<?xml version="1.0"encoding="utf-8"?> <?xml-model href="rfc7991bis.rnc"?>encoding="UTF-8"?> <!--Required for schema validation and schema-aware editing --> <!-- <?xml-stylesheet type="text/xsl" href="rfc2629.xslt" ?> --> <!-- This third-party XSLT can be enabled for direct transformationsdraft submitted inXML processors, including most browsersxml v3 --> <!DOCTYPE rfc [ <!ENTITY nbsp " "> <!ENTITY zwsp "​"> <!ENTITY nbhy "‑"> <!ENTITY wj "⁠"> ]><!-- If further character entities are required then they should be added to the DOCTYPE above. Use of an external entity file is not recommended. --><rfc xmlns:xi="http://www.w3.org/2001/XInclude" category="info" docName="draft-ietf-v6ops-dhcp-pd-per-device-08" number="9663" ipr="trust200902" obsoletes="" updates="" submissionType="IETF" consensus="true" xml:lang="en"version="3">version="3" sortRefs="true" symRefs="true"> <front> <title abbrev="Prefix perclient usingClient Using DHCPv6-PD">Using DHCPv6PD">Using DHCPv6-PDPrefix Delegation (DHCPv6-PD) to Allocate Unique IPv6PrefixPrefixes per Client in Large Broadcast Networks</title> <seriesInfoname="Internet-Draft" value="draft-ietf-v6ops-dhcp-pd-per-device"/>name="RFC" value="9663"/> <author initials="L." surname="Colitti" fullname="Lorenzo Colitti"> <organization>Google, LLC</organization> <address> <postal> <street>Shibuya 3-21-3</street> <country>Japan</country> </postal> <email>lorenzo@google.com</email> </address> </author> <author fullname="Jen Linkova" initials="J" role="editor" surname="Linkova"> <organization>Google</organization> <address> <postal><!-- Reorder these if your country does things differently --><street>1 Darling Island Rd</street> <city>Pyrmont</city><region>NSW</region><region>New South Wales</region> <code>2009</code><country>AU</country><country>Australia</country> </postal> <email>furry13@gmail.com</email> <email>furry@google.com</email> </address> </author> <author fullname="Xiao Ma" initials="X" role="editor" surname="Ma"> <organization>Google</organization> <address> <postal> <street>Shibuya 3-21-3</street> <country>Japan</country> </postal> <email>xiaom@google.com</email> </address> </author> <date month="October" year="2024"/><area>OPS Area</area> <workgroup>v6ops Working Group</workgroup><area>OPS</area> <workgroup>v6ops</workgroup> <keyword>IPv6</keyword> <keyword>SLAAC</keyword> <keyword>DHCPv6-PD</keyword> <abstract> <t>This document discusses an IPv6 deployment scenario when individual nodes connected to large broadcast networks (such as enterprise networks or public Wi-Fi networks) are allocated unique prefixes via DHCPv6 Prefix Delegation(DHCPv6-PD, RFC8415).</t>(DHCPv6-PD), as specified in RFC 8415.</t> </abstract> </front> <middle> <section> <name>Introduction</name><t>Often<t>Often, broadcast networks such as enterprise or public Wi-Fi deployments place many devices on a shared link with a single on-link prefix. This document describes an alternative deployment model where individual devices obtain prefixes from the network. This provides two important advantages.</t> <t>First, it offers better scalability. Unlike IPv4, IPv6 allows hosts to have multiple addresses, and this is the case in most deployments (see <xref target="appendix"/> for more details). However, increasing the number of addresses introduces scalability issues on the network infrastructure. Network devices need to maintain various types oftables/hashestables and hashes (Neighbor Cache on first-hop routers, Neighbor Discovery Proxy caches on Layer 2devicesdevices, etc.). OnVXLAN <xref target="RFC7348"/>Virtual eXtensible Local Area Network (VXLAN) networks <xref target="RFC7348"/>, each address might be represented as aroute so 8route. This means, for example, that if every client has 10 addresses instead of1 requiresone, thedevices tonetwork must support810 times more routes, etc. If an infrastructuredevicedevice's resources are exhausted, the device might drop some IPv6 addresses from the corresponding tables, while the address owner mightbestill be using the address to send traffic. This leads to trafficblackholingbeing discarded and a degraded customer experience. Providing every host with one prefix allows the network to maintain only one entry per device, while still providing the device the ability to use an arbitrary number of addresses. </t> <t>Second,itthis deployment model provides the ability to extend the network. In IPv4, a device that connects to the network can provide connectivity to subtended devices by using NAT. With DHCPv6 Prefix Delegation(DHCPv6-PD,(DHCPv6-PD) <xreftarget="RFC8415"/>),target="RFC8415"/>, such a device can similarly extend the network, but unlike IPv4 NAT, it can provide its subtended devices with full end-to-end connectivity.