtering at borders 1457 It is desirable that there are mechanisms to detect different types 1458 of borders within the homenet, as discussed previously, and further 1459 mechanisms to then apply different types of filtering policies at 1460 those borders, e.g., whether naming and service discovery should pass 1461 a given border. Any such policies should be able to be easily 1462 applied by typical home users, e.g., to give a user in a guest 1463 network access to media services in the home, or access to a printer. 1464 Simple mechanisms to apply policy changes, or associations between 1465 devices, will be required. 1467 There are cases where full internal connectivity may not be 1468 desirable, e.g., in certain utility networking scenarios, or where 1469 filtering is required for policy reasons against guest network 1470 subnet(s). Some scenarios/models may as a result involve running 1471 isolated subnet(s) with their own CERs. In such cases connectivity 1472 would only be expected within each isolated network (though traffic 1473 may potentially pass between them via external providers). 1475 LLNs provide an another example of where there may be secure 1476 perimeters inside the homenet. Constrained LLN nodes may implement 1477 network key security but may depend on access policies enforced by 1478 the LLN border router. 1480 Considerations for differentiating neighbouring homenets are 1481 discussed in Section 3.3.1. 1483 3.6.3. Partial Effectiveness of NAT and Firewalls 1485 Security by way of obscurity (address translation) or through 1486 firewalls (filtering) is at best only partially effective. The very 1487 poor security track record of home computer, home networking and 1488 business PC computers and networking is testimony to this. A 1489 security compromise behind the firewall of any device exposes all 1490 others, making an entire network that relies on obscurity or a 1491 firewall as vulnerable as the most insecure device on the private 1492 side of the network. 1494 However, given current evidence of home network products with very 1495 poor default device security, putting a firewall in place does 1496 provide some level of protection. The use of firewalls today, 1497 whether a good practice or not, is common practice and the capability 1498 to afford protection via a 'default deny' setting, even if marginally 1499 effective, should not be lost. Thus, while it is highly desirable 1500 that all hosts in a homenet be adequately protected by built-in 1501 security functions, it should also be assumed that all CERs will 1502 continue to support appropriate perimeter defence functions, as per 1503 [RFC7084]. 1505 3.6.4. Exfiltration concerns 1507 As homenets become more complex, with more devices, and with service 1508 discovery potentially enabled across the whole home, there are 1509 potential concerns over the leakage of information should devices use 1510 discovery protocols to gather information and report it to equipment 1511 vendors or application service providers. 1513 While it is not clear how such exfiltration could be easily avoided, 1514 the threat should be recognised, be it from a new piece of hardware 1515 or some 'app' installed on a personal device. 1517 3.6.5. Device capabilities 1519 In terms of the devices, homenet hosts should implement their own 1520 security policies in accordance to their computing capabilities. 1521 They should have the means to request transparent communications to 1522 be able to be initiated to them through security filters in the 1523 homenet, either for all ports or for specific services. Users should 1524 have simple methods to associate devices to services that they wish 1525 to operate transparently through (CER) borders. 1527 3.6.6. ULAs as a hint of connection origin 1529 As noted in Section 3.6, if appropriate filtering is in place on the 1530 CER(s), as mandated by RFC 6204 requirement S-2, a ULA source address 1531 may be taken as an indication of locally sourced traffic. This 1532 indication could then be used with security settings to designate 1533 between which nodes a particular application is allowed to 1534 communicate, provided ULA address space is filtered appropriately at 1535 the boundary of the realm. 1537 3.7. Naming and Service Discovery 1539 The homenet requires devices to be able to determine and use unique 1540 names by which they can be accessed on the network, and which are not 1541 used by other devices on the network. Users and devices will need to 1542 be able to discover devices and services available on the network, 1543 e.g., media servers, printers, displays or specific home automation 1544 devices. Thus naming and service discovery must be supported in the 1545 homenet, and, given the nature of typical home network users, the 1546 service(s) providing this function must as far as possible support 1547 unmanaged operation. 1549 The naming system will be required to work internally or externally, 1550 be the user within the homenet or outside it, i.e., the user should 1551 be able to refer to devices by name, and potentially connect to them, 1552 wherever they may be. The most natural way to think about such 1553 naming and service discovery is to enable it to work across the 1554 entire homenet residence (site), disregarding technical borders such 1555 as subnets but respecting policy borders such as those between guest 1556 and other internal network realms. Remote access may be desired by 1557 the homenet residents while travelling, but also potentially by 1558 manufacturers or other 'benevolent' third parties. 