idnits 2.17.1 draft-ietf-homenet-hybrid-proxy-zeroconf-00.txt: Checking boilerplate required by RFC 5378 and the IETF Trust (see https://trustee.ietf.org/license-info): ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/1id-guidelines.txt: ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/checklist : ---------------------------------------------------------------------------- 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 (March 5, 2015) is 3339 days in the past. Is this intentional? Checking references for intended status: Proposed Standard ---------------------------------------------------------------------------- (See RFCs 3967 and 4897 for information about using normative references to lower-maturity documents in RFCs) -- Looks like a reference, but probably isn't: '1' on line 564 -- Looks like a reference, but probably isn't: '0' on line 448 -- Looks like a reference, but probably isn't: '2' on line 440 -- Looks like a reference, but probably isn't: '3' on line 446 -- Looks like a reference, but probably isn't: '4' on line 446 == Outdated reference: A later version (-10) exists of draft-ietf-homenet-hncp-03 -- Obsolete informational reference (is this intentional?): RFC 3633 (Obsoleted by RFC 8415) Summary: 0 errors (**), 0 flaws (~~), 2 warnings (==), 7 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Homenet Working Group M. Stenberg 3 Internet-Draft 4 Intended status: Standards Track March 5, 2015 5 Expires: September 6, 2015 7 Auto-Configuration of a Network of Hybrid Unicast/Multicast DNS-Based 8 Service Discovery Proxy Nodes 9 draft-ietf-homenet-hybrid-proxy-zeroconf-00 11 Abstract 13 This document describes how a proxy functioning between Unicast DNS- 14 Based Service Discovery and Multicast DNS can be automatically 15 configured using an arbitrary network-level state sharing mechanism. 17 Status of This Memo 19 This Internet-Draft is submitted in full conformance with the 20 provisions of BCP 78 and BCP 79. 22 Internet-Drafts are working documents of the Internet Engineering 23 Task Force (IETF). Note that other groups may also distribute 24 working documents as Internet-Drafts. The list of current Internet- 25 Drafts is at http://datatracker.ietf.org/drafts/current/. 27 Internet-Drafts are draft documents valid for a maximum of six months 28 and may be updated, replaced, or obsoleted by other documents at any 29 time. It is inappropriate to use Internet-Drafts as reference 30 material or to cite them other than as "work in progress." 32 This Internet-Draft will expire on September 6, 2015. 34 Copyright Notice 36 Copyright (c) 2015 IETF Trust and the persons identified as the 37 document authors. All rights reserved. 39 This document is subject to BCP 78 and the IETF Trust's Legal 40 Provisions Relating to IETF Documents 41 (http://trustee.ietf.org/license-info) in effect on the date of 42 publication of this document. Please review these documents 43 carefully, as they describe your rights and restrictions with respect 44 to this document. Code Components extracted from this document must 45 include Simplified BSD License text as described in Section 4.e of 46 the Trust Legal Provisions and are provided without warranty as 47 described in the Simplified BSD License. 49 Table of Contents 51 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 2 52 2. Requirements language . . . . . . . . . . . . . . . . . . . . 3 53 3. Hybrid proxy - what to configure . . . . . . . . . . . . . . 3 54 3.1. Conflict resolution within network . . . . . . . . . . . 4 55 3.2. Per-link DNS-SD forward zone names . . . . . . . . . . . 4 56 3.3. Reasonable defaults . . . . . . . . . . . . . . . . . . . 5 57 3.3.1. Network-wide unique link name (scheme 1) . . . . . . 5 58 3.3.2. Node name (scheme 2) . . . . . . . . . . . . . . . . 5 59 3.3.3. Link name (scheme 2) . . . . . . . . . . . . . . . . 5 60 4. TLVs . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 61 4.1. DNS Delegated Zone TLV . . . . . . . . . . . . . . . . . 5 62 4.2. Domain Name TLV . . . . . . . . . . . . . . . . . . . . . 7 63 4.3. Node Name TLV . . . . . . . . . . . . . . . . . . . . . . 7 64 5. Desirable behavior . . . . . . . . . . . . . . . . . . . . . 7 65 5.1. DNS search path in DHCP requests . . . . . . . . . . . . 7 66 5.2. Hybrid proxy . . . . . . . . . . . . . . . . . . . . . . 