</t> <t>Another method of deploying unique prefixes per device is documented in <xref target="RFC8273"/>. Similarly, the standard deployment model in cellular IPv6 networks <xref target="RFC6459"/> provides a unique prefix to every device. However, providing a unique prefix per device is very uncommon in enterprise-style networks, where nodes are usually connected to broadcastsegments/VLANssegments such as VLANs and each link has a single on-link prefix assigned. This document takes a similar approach to <xref target="RFC8273"/>, but allocates the prefix usingDHCPv6-PD. </t>DHCPv6-PD.</t> <t>This document focuses on thebehaviourbehavior of the network. Hostbehaviourbehavior is not defined in thisdocument. </t>document.</t> </section> <section> <name>Requirements Language</name><t>The<t> The key words"MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY","<bcp14>MUST</bcp14>", "<bcp14>MUST NOT</bcp14>", "<bcp14>REQUIRED</bcp14>", "<bcp14>SHALL</bcp14>", "<bcp14>SHALL NOT</bcp14>", "<bcp14>SHOULD</bcp14>", "<bcp14>SHOULD NOT</bcp14>", "<bcp14>RECOMMENDED</bcp14>", "<bcp14>NOT RECOMMENDED</bcp14>", "<bcp14>MAY</bcp14>", and"OPTIONAL""<bcp14>OPTIONAL</bcp14>" in this document are to be interpreted as described inBCP 14BCP 14 <xref target="RFC2119"/> <xref target="RFC8174"/> when, and only when, they appear in all capitals, as shownhere.</t>here. </t> </section> <section> <name>Terminology</name><t> Node: a<dl> <dt>Node:</dt><dd>a device that implementsIPv6,IPv6 <xreftarget="RFC8200"/>. </t> <t> Host: anytarget="RFC8200"/></dd> <dt>Host:</dt><dd>any node that is not arouter,router <xreftarget="RFC8200"/>. </t> <t> Client: atarget="RFC8200"/></dd> <dt>Client:</dt><dd>a nodewhichthat connects to a network and acquires addresses. The node may wish to obtain addresses for its own use, or it may be a router that wishes to extend the network to its physical or virtual subsystems, or both. It may be either a host or a router as defined by <xreftarget="RFC8200"/>. </t> <t> ND: Neighbor Discovery, <xref target="RFC4861"/>. </t> <t> NUD: Neighbor Unreachability Detection, <xref target="RFC4861"/>. </t> <t> PIO:target="RFC8200"/>.</dd> <dt>AP:</dt><dd>(wireless) Access Point</dd> <dt>DHCPv6 IA_NA:</dt><dd>Identity Association for Non-temporary Addresses (<xref target="RFC8415" sectionFormat="of" section="21.4"/>)</dd> <dt>DHCPv6 IA_PD:</dt><dd>Identity Association for PrefixInformation Option, <xref target="RFC4862"/>. </t> <t> SLAAC: IPv6 Stateless Address Autoconfiguration, <xref target="RFC4862"/>. </t> <t> DHCPv6-PD: DHCPv6Delegation (<xreftarget="RFC8415"/>)target="RFC8415" sectionFormat="of" section="21.21"/>)</dd> <dt>DHCPv6-PD:</dt><dd>DHCPv6 Prefix Delegation <xref target="RFC8415"/>; a mechanism to delegate IPv6 prefixes toclients. </t>clients.</dd> <dt>ND:</dt><dd>Neighbor Discovery <xref target="RFC4861"/></dd> <dt>NUD:</dt><dd>Neighbor Unreachability Detection <xref target="RFC4861"/></dd> <dt>PIO:</dt><dd>Prefix Information Option <xref target="RFC4862"/></dd> <dt>SLAAC:</dt><dd>IPv6 Stateless Address Autoconfiguration <xref target="RFC4862"/></dd> </dl> </section> <section> <name>Design Principles</name><t> Instead<t>Instead of all clients on a given link forming addresses from the same shared prefix assigned to thatlink:link, this deployment model operates as described below: </t> <ul><li> A<li>A device acts as a DHCPv6-PD client and requests a prefix via DHCPv6-PD by sending an IA_PDrequest. </li> <li> Therequest.