1560 3.7.1. Discovering services 1562 Users will typically perform service discovery through graphical user 1563 interfaces (GUIs) that allow them to browse services on their network 1564 in an appropriate and intuitive way. Devices may also need to 1565 discover other devices, without any user intervention or choice. 1566 Either way, such interfaces are beyond the scope of this document, 1567 but the interface should have an appropriate application programming 1568 interface (API) for the discovery to be performed. 1570 Such interfaces may also typically hide the local domain name element 1571 from users, especially where only one name space is available. 1572 However, as we discuss below, in some cases the ability to discover 1573 available domains may be useful. 1575 We note that current zero-configuration service discovery protocols 1576 are generally aimed at single subnets. There is thus a choice to 1577 make for multi-subnet homenets as to whether such protocols should be 1578 proxied or extended to operate across a whole homenet. In this 1579 context, that may mean bridging a link-local method, taking care to 1580 avoid packets entering looping paths, or extending the scope of 1581 multicast traffic used for the purpose. It may mean that some proxy 1582 or hybrid service is utilised, perhaps co-resident on the CER. Or it 1583 may be that a new approach is preferable, e.g., flooding information 1584 around the homenet as attributes within the routing protocol (which 1585 could allow per-prefix configuration). However, we should prefer 1586 approaches that are backwardly compatible, and allow current 1587 implementations to continue to be used. Note that this document does 1588 not mandate a particular solution, rather it expresses the principles 1589 that should be used for a homenet naming and service discovery 1590 environment. 1592 One of the primary challenges facing service discovery today is lack 1593 of interoperability due to the ever increasing number of service 1594 discovery protocols available. While it is conceivable for consumer 1595 devices to support multiple discovery protocols, this is clearly not 1596 the most efficient use of network and computational resources. One 1597 goal of the homenet architecture should be a path to service 1598 discovery protocol interoperability either through a standards based 1599 translation scheme, hooks into current protocols to allow some for of 1600 communication among discovery protocols, extensions to support a 1601 central service repository in the homenet, or simply convergence 1602 towards a unified protocol suite. 1604 3.7.2. Assigning names to devices 1606 Given the large number of devices that may be networked in the 1607 future, devices should have a means to generate their own unique 1608 names within a homenet, and to detect clashes should they arise, 1609 e.g., where a second device of the same type/vendor as an existing 1610 device with the same default name is deployed, or where a new subnet 1611 is added to the homenet which already has a device of the same name. 1612 It is expected that a device should have a fixed name while within 1613 the scope of the homenet. 1615 Users will also want simple ways to (re)name devices, again most 1616 likely through an appropriate and intuitive interface that is beyond 1617 the scope of this document. Note the name a user assigns to a device 1618 may be a label that is stored on the device as an attribute of the 1619 device, and may be distinct from the name used in a name service, 1620 e.g., 'Study Laser Printer' as opposed to printer2.. 1622 3.7.3. The homenet name service 1624 The homenet name service should support both lookups and discovery. 1625 A lookup would operate via a direct query to a known service, while 1626 discovery may use multicast messages or a service where applications 1627 register in order to be found. 1629 It is highly desirable that the homenet name service must at the very 1630 least co-exist with the Internet name service. There should also be 1631 a bias towards proven, existing solutions. The strong implication is 1632 thus that the homenet service is DNS-based, or DNS-compatible. There 1633 are naming protocols that are designed to be configured and operate 1634 Internet-wide, like unicast-based DNS, but also protocols that are 1635 designed for zero-configuration local environments, like mDNS 1636 [RFC6762]. 1638 When DNS is used as the homenet name service, it typically includes 1639 both a resolving service and an authoritative service. The 1640 authoritative service hosts the homenet related zone. One approach 1641 when provisioning such a name service, which is designed to 1642 facilitate name resolution from the global Internet, is to run an 1643 authoritative name service on the CER and a secondary authoritative 1644 name service provided by the ISP or perhaps an external third party. 