8 67 5.3. Hybrid proxy network zeroconf daemon . . . . . . . . . . 8 68 6. Security Considerations . . . . . . . . . . . . . . . . . . . 8 69 7. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 9 70 8. References . . . . . . . . . . . . . . . . . . . . . . . . . 9 71 8.1. Normative references . . . . . . . . . . . . . . . . . . 9 72 8.2. Informative references . . . . . . . . . . . . . . . . . 9 73 Appendix A. Example configuration . . . . . . . . . . . . . . . 10 74 A.1. Used topology . . . . . . . . . . . . . . . . . . . . . . 10 75 A.2. Zero-configuration steps . . . . . . . . . . . . . . . . 10 76 A.3. TLV state . . . . . . . . . . . . . . . . . . . . . . . . 11 77 A.4. DNS zone . . . . . . . . . . . . . . . . . . . . . . . . 12 78 A.5. Interaction with hosts . . . . . . . . . . . . . . . . . 12 79 Appendix B. Implementation . . . . . . . . . . . . . . . . . . . 12 80 Appendix C. Why not just proxy Multicast DNS? . . . . . . . . . 13 81 C.1. General problems . . . . . . . . . . . . . . . . . . . . 13 82 C.2. Stateless proxying problems . . . . . . . . . . . . . . . 14 83 C.3. Stateful proxying problems . . . . . . . . . . . . . . . 14 84 Appendix D. Acknowledgements . . . . . . . . . . . . . . . . . . 14 85 Author's Address . . . . . . . . . . . . . . . . . . . . . . . . 15 87 1. Introduction 89 Section 3 ("Hybrid Proxy Operation") of [I-D.cheshire-dnssd-hybrid] 90 describes how to translate queries from Unicast DNS-Based Service 91 Discovery described in [RFC6763] to Multicast DNS described in 92 [RFC6762], and how to filter the responses and translate them back to 93 unicast DNS. 95 This document describes what sort of configuration the participating 96 hybrid proxy servers require, as well as how it can be provided using 97 any network-wide state sharing mechanism such as link-state routing 98 protocol or Home Networking Control Protocol [I-D.ietf-homenet-hncp]. 99 The document also describes a naming scheme which does not even need 100 to be same across the whole covered network to work as long as the 101 specified conflict resolution works. The scheme can be used to 102 provision both forward and reverse DNS zones which employ hybrid 103 proxy for heavy lifting. 105 This document does not go into low level encoding details of the 106 Type-Length-Value (TLV) data that we want synchronized across a 107 network. Instead, we just specify what needs to be available, and 108 assume every node that needs it has it available. 110 We go through the mandatory specification of the language used in 111 Section 2, then describe what needs to be configured in hybrid 112 proxies and participating DNS servers across the network in 113 Section 3. How the data is exchanged using arbitrary TLVs is 114 described in Section 4. Finally, some overall notes on desired 115 behavior of different software components is mentioned in Section 5. 117 2. Requirements language 119 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 120 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this 121 document are to be interpreted as described in [RFC2119]. 123 3. Hybrid proxy - what to configure 125 Beyond the low-level translation mechanism between unicast and 126 multicast service discovery, the hybrid proxy draft 127 [I-D.cheshire-dnssd-hybrid] describes just that there have to be NS 128 records pointing to hybrid proxy responsible for each link within the 129 covered network. 131 In zero-configuration case, choosing the links to be covered is also 132 non-trivial choice; we can use the border discovery functionality (if 133 available) to determine internal and external links. Or we can use 134 some other protocol's presence (or lack of it) on a link to determine 135 internal links within the covered network, and some other signs 136 (depending on the deployment) such as DHCPv6 Prefix Delegation (as 137 described in [RFC3633]) to determine external links that should not 138 be covered. 