1646 Where zero configuration name services are used, it is desirable that 1647 these can also coexist with the Internet name service. In 1648 particular, where the homenet is using a global name space, it is 1649 desirable that devices have the ability, where desired, to add 1650 entries to that name space. There should also be a mechanism for 1651 such entries to be removed or expired from the global name space. 1653 To protect against attacks such as cache poisoning, where an attacker 1654 is able to insert a bogus DNS entry in the local cache, it is 1655 desirable to support appropriate name service security methods, 1656 including DNS Security Extensions (DNSSEC) [RFC4033], on both the 1657 authoritative server and the resolver sides. Where DNS is used, the 1658 homenet router or naming service must not prevent DNSSEC from 1659 operating. 1661 While this document does not specify hardware requirements, it is 1662 worth noting briefly here that e.g., in support of DNSSEC, 1663 appropriate homenet devices should have good random number generation 1664 capability, and future homenet specifications should indicate where 1665 high quality random number generators, i.e., with decent entropy, are 1666 needed. 1668 Finally, the impact of a change in CER must be considered. It would 1669 be desirable to retain any relevant state (configuration) that was 1670 held in the old CER. This might imply that state information should 1671 be distributed in the homenet, to be recoverable by/to the new CER, 1672 or to the homenet's ISP or a third party externally provided service 1673 by some means. 1675 3.7.4. Name spaces 1677 If access to homenet devices is required remotely from anywhere on 1678 the Internet, then at least one globally unique name space is 1679 required, though the use of multiple name spaces should not be 1680 precluded. One approach is that the name space(s) used for the 1681 homenet would be served authoritatively by the homenet, most likely 1682 by a server resident on the CER. Such name spaces may be acquired by 1683 the user or provided/generated by their ISP or an alternative 1684 externally provided service. It is likely that the default case is 1685 that a homenet will use a global domain provided by the ISP, but 1686 advanced users wishing to use a name space that is independent of 1687 their provider in the longer term should be able to acquire and use 1688 their own domain name. For users wanting to use their own 1689 independent domain names, such services are already available. 1691 Devices may also be assigned different names in different name 1692 spaces, e.g., by third parties who may manage systems or devices in 1693 the homenet on behalf of the resident(s). Remote management of the 1694 homenet is out of scope of this document. 1696 If however a global name space is not available, the homenet will 1697 need to pick and use a local name space which would only have meaning 1698 within the local homenet (i.e., it would not be used for remote 1699 access to the homenet). The .local name space currently has a 1700 special meaning for certain existing protocols which have link-local 1701 scope, and is thus not appropriate for multi-subnet home networks. A 1702 different name space is thus required for the homenet. 1704 One approach for picking a local name space is to use an Ambiguous 1705 Local Qualified Domain Name (ALQDN) space, such as .sitelocal (or an 1706 appropriate name reserved for the purpose). While this is a simple 1707 approach, there is the potential in principle for devices that are 1708 bookmarked somehow by name by an application in one homenet to be 1709 confused with a device with the same name in another homenet. In 1710 practice however the underlying service discovery protocols should be 1711 capable of handling moving to a network where a new device is using 1712 the same name as a device used previously in another homenet. 1714 An alternative approach for a local name space would be to use a 1715 Unique Locally Qualified Domain Name (ULQDN) space such as 1716 ..sitelocal. The could be generated in 1717 a variety of ways, one potentially being based on the local /48 ULA 1718 prefix being used across the homenet. Such a should 1719 survive a cold restart, i.e., be consistent after a network power- 1720 down, or, if a value is not set on startup, the CER or device running 1721 the name service should generate a default value. It would be 1722 desirable for the homenet user to be able to override the 1723 with a value of their choice, but that would increase 1724 the likelihood of a name conflict. Any generated 1725 should not be predictable; thus adding a salt/hash function would be 1726 desirable. 1728 In the (likely) event that the homenet is accessible from outside the 1729 homenet (using the global name space), it is vital that the homenet 1730 name space follow the rules and conventions of the global name space. 1731 In this mode of operation, names in the homenet (including those 1732 automatically generated by devices) must be usable as labels in the 1733 global name space. [RFC5890] describes considerations for 1734 Internationalizing Domain Names in Applications (IDNA). 1736 Also, with the introduction of new 'dotless' top level domains, there 1737 is also potential for ambiguity between, for example, a local host 1738 called 'computer' and (if it is registered) a .computer gTLD. Thus 1739 qualified names should always be used, whether these are exposed to 1740 the user or not. The IAB has issued a statement which explains why 1741 dotless domains should be considered harmful [IABdotless]. 1743 There may be use cases where either different name spaces may be 1744 desired for different realms in the homenet, or for segmentation of a 1745 single name space within the homenet. Thus hierarchical name space 1746 management is likely to be required. There should also be nothing to 1747 prevent individual device(s) being independently registered in 1748 external name spaces. 1750 It may be the case that if there are two or more CERs serving the 1751 home network, that if each has name space delegated from a different 1752 ISP there is the potential for devices in the home to have multiple 1753 fully qualified names under multiple domains. 1755 Where a user is in a remote network wishing to access devices in 1756 their home network, there may be a requirement to consider the domain 1757 search order presented where multiple associated name spaces exist. 1758 This also implies that a domain discovery function is desirable. 1760 It may be the case that not all devices in the homenet are made 1761 available by name via an Internet name space, and that a 'split view' 1762 (as described in [RFC6950] Section 4) is preferred for certain 1763 devices, whereby devices inside the homenet see different DNS 1764 responses to those outside. 1766 Finally, this document makes no assumption about the presence or 1767 omission of a reverse lookup service. There is an argument that it 1768 may be useful for presenting logging information to users with 1769 meaningful device names rather than literal addresses. There are 1770 also some services, most notably email mail exchangers, where some 1771 operators have chosen to require a valid reverse lookup before 1772 accepting connections. 1774 3.7.5. Independent operation 1776 Name resolution and service discovery for reachable devices must 1777 continue to function if the local network is disconnected from the 1778 global Internet, e.g., a local media server should still be available 1779 even if the Internet link is down for an extended period. This 1780 implies the local network should also be able to perform a complete 1781 restart in the absence of external connectivity, and have local 1782 naming and service discovery operate correctly. 1784 The approach described above of a local authoritative name service 1785 with a cache would allow local operation for sustained ISP outages. 1787 Having an independent local trust anchor is desirable, to support 1788 secure exchanges should external connectivity be unavailable. 1790 A change in ISP should not affect local naming and service discovery. 1791 However, if the homenet uses a global name space provided by the ISP, 1792 then this will obviously have an impact if the user changes their 1793 network provider. 1795 3.7.6. Considerations for LLNs 1797 In some parts of the homenet, in particular LLNs or any devices where 1798 battery power is used, devices may be sleeping, in which case a proxy 1799 for such nodes may be required, that could respond (for example) to 1800 multicast service discovery requests. Those same devices or parts of 1801 the network may have less capacity for multicast traffic that may be 1802 flooded from other parts of the network. In general, message 1803 utilisation should be efficient considering the network technologies 1804 and constrained devices that the service may need to operate over. 1806 There are efforts underway to determine naming and discovery 1807 solutions for use by the Constrained Application Protocol (CoAP) 1808 [I-D.ietf-core-coap] in LLN networks. These are outside the scope of 1809 this document. 1811 3.7.7. DNS resolver discovery 1813 Automatic discovery of a name service to allow client devices in the 1814 homenet to resolve external domains on the Internet is required, and 1815 such discovery must support clients that may be a number of router 1816 hops away from the name service. Similarly it may be desirable to 1817 convey any DNS domain search list that may be in effect for the 1818 homenet. 1820 3.7.8. Devices roaming to/from the homenet 1822 It is likely that some devices which have registered names within the 1823 homenet Internet name space and that are mobile will attach to the 1824 Internet at other locations and acquire an IP address at those 1825 locations. Devices may move between different homenets. In such 1826 cases it is desirable that devices may be accessed by the same name 1827 as is used in their home network. 1829 Solutions to this problem are not discussed in this document. They 1830 may include use of Mobile IPv6 or Dynamic DNS, either of which would 1831 put additional requirements on to the homenet, or establishment of a 1832 (VPN) tunnel to a server in the home network. 1834 3.8. Other Considerations 1836 This section discusses two other considerations for home networking 1837 that the architecture should not preclude, but that this text is 1838 neutral towards. 