140 For each covered link we want forward DNS zone delegation to an 141 appropriate node which is connected to a link, and running hybrid 142 proxy. Therefore the links' forward DNS zone names should be unique 143 across the network. We also want to populate reverse DNS zone 144 similarly for each IPv4 or IPv6 prefix in use. 146 There should be DNS-SD browse domain list provided for the network's 147 domain which contains each physical link only once, regardless of how 148 many nodes and hybrid proxy implementations are connected to it. 150 Yet another case to consider is the list of DNS-SD domains that we 151 want hosts to enumerate for browse domain lists. Typically, it 152 contains only the local network's domain, but there may be also other 153 networks we may want to pretend to be local but are in different 154 scope, or controlled by different organization. For example, a home 155 user might see both home domain's services (TBD-TLD), as well as 156 ISP's services under isp.example.com. 158 3.1. Conflict resolution within network 160 Any naming-related choice on node may have conflicts in the network 161 given that we require only distributed loosely synchronized database. 162 We assume only that the underlying protocol used for synchronization 163 has some concept of precedence between nodes originating conflicting 164 information, and in case of conflict, the higher precedence node MUST 165 keep the name they have chosen. The one(s) with lower precedence 166 MUST either try different one (that is not in use at all according to 167 the current link state information), or choose not to publish the 168 name altogether. 170 If a node needs to pick a different name, any algorithm works, 171 although simple algorithm choice is just like the one described in 172 Multicast DNS[RFC6762]: append -2, -3, and so forth, until there are 173 no conflicts in the network for the given name. 175 3.2. Per-link DNS-SD forward zone names 177 How to name the links of a whole network in automated fashion? Two 178 different approaches seem obvious: 180 1. Unique link name based - (unique-link).(domain). 182 2. Node and link name - (link).(unique-node).(domain). 184 The first choice is appealing as it can be much more friendly 185 (especially given manual configuration). For example, it could mean 186 just lan.example.com and wlan.example.com for a simple home network. 187 The second choice, on the other hand, has a nice property of being 188 local choice as long as node name can be made unique. 190 The type of naming scheme to use can be left as implementation 191 option. And the actual names themselves SHOULD be also overridable, 192 if the end-user wants to customize them in some way. 194 3.3. Reasonable defaults 196 Note that any manual configuration, which SHOULD be possible, MUST 197 override the defaults provided here or chosen by the creator of the 198 implementation. 200 3.3.1. Network-wide unique link name (scheme 1) 202 It is not obvious how to produce network-wide unique link names for 203 the (unique-link).(domain) scheme. One option would be to base it on 204 type of physical network layer, and then hope that the number of the 205 networks won't be significant enough to confuse (e.g. "lan", or 206 "wlan"). 208 The network-wide unique link names should be only used in small 209 networks. Given a larger network, after conflict resolution, 210 identifying which link is 'lan-42.example.com' may be challenging. 212 3.3.2. Node name (scheme 2) 214 Our recommendation is to use some short form which indicates the type 215 of node it is, for example, "openwrt.example.com". As the name is 216 visible to users, it should be kept as short as possible. In theory 217 even more exact model could be helpful, for example, "openwrt- 218 buffalo-wzr-600-dhr.example.com". In practice providing some other 219 records indicating exact node information (and access to management 220 UI) is more sensible. 222 3.3.3. Link name (scheme 2) 224 Recommendation for (link) portion of (link).(node).(domain) is to use 225 physical network layer type as base, or possibly even just interface 226 name on the node if it's descriptive enough. For example, 227 "eth0.openwrt.example.com" and "wlan0.openwrt.example.com" may be 228 good enough. 230 4. TLVs 232 To implement this specification fully, support for following three 233 different TLVs is needed. However, only the DNS Delegated Zone TLVs 234 MUST be supported, and the other two SHOULD be supported. 236 4.1. DNS Delegated Zone TLV 238 This TLV is effectively a combined NS and A/AAAA record for a zone. 239 It MUST be supported by implementations conforming to this 240 specification. Implementations SHOULD provide forward zone per link 241 (or optimizing a bit, zone per link with Multicast DNS traffic). 