1840 3.8.1. Quality of Service 1842 Support for Quality of Service in a multi-service homenet may be a 1843 requirement, e.g., for a critical system (perhaps healthcare 1844 related), or for differentiation between different types of traffic 1845 (file sharing, cloud storage, live streaming, VoIP, etc). Different 1846 link media types may have different such properties or capabilities. 1848 However, homenet scenarios should require no new Quality of Service 1849 protocols. A DiffServ [RFC2475] approach with a small number of 1850 predefined traffic classes may generally be sufficient, though at 1851 present there is little experience of Quality of Service deployment 1852 in home networks. It is likely that QoS, or traffic prioritisation, 1853 methods will be required at the CER, and potentially around 1854 boundaries between different link media types (where for example some 1855 traffic may simply not be appropriate for some media, and need to be 1856 dropped to avoid overloading the constrained media). 1858 There may also be complementary mechanisms that could be beneficial 1859 to application performance and behaviour in the homenet domain, such 1860 as ensuring proper buffering algorithms are used as described in 1861 [Gettys11]. 1863 3.8.2. Operations and Management 1865 In this section we briefly review some initial considerations for 1866 operations and management in the type of homenet described in this 1867 document. It is expected that a separate document will define an 1868 appropriate operations and management framework for such homenets. 1870 As described in this document, the homenet should have the general 1871 goal of being self-organising and configuring from the network layer 1872 perspective, e.g. prefixes should be able to be assigned to router 1873 interfaces. Further, applications running on devices should be able 1874 to use zero configuration service discovery protocols to discover 1875 services of interest to the home user. In contrast, a home user 1876 would not be expected, for example, to have to assign prefixes to 1877 links, or manage the DNS entries for the home network. Such expert 1878 operation should not be precluded, but it is not the norm. 1880 The user may still be required to, or wish to, perform some 1881 configuration of the network and the devices on it. Examples might 1882 include entering a security key to enable access to their wireless 1883 network, or choosing to give a 'friendly name' to a device presented 1884 to them through service discovery. Configuration of link layer and 1885 application layer services is out of scope of this architectural 1886 principles document, but are likely to be required in an operational 1887 homenet. 1889 While not being expected to actively configure the networking 1890 elements of their homenet, users may be interested in being able to 1891 view the status of their networks and the devices connected to it, in 1892 which case appropriate network monitoring protocols will be required 1893 to allow them to view their network, and its status, e.g. via a web 1894 interface or equivalent. While the user may not understand how the 1895 network operates, it is reasonable to assume they are interested in 1896 understanding what faults or problems may exist on it. Such 1897 monitoring may extend to other devices on the network, e.g. storage 1898 devices, or web cameras, but such devices are beyond the scope of 1899 this document. 1901 It may also be the case that an ISP, or a third party, might wish to 1902 offer a remote management service for the homenet on behalf of the 1903 user, or to be able to assist the user in event of some problem they 1904 are experiencing, in which case appropriate management and monitoring 1905 protocols would be required. 1907 Specifying the required protocols to facilitate homenet management 1908 and monitoring is out of scope of this document. As stated above, it 1909 is expected that a separate document will be produced to describe the 1910 operations and management framework for the types of home network 1911 presented in this document. 1913 As a final point, we note that it is desirable that all network 1914 management and monitoring functions should be available over IPv6 1915 transport, even where the homenet is dual-stack. 1917 3.9. Implementing the Architecture on IPv6 1919 This architecture text encourages re-use of existing protocols. Thus 1920 the necessary mechanisms are largely already part of the IPv6 1921 protocol set and common implementations, though there are some 1922 exceptions. 1924 For automatic routing, it is expected that solutions can be found 1925 based on existing protocols. Some relatively smaller updates are 1926 likely to be required, e.g., a new mechanism may be needed in order 1927 to turn a selected protocol on by default, a mechanism may be 1928 required to automatically assign prefixes to links within the 1929 homenet. 1931 Some functionality, if required by the architecture, may need more 1932 significant changes or require development of new protocols, e.g., 1933 support for multihoming with multiple exit routers would likely 1934 require extensions to support source and destination address based 1935 routing within the homenet. 