243 Implementations MAY provide reverse zone per prefix using this same 244 mechanism. If multiple nodes advertise same reverse zone, it should 245 be assumed that they all have access to the link with that prefix. 246 However, as noted in Section 5.3, mainly only the node with highest 247 precedence on the link should publish this TLV. 249 Contents: 251 o Address field is IPv6 address (e.g. 2001:db8::3) or IPv4 address 252 mapped to IPv6 address (e.g. ::FFFF:192.0.2.1) where the 253 authoritative DNS server for Zone can be found. If the address 254 field is all zeros, the Zone is under global DNS hierarchy and can 255 be found using normal recursive name lookup starting at the 256 authoritative root servers (This is mostly relevant with the S bit 257 below). 259 o S-bit indicates that this delegated zone consists of a full DNS-SD 260 domain, which should be used as base for DNS-SD domain enumeration 261 (that is, (field)._dns-sd._udp.(zone) exists). Forward zones MAY 262 have this set. Reverse zones MUST NOT have this set. This can be 263 used to provision DNS search path to hosts for non-local services 264 (such as those provided by ISP, or other manually configured 265 service providers). 267 o B-bit indicates that this delegated zone should be included in 268 network's DNS-SD browse list of domains at b._dns- 269 sd._udp.(domain). Local forward zones SHOULD have this set. 270 Reverse zones SHOULD NOT have this set. 272 o L-bit indicates that this delegated zone should be included in the 273 network's DNS-SD legacy browse list of domains at lb._dns- 274 sd._udp.(DOMAIN-NAME). Local forward zones SHOULD have this bit 275 set, reverse zones SHOULD NOT. 277 o Zone is the label sequence of the zone, encoded according to 278 section 3.1. ("Name space definitions") of [RFC1035]. Note that 279 name compression is not required here (and would not have any 280 point in any case), as we encode the zones one by one. The zone 281 MUST end with an empty label. 283 In case of a conflict (same zone being advertised by multiple parties 284 with different address or bits), conflict should be addressed 285 according to Section 3.1. 287 4.2. Domain Name TLV 289 This TLV is used to indicate the base (domain) to be used for the 290 network. If multiple nodes advertise different ones, the conflict 291 resolution rules in Section 3.1 should result in only the one with 292 highest precedence advertising one, eventually. In case of such 293 conflict, user SHOULD be notified somehow about this, if possible, 294 using the configuration interface or some other notification 295 mechanism for the nodes. Like the Zone field in Section 4.1, the 296 Domain Name TLV's contents consist of a single DNS label sequence. 298 This TLV SHOULD be supported if at all possible. It may be derived 299 using some future DHCPv6 option, or be set by manual configuration. 300 Even on nodes without manual configuration options, being able to 301 read the domain name provided by a different node could make the user 302 experience better due to consistent naming of zones across the 303 network. 305 By default, if no node advertises domain name TLV, hard-coded default 306 (TBD) should be used. 308 4.3. Node Name TLV 310 This TLV is used to advertise a node's name. After the conflict 311 resolution procedure described in Section 3.1 finishes, there should 312 be exactly zero to one nodes publishing each node name. The contents 313 of the TLV should be a single DNS label. 315 This TLV SHOULD be supported if at all possible. If not supported, 316 and another node chooses to use the (link).(node) naming scheme with 317 this node's name, the contents of the network's domain may look 318 misleading (but due to conflict resolution of per-link zones, still 319 functional). 321 If the node name has been configured manually, and there is a 322 conflict, user SHOULD be notified somehow about this, if possible, 323 using the configuration interface or some other notification 324 mechanism for the nodes. 326 5. Desirable behavior 328 5.1. DNS search path in DHCP requests 330 The nodes following this specification SHOULD provide the used 331 (domain) as one item in the search path to it's hosts, so that DNS-SD 332 browsing will work correctly. They also SHOULD include any DNS 333 Delegated Zone TLVs' zones, that have S bit set. 335 5.2. Hybrid proxy 337 The hybrid proxy implementation SHOULD support both forward zones, 338 and IPv4 and IPv6 reverse zones. It SHOULD also detect whether or 339 not there are any Multicast DNS entities on a link, and make that 340 information available to the network zeroconf daemon (if implemented 341 separately). This can be done by (for example) passively monitoring 342 traffic on all covered links, and doing infrequent service 343 enumerations on links that seem to be up, but without any Multicast 344 DNS traffic (if so desired). 346 Hybrid proxy nodes MAY also publish it's own name via Multicast DNS 347 (both forward A/AAAA records, as well as reverse PTR records) to 348 facilitate applications that trace network topology. 350 5.3. Hybrid proxy network zeroconf daemon 352 The daemon should avoid publishing TLVs about links that have no 353 Multicast DNS traffic to keep the DNS-SD browse domain list as 354 concise as possible. It also SHOULD NOT publish delegated zones for 355 links for which zones already exist by another node with higher 356 precedence. 358 The daemon (or other entity with access to the TLVs) SHOULD generate 359 zone information for DNS implementation that will be used to serve 360 the (domain) zone to hosts. Domain Name TLV described in Section 4.2 361 should be used as base for the zone, and then all DNS Delegated Zones 362 described in Section 4.1 should be used to produce the rest of the 363 entries in zone (see Appendix A.4 for example interpretation of the 364 TLVs in Appendix A.3. 366 6. Security Considerations 368 There is a trade-off between security and zero-configuration in 369 general; if used network state synchronization protocol is not 370 authenticated (and in zero-configuration case, it most likely is 371 not), it is vulnerable to local spoofing attacks. We assume that 372 this scheme is used either within (lower layer) secured networks, or 373 with not-quite-zero-configuration initial set-up. 375 If some sort of dynamic inclusion of links to be covered using border 376 discovery or such is used, then effectively service discovery will 377 share fate with border discovery (and also security issues if any). 379 7. IANA Considerations 381 This document has no actions for IANA. 383 8. References 385 8.1. Normative references 387 [I-D.cheshire-dnssd-hybrid] 388 Cheshire, S., "Hybrid Unicast/Multicast DNS-Based Service 389 Discovery", draft-cheshire-dnssd-hybrid-01 (work in 390 progress), January 2014. 392 [RFC1035] Mockapetris, P., "Domain names - implementation and 393 specification", STD 13, RFC 1035, November 1987. 395 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 396 Requirement Levels", BCP 14, RFC 2119, March 1997. 398 [RFC6762] Cheshire, S. and M. Krochmal, "Multicast DNS", RFC 6762, 399 February 2013. 401 [RFC6763] Cheshire, S. and M. Krochmal, "DNS-Based Service 402 Discovery", RFC 6763, February 2013. 404 8.2. Informative references 406 [I-D.ietf-homenet-hncp] 407 Stenberg, M., Barth, S., and P. Pfister, "Home Networking 408 Control Protocol", draft-ietf-homenet-hncp-03 (work in 409 progress), January 2015. 411 [RFC3633] Troan, O. and R. Droms, "IPv6 Prefix Options for Dynamic 412 Host Configuration Protocol (DHCP) version 6", RFC 3633, 413 December 2003. 415 [RFC3646] Droms, R., "DNS Configuration options for Dynamic Host 416 Configuration Protocol for IPv6 (DHCPv6)", RFC 3646, 417 December 2003. 419 8.3. URIs 421 [1] https://github.com/sbyx/hnetd/ 423 Appendix A. Example configuration 425 A.1. Used topology 427 Let's assume home network that looks like this: 429 |[0] 430 +-----+ 431 | CER | 432 +-----+ 433 [1]/ \[2] 434 / \ 435 +-----+ +-----+ 436 | IR1 |-| IR2 | 437 +-----+ +-----+ 438 |[3]| |[4]| 440 We're not really interested about links [0], [1] and [2], or the 441 links between IRs. Given the optimization described in Section 4.1, 442 they should not produce anything to network's Multicast DNS state 443 (and therefore to DNS either) as there isn't any Multicast DNS 444 traffic there. 446 The user-visible set of links are [3] and [4]; each consisting of a 447 LAN and WLAN link. We assume that ISP provides 2001:db8:1234::/48 448 prefix to be delegated in the home via [0]. 450 A.2. Zero-configuration steps 452 Given implementation that chooses to use the second naming scheme 453 (link).(node).