1937 Some protocol changes are however required in the architecture, e.g., 1938 for name resolution and service discovery, extensions to existing 1939 zero configuration link-local name resolution protocols are needed to 1940 enable them to work across subnets, within the scope of the home 1941 network site. 1943 Some of the hardest problems in developing solutions for home 1944 networking IPv6 architectures include discovering the right borders 1945 where the 'home' domain ends and the service provider domain begins, 1946 deciding whether some of the necessary discovery mechanism extensions 1947 should affect only the network infrastructure or also hosts, and the 1948 ability to turn on routing, prefix delegation and other functions in 1949 a backwards compatible manner. 1951 4. Conclusions 1953 This text defines principles and requirements for a homenet 1954 architecture. The principles and requirements documented here should 1955 be observed by any future texts describing homenet protocols for 1956 routing, prefix management, security, naming or service discovery. 1958 5. Security Considerations 1960 Security considerations for the homenet architecture are discussed in 1961 Section 3.6 above. 1963 6. IANA Considerations 1965 This document has no actions for IANA. 1967 7. References 1969 7.1. Normative References 1971 [RFC2460] Deering, S. and R. Hinden, "Internet Protocol, Version 6 1972 (IPv6) Specification", RFC 2460, December 1998. 1974 [RFC3633] Troan, O. and R. Droms, "IPv6 Prefix Options for Dynamic 1975 Host Configuration Protocol (DHCP) version 6", RFC 3633, 1976 December 2003. 1978 [RFC4193] Hinden, R. and B. Haberman, "Unique Local IPv6 Unicast 1979 Addresses", RFC 4193, October 2005. 1981 [RFC4291] Hinden, R. and S. Deering, "IP Version 6 Addressing 1982 Architecture", RFC 4291, February 2006. 1984 7.2. Informative References 1986 [RFC1918] Rekhter, Y., Moskowitz, R., Karrenberg, D., Groot, G., and 1987 E. Lear, "Address Allocation for Private Internets", 1988 BCP 5, RFC 1918, February 1996. 1990 [RFC2475] Blake, S., Black, D., Carlson, M., Davies, E., Wang, Z., 1991 and W. Weiss, "An Architecture for Differentiated 1992 Services", RFC 2475, December 1998. 1994 [RFC2775] Carpenter, B., "Internet Transparency", RFC 2775, 1995 February 2000. 1997 [RFC2827] Ferguson, P. and D. Senie, "Network Ingress Filtering: 1998 Defeating Denial of Service Attacks which employ IP Source 1999 Address Spoofing", BCP 38, RFC 2827, May 2000. 2001 [RFC3002] Mitzel, D., "Overview of 2000 IAB Wireless Internetworking 2002 Workshop", RFC 3002, December 2000. 2004 [RFC3022] Srisuresh, P. and K. Egevang, "Traditional IP Network 2005 Address Translator (Traditional NAT)", RFC 3022, 2006 January 2001. 2008 [RFC4033] Arends, R., Austein, R., Larson, M., Massey, D., and S. 2009 Rose, "DNS Security Introduction and Requirements", 2010 RFC 4033, March 2005. 2012 [RFC4191] Draves, R. and D. Thaler, "Default Router Preferences and 2013 More-Specific Routes", RFC 4191, November 2005. 2015 [RFC4192] Baker, F., Lear, E., and R. Droms, "Procedures for 2016 Renumbering an IPv6 Network without a Flag Day", RFC 4192, 2017 September 2005. 2019 [RFC4607] Holbrook, H. and B. Cain, "Source-Specific Multicast for 2020 IP", RFC 4607, August 2006. 2022 [RFC4862] Thomson, S., Narten, T., and T. Jinmei, "IPv6 Stateless 2023 Address Autoconfiguration", RFC 4862, September 2007. 2025 [RFC4864] Van de Velde, G., Hain, T., Droms, R., Carpenter, B., and 2026 E. Klein, "Local Network Protection for IPv6", RFC 4864, 2027 May 2007. 2029 [RFC4941] Narten, T., Draves, R., and S. Krishnan, "Privacy 2030 Extensions for Stateless Address Autoconfiguration in 2031 IPv6", RFC 4941, September 2007. 2033 [RFC5533] Nordmark, E. and M. Bagnulo, "Shim6: Level 3 Multihoming 2034 Shim Protocol for IPv6", RFC 5533, June 2009. 2036 [RFC5890] Klensin, J., "Internationalized Domain Names for 2037 Applications (IDNA): Definitions and Document Framework", 2038 RFC 5890, August 2010. 2040 [RFC5969] Townsley, W. and O. Troan, "IPv6 Rapid Deployment on IPv4 2041 Infrastructures (6rd) -- Protocol Specification", 2042 RFC 5969, August 2010. 2044 [RFC6092] Woodyatt, J., "Recommended Simple Security Capabilities in 2045 Customer Premises Equipment (CPE) for Providing 2046 Residential IPv6 Internet Service", RFC 6092, 2047 January 2011. 2049 [RFC6144] Baker, F., Li, X., Bao, C., and K. Yin, "Framework for 2050 IPv4/IPv6 Translation", RFC 6144, April 2011. 2052 [RFC6145] Li, X., Bao, C., and F. Baker, "IP/ICMP Translation 2053 Algorithm", RFC 6145, April 2011. 2055 [RFC6177] Narten, T., Huston, G., and L. Roberts, "IPv6 Address 2056 Assignment to End Sites", BCP 157, RFC 6177, March 2011. 2058 [RFC6204] Singh, H., Beebee, W., Donley, C., Stark, B., and O. 2059 Troan, "Basic Requirements for IPv6 Customer Edge 2060 Routers", RFC 6204, April 2011. 2062 [RFC6296] Wasserman, M. and F. Baker, "IPv6-to-IPv6 Network Prefix 2063 Translation", RFC 6296, June 2011. 2065 [RFC6333] Durand, A., Droms, R., Woodyatt, J., and Y. Lee, "Dual- 2066 Stack Lite Broadband Deployments Following IPv4 2067 Exhaustion", RFC 6333, August 2011. 2069 [RFC6555] Wing, D. and A. Yourtchenko, "Happy Eyeballs: Success with 2070 Dual-Stack Hosts", RFC 6555, April 2012. 2072 [RFC6724] Thaler, D., Draves, R., Matsumoto, A., and T. Chown, 2073 "Default Address Selection for Internet Protocol Version 6 2074 (IPv6)", RFC 6724, September 2012. 