(domain), and no configuration whatsoever, here's what 454 happens (the steps are interleaved in practice but illustrated here 455 in order): 457 1. Network-level state synchronization protocol runs, nodes get 458 effective precedences. For ease of illustration, CER winds up 459 with 2, IR1 with 3, and IR2 with 1. 461 2. Prefix delegation takes place. IR1 winds up with 462 2001:db8:1234:11::/64 for LAN and 2001:db8:1234:12::/64 for WLAN. 463 IR2 winds up with 2001:db8:1234:21::/64 for LAN and 464 2001:db8:1234:22::/64 for WLAN. 466 3. IR1 is assumed to be reachable at 2001:db8:1234:11::1 and IR2 at 467 2001:db8:1234:21::1. 469 4. Each node wants to be called 'node' due to lack of branding in 470 drafts. They announce that using the node name TLV defined in 471 Section 4.3. They also advertise their local zones, but as that 472 information may change, it's omitted here. 474 5. Conflict resolution ensues. As IR1 has precedence over the rest, 475 it becomes "node". CER and IR2 have to rename, and (depending on 476 timing) one of them becomes "node-2" and other one "node-3". Let 477 us assume IR2 is "node-2". During conflict resolution, each node 478 publishes TLVs for it's own set of delegated zones. 480 6. CER learns ISP-provided domain "isp.example.com" using DHCPv6 481 domain list option defined in [RFC3646]. The information is 482 passed along as S-bit enabled delegated zone TLV. 484 A.3. TLV state 486 Once there is no longer any conflict in the system, we wind up with 487 following TLVs (NN is used as abbreviation for Node Name, and DZ for 488 Delegated Zone TLVs): 490 (from CER) 491 DZ {s=1,zone="isp.example.com"} 493 (from IR1) 494 NN {name="node"} 496 DZ {address=2001:db8:1234:11::1, b=1, 497 zone="lan.node.example.com."} 498 DZ {address=2001:db8:1234:11::1, 499 zone="1.1.0.0.4.3.2.1.8.b.d.0.1.0.0.2.ip6.arpa."} 501 DZ {address=2001:db8:1234:11::1, b=1, 502 zone="wlan.node.example.com."} 503 DZ {address=2001:db8:1234:11::1, 504 zone="2.1.0.0.4.3.2.1.8.b.d.0.1.0.0.2.ip6.arpa."} 506 (from IR2) 507 NN {name="node-2"} 509 DZ {address=2001:db8:1234:21::1, b=1, 510 zone="lan.node-2.example.com."} 511 DZ {address=2001:db8:1234:21::1, 512 zone="1.2.0.0.4.3.2.1.8.b.d.0.1.0.0.2.ip6.arpa."} 514 DZ {address=2001:db8:1234:21::1, b=1, 515 zone="wlan.node-2.example.com."} 516 DZ {address=2001:db8:1234:21::1, 517 zone="2.2.0.0.4.3.2.1.8.b.d.0.1.0.0.2.ip6.arpa."} 519 A.4. DNS zone 521 In the end, we should wind up with following zone for (domain) which 522 is example.com in this case, available at all nodes, just based on 523 dumping the delegated zone TLVs as NS+AAAA records, and optionally 524 domain list browse entry for DNS-SD: 526 b._dns_sd._udp PTR lan.node 527 b._dns_sd._udp PTR wlan.node 529 b._dns_sd._udp PTR lan.node-2 530 b._dns_sd._udp PTR wlan.node-2 532 node AAAA 2001:db8:1234:11::1 533 node-2 AAAA 2001:db8:1234:21::1 535 node NS node 536 node-2 NS node-2 538 1.1.0.0.4.3.2.1.8.b.d.0.1.0.0.2.ip6.arpa. NS node.example.com. 539 2.1.0.0.4.3.2.1.8.b.d.0.1.0.0.2.ip6.arpa. NS node.example.com. 540 1.2.0.0.4.3.2.1.8.b.d.0.1.0.0.2.ip6.arpa. NS node-2.example.com. 541 2.2.0.0.4.3.2.1.8.b.d.0.1.0.0.2.ip6.arpa. NS node-2.example.com. 543 Internally, the node may interpret the TLVs as it chooses to, as long 544 as externally defined behavior follows semantics of what's given in 545 the above. 547 A.5. Interaction with hosts 549 So, what do the hosts receive from the nodes? Using e.g. DHCPv6 DNS 550 options defined in [RFC3646], DNS server address should be one (or 551 multiple) that point at DNS server that has the zone information 552 described in Appendix A.4. Domain list provided to hosts should 553 contain both "example.com" (the hybrid-enabled domain), as well as 554 the externally learned domain "isp.example.com". 556 When hosts start using DNS-SD, they should check both b._dns- 557 sd._udp.example.com, as well as b._dns-sd._udp.isp.example.com for 558 list of concrete domains to browse, and as a result services from two 559 different domains will seem to be available. 561 Appendix B. Implementation 563 There is an prototype implementation of this draft at hnetd github 564 repository [1] which contains variety of other homenet WG-related 565 things' implementation too. 567 Appendix C. Why not just proxy Multicast DNS? 569 Over the time number of people have asked me about how, why, and if 570 we should proxy (originally) link-local Multicast DNS over multiple 571 links. 573 At some point I meant to write a draft about this, but I think I'm 574 too lazy; so some notes left here for general amusement of people 575 (and to be removed if this ever moves beyond discussion piece). 577 C.1. General problems 579 There are two main reasons why Multicast DNS is not proxyable in the 580 general case. 582 First reason is the conflict resolution depends on the RRsets staying 583 constant. That is not possible across multiple links (due to e.g. 584 link-local addresses having to be filtered). Therefore, conflict 585 resolution breaks, or at least requires ugly hacks to work around. 587 A simple, but not really working workaround for this is to make sure 588 that in conflict resolution, propagated resources always loses. 589 Given that the proxy function only removes records, the result SHOULD 590 be consistently original set of records winning. Even with that, the 591 conflict resolution will effectively cease working, allowing for two 592 instances of same name to exist (as both think they 'own' the name 593 due to locally seen higher precedence). 595 Given some more extra logic, it is possible to make this work by 596 having proxies be aware of both the original record sets, and 597 effectively enforcing the correct conflict resolution results by (for 598 example) passing the unfiltered packets to the losing party just to 599 make sure they renumber, or by altering the RR sets so that they will 600 consistently win (by inserting some lower rrclass/rrtype records). 601 As the conflicts happen only in rrclass=1/rrtype=28, it is easy 602 enough to add e.g. extra TXT record (rrtype 16) to force precedence 603 even when removing the later rrtype 28 record. Obviously, this new 604 RRset must never wind up near the host with the higher precedence, or 605 it will cause spurious renaming loops. 607 Second reason is timing, which is relatively tight in the conflict 608 resolution phase, especially given lossy and/or high latency 609 networks. 611 C.2. Stateless proxying problems 613 In general, typical stateless proxy has to involve flooding, as 614 Multicast DNS assumes that most messages are received by every host. 615 And it won't scale very well, as a result. 617 The conflict resolution is also harder without state. It may result 618 in Multicast DNS responder being in constant probe-announce loop, 619 when it receives altered records, notes that it's the one that should 620 own the record. Given stateful proxying, this would be just a 621 transient problem but designing stateless proxy that won't cause this 622 is non-trivial exercise. 624 C.3. Stateful proxying problems 626 One option is to write proxy that learns state from one link, and 627 propagates it in some way to other links in the network. 629 A big problem with this case lies in the fact that due to conflict 630 resolution concerns above, it is easy to accidentally send packets 631 that will (possibly due to host mobility) wind up at the originator 632 of the service, who will then perform renaming. That can be 633 alleviated, though, given clever hacks with conflict resolution 634 order. 636 The stateful proxying may be also too slow to occur within the 637 timeframe allocated for announcing, leading to excessive later 638 renamings based on delayed finding of duplicate services with same 639 name 641 A work-around exists for this though; if the game doesn't work for 642 you, don't play it. One option would be simply not to propagate ANY 643 records for which conflict has seen even once. This would work, but 644 result in rather fragile, lossy service discovery infrastructure. 646 There are some other small nits too; for example, Passive Observation 647 Of Failure (POOF) will not work given stateful proxying. Therefore, 648 it leads to requiring somewhat shorter TTLs, perhaps. 650 Appendix D. Acknowledgements 652 Thanks to Stuart Cheshire for the original hybrid proxy draft and 653 interesting discussion in Orlando, where I was finally convinced that 654 stateful Multicast DNS proxying is a bad idea. 656 Also thanks to Mark Baugher, Ole Troan, Shwetha Bhandari and Gert 657 Doering for review comments. 659 Author's Address 661 Markus Stenberg 662 Helsinki 00930 663 Finland 665 Email: markus.stenberg@iki.fi