2076 [RFC6762] Cheshire, S. and M. Krochmal, "Multicast DNS", RFC 6762, 2077 February 2013. 2079 [RFC6824] Ford, A., Raiciu, C., Handley, M., and O. Bonaventure, 2080 "TCP Extensions for Multipath Operation with Multiple 2081 Addresses", RFC 6824, January 2013. 2083 [RFC6887] Wing, D., Cheshire, S., Boucadair, M., Penno, R., and P. 2084 Selkirk, "Port Control Protocol (PCP)", RFC 6887, 2085 April 2013. 2087 [RFC6950] Peterson, J., Kolkman, O., Tschofenig, H., and B. Aboba, 2088 "Architectural Considerations on Application Features in 2089 the DNS", RFC 6950, October 2013. 2091 [RFC7084] Singh, H., Beebee, W., Donley, C., and B. Stark, "Basic 2092 Requirements for IPv6 Customer Edge Routers", RFC 7084, 2093 November 2013. 2095 [RFC7157] Troan, O., Miles, D., Matsushima, S., Okimoto, T., and D. 2096 Wing, "IPv6 Multihoming without Network Address 2097 Translation", RFC 7157, March 2014. 2099 [I-D.ietf-core-coap] 2100 Shelby, Z., Hartke, K., and C. Bormann, "Constrained 2101 Application Protocol (CoAP)", draft-ietf-core-coap-18 2102 (work in progress), June 2013. 2104 [IABdotless] 2105 "IAB Statement: Dotless Domains Considered Harmful", 2106 February 2013, . 2110 [Gettys11] 2111 Gettys, J., "Bufferbloat: Dark Buffers in the Internet", 2112 March 2011, 2113 . 2115 Appendix A. Acknowledgments 2117 The authors would like to thank Aamer Akhter, Mikael Abrahamsson, 2118 Mark Andrews, Dmitry Anipko, Ran Atkinson, Fred Baker, Ray Bellis, 2119 Teco Boot, John Brzozowski, Cameron Byrne, Brian Carpenter, Stuart 2120 Cheshire, Julius Chroboczek, Lorenzo Colitti, Robert Cragie, Elwyn 2121 Davies, Ralph Droms, Lars Eggert, Jim Gettys, Olafur Gudmundsson, 2122 Wassim Haddad, Joel M. Halpern, David Harrington, Lee Howard, Ray 2123 Hunter, Joel Jaeggli, Heather Kirksey, Ted Lemon, Acee Lindem, Kerry 2124 Lynn, Daniel Migault, Erik Nordmark, Michael Richardson, Mattia 2125 Rossi, Barbara Stark, Markus Stenberg, Sander Steffann, Don Sturek, 2126 Andrew Sullivan, Dave Taht, Dave Thaler, Michael Thomas, Mark 2127 Townsley, JP Vasseur, Curtis Villamizar, Dan Wing, Russ White, and 2128 James Woodyatt for their comments and contributions within homenet WG 2129 meetings and on the WG mailing list. An acknowledgement generally 2130 means that person's text made it in to the document, or was helpful 2131 in clarifying or reinforcing an aspect of the document. It does not 2132 imply that each contributor agrees with every point in the document. 2134 Appendix B. Changes 2136 This section will be removed in the final version of the text. 2138 B.1. Version 14 2140 Changes made include: 2142 o Changes for Adrian Farrell discuss/comment. 2144 o Very minor wordsmithing requested by Benoit for OAM text. 2146 o Very minor wordsmithing from IETF89 session. 2148 o Added note to support SSM. 2150 o Emphasised at most one routing protocol in use, possibly none. 2152 B.2. Version 13 2154 Changes made include: 2156 o Changes to address last outstanding IESG DISCUSSes/COMMENTs. 2158 B.3. Version 12 2160 Changes made include: 2162 o Fixed minor typo nits introduced in -11. 2164 o Elwyn Davies' gen-art review comments addressed. 2166 o Some further IESG DISCUSSes/COMMENTs addressed. 2168 B.4. Version 11 (after IESG review) 2170 Changes made include: 2172 o Jouni Korhonen's OPSDIR review comments addressed. 2174 o Elwyn Davies' gen-art review comments addressed. 2176 o Considered secdir review by Samiel Weiler; many points addressed. 2178 o Considered APPSDIR review. 2180 o Addressed a large number of IESG comments and discusses. 2182 B.5. Version 10 (after AD review) 2184 Changes made include: 2186 o Minor changes/clarifications resulting from AD review 2188 B.6. Version 09 (after WGLC) 2190 Changes made include: 2192 o Added note about multicast into or out of site 2194 o Removed further personal draft references, replaced with covering 2195 text 2197 o Routing functionality text updated to avoid ambiguity 2198 o Added note that devices away from homenet may tunnel home (via 2199 VPN) 2201 o Added note that homenets more exposed to provider renumbering than 2202 with IPv4 and NAT 2204 o Added note about devices that may be ULA-only until configured to 2205 be globally addressable 2207 o Removed paragraph about broken CERs that do not work with prefixes 2208 other than /64 2210 o Noted no recommendation on methods to convey prefix information is 2211 made in this text 2213 o Stated that this text does not recommend how to form largest 2214 possible subnets 2216 o Added text about homenet evolution and handling disparate media 2217 types 2219 o Rephrased NAT/firewall text on marginal effectiveness 2221 o Emphasised that multihoming may be to any number of ISPs 2223 B.7. Version 08 2225 Changes made include: 2227 o Various clarifications made in response to list comments 2229 o Added note on ULAs with IPv4, where no GUAs in use 2231 o Added note on naming and internationalisation (IDNA) 2233 o Added note on trust relationships when adding devices 2235 o Added note for MPTCP 2237 o Added various naming and SD notes 2239 o Added various notes on delegated ISP prefixes 2241 B.8. Version 07 2243 Changes made include: 2245 o Removed reference to NPTv6 in section 3.2.4. Instead now say it 2246 has an architectural cost to use in the earlier section, and thus 2247 it is not recommended for use in the homenet architecture. 2249 o Removed 'proxy or extend?' section. Included shorter text in main 2250 body, without mandating either approach for service discovery. 2252 o Made it clearer that ULAs are expected to be used alongside 2253 globals. 2255 o Removed reference to 'advanced security' as described in 2256 draft-vyncke-advanced-ipv6-security. 2258 o Balanced the text between ULQDN and ALQDN. 2260 o Clarify text does not assume default deny or allow on CER, but 2261 that either mode may be enabled. 2263 o Removed ULA-C reference for 'simple' addresses. Instead only 2264 suggested service discovery to find such devices. 2266 o Reiterated that single/multiple CER models to be supported for 2267 multihoming. 2269 o Reordered section 3.3 to improve flow. 2271 o Added recommendation that homenet is not allocated less than /60, 2272 and a /56 is preferable. 2274 o Tidied up first few intro sections. 2276 o Other minor edits from list feedback. 2278 B.9. Version 06 2280 Changes made include: 2282 o Stated that unmanaged goal is 'as far as possible'. 2284 o Added note about multiple /48 ULAs potentially being in use. 2286 o Minor edits from list feedback. 2288 B.10. Version 05 2290 Changes made include: 2292 o Some significant changes to naming and SD section. 2294 o Removed some expired drafts. 2296 o Added notes about issues caused by ISP only delegating a /64. 2298 o Recommended against using prefixes longer than /64. 2300 o Suggested CER asks for /48 by DHCP PD, even if it only receives 2301 less. 2303 o Added note about DS-Lite but emphasised transition is out of 2304 scope. 2306 o Added text about multicast routing. 2308 B.11. Version 04 2310 Changes made include: 2312 o Moved border section from IPv6 differences to principles section. 2314 o Restructured principles into areas. 2316 o Added summary of naming and service discovery discussion from WG 2317 list. 2319 B.12. Version 03 2321 Changes made include: 2323 o Various improvements to the readability. 2325 o Removed bullet lists of requirements, as requested by chair. 2327 o Noted 6204bis has replaced advanced-cpe draft. 2329 o Clarified the topology examples are just that. 2331 o Emphasised we are not targetting walled gardens, but they should 2332 not be precluded. 2334 o Also changed text about requiring support for walled gardens. 2336 o Noted that avoiding falling foul of ingress filtering when 2337 multihomed is desirable. 2339 o Improved text about realms, detecting borders and policies at 2340 borders. 2342 o Stated this text makes no recommendation about default security 2343 model. 2345 o Added some text about failure modes for users plugging things 2346 arbitrarily. 2348 o Expanded naming and service discovery text. 2350 o Added more text about ULAs. 2352 o Removed reference to version 1 on chair feedback. 2354 o Stated that NPTv6 adds architectural cost but is not a homenet 2355 matter if deployed at the CER. This text only considers the 2356 internal homenet. 2358 o Noted multihoming is supported. 2360 o Noted routers may not by separate devices, they may be embedded in 2361 devices. 2363 o Clarified simple and advanced security some more, and RFC 4864 and 2364 6092. 2366 o Stated that there should be just one secret key, if any are used 2367 at all. 2369 o For multihoming, support multiple CERs but note that routing to 2370 the correct CER to avoid ISP filtering may not be optimal within 2371 the homenet. 2373 o Added some ISPs renumber due to privacy laws. 2375 o Removed extra repeated references to Simple Security. 2377 o Removed some solution creep on RIOs/RAs. 2379 o Load-balancing scenario added as to be supported. 2381 B.13. Version 02 2383 Changes made include: 2385 o Made the IPv6 implications section briefer. 2387 o Changed Network Models section to describe properties of the 2388 homenet with illustrative examples, rather than implying the 2389 number of models was fixed to the six shown in 01. 2391 o Text to state multihoming support focused on single CER model. 2392 Multiple CER support is desirable, but not required. 2394 o Stated that NPTv6 not supported. 2396 o Added considerations section for operations and management. 2398 o Added bullet point principles/requirements to Section 3.4. 2400 o Changed IPv6 solutions must not adversely affect IPv4 to should 2401 not. 2403 o End-to-end section expanded to talk about "Simple Security" and 2404 borders. 2406 o Extended text on naming and service discovery. 2408 o Added reference to RFC 2775, RFC 6177. 2410 o Added reference to the new xmDNS draft. 2412 o Added naming/SD requirements from Ralph Droms. 2414 Authors' Addresses 2416 Tim Chown (editor) 2417 University of Southampton 2418 Highfield 2419 Southampton, Hampshire SO17 1BJ 2420 United Kingdom 2422 Email: tjc@ecs.soton.ac.uk 2424 Jari Arkko 2425 Ericsson 2426 Jorvas 02420 2427 Finland 2429 Email: jari.arkko@piuha.net 2430 Anders Brandt 2431 Sigma Designs 2432 Emdrupvej 26A, 1 2433 Copenhagen DK-2100 2434 Denmark 2436 Email: abr@sdesigns.dk 2438 Ole Troan 2439 Cisco Systems, Inc. 2440 Drammensveien 145A 2441 Oslo N-0212 2442 Norway 2444 Email: ot@cisco.com 2446 Jason Weil 2447 Time Warner Cable 2448 13820 Sunrise Valley Drive 2449 Herndon, VA 20171 2450 USA 2452 Email: jason.weil@twcable.com