</li> <li>The server delegates a prefix to the client and the delegated prefix is installed into the routing table of the first-hop router as a route pointing to the client's link-local address. The first-hop router can act as a DHCPv6 relay and snoop DHCPv6 Reply messages from an off-link DHCPv6 server, or it can act as a DHCPv6 server itself. In bothcasescases, it can install the route locally, and if the network is running a dynamic routing protocol, distribute the route or the entire prefix pool into the protocol.</li> <li>For the router and all other infrastructure devices, the delegated prefix is considered off-link, so traffic to that prefix does not trigger any ND packets, other than the minimum ND required to sustain Neighbor Unreachability Detection (NUD) for the client's link-localaddress. </li> <li> Theaddress.</li> <li>The device can use the delegated prefix in various ways. For example, it can form addresses, as described in<xref target="RFC7084"/>requirementWAA-7.WAA-7 of <xref target="RFC7084"/>. It can also extend the network, as described in <xref target="RFC7084"/> or <xreftarget="RFC7278"/>. </li>target="RFC7278"/>.</li> </ul><t> An<t>An example scenario is shown inFigure 1.<xref target="fig1"/>. Note that the prefix lengths used in the example are /64 because that is the prefix length currently supported by SLAAC and is not otherwise required by the proposed deploymentmodel. </t>model.</t> <figure anchor="fig1"><name> An<name>An Example Topology with Two First-HopRouters. </name>Routers</name> <artset> <artwork type="svg"> <svg xmlns="http://www.w3.org/2000/svg" version="1.1"height="928" width="856"height="784" width="576" viewBox="0 0856 928"576 784" class="diagram" text-anchor="middle" font-family="monospace" font-size="13px"> <path d="M8,1288,112 L8,288"8,224" fill="none" stroke="black"/> <path d="M8,73624,16 L8,912"24,64" fill="none" stroke="black"/> <path d="M24,33624,272 L24,432"24,336" fill="none" stroke="black"/> <path d="M24,81624,608 L24,880"24,768" fill="none" stroke="black"/> <path d="M48,4840,672 L48,176"40,736" fill="none" stroke="black"/> <path d="M48,22456,344 L48,336"56,416" fill="none" stroke="black"/> <path d="M48,44056,464 L48,592"56,608" fill="none" stroke="black"/> <path d="M48,64064,160 L48,736"64,272" fill="none" stroke="black"/> <path d="M168,8072,72 L168,192"72,128" fill="none" stroke="black"/> <path d="M168,256128,64 L168,328"128,160" fill="none" stroke="black"/> <path d="M168,432128,200 L168,592"128,264" fill="none" stroke="black"/> <path d="M168,640128,336 L168,728"128,496" fill="none" stroke="black"/> <path d="M216,816128,544 L216,880"128,600" fill="none" stroke="black"/> <path d="M224,528152,224 L224,736"152,272" fill="none" stroke="black"/> <path d="M232,816168,384 L232,880"168,608" fill="none" stroke="black"/> <path d="M248,288176,672 L248,336"176,736" fill="none" stroke="black"/> <path d="M256,16184,416 L256,96"184,480" fill="none" stroke="black"/> <path d="M304,432192,672 L304,528"192,736" fill="none" stroke="black"/> <path d="M320,576216,336 L320,656"216,384" fill="none" stroke="black"/> <path d="M352,240240,176 L352,328"240,264" fill="none" stroke="black"/> <path d="M424,816272,512 L424,880"272,576" fill="none" stroke="black"/> <path d="M432,336280,272 L432,432"280,336" fill="none" stroke="black"/> <path d="M440,736304,64 L440,912"304,112" fill="none" stroke="black"/> <path d="M448,96312,384 L448,128"312,416" fill="none" stroke="black"/> <path d="M456,512320,272 L456,576"320,336" fill="none" stroke="black"/> <path d="M480,336328,672 L480,432"328,736" fill="none" stroke="black"/> <path d="M480,864344,176 L480,912"344,264" fill="none" stroke="black"/> <path d="M520,240352,608 L520,328"352,768" fill="none" stroke="black"/> <path d="M536,688376,576 L536,768"376,672" fill="none" stroke="black"/> <path d="M584,832392,336 L584,864"392,384" fill="none" stroke="black"/> <path d="M608,768400,224 L608,832"400,272" fill="none" stroke="black"/> <path d="M616,576400,704 L616,656"400,752" fill="none" stroke="black"/> <path d="M632,432424,416 L632,528"424,480" fill="none" stroke="black"/> <path d="M640,768440,384 L640,800"440,512" fill="none" stroke="black"/> <path d="M656,16456,672 L656,96"456,704" fill="none" stroke="black"/> <path d="M656,864464,160 L656,912"464,272" fill="none" stroke="black"/> <path d="M680,288464,344 L680,336"464,400" fill="none" stroke="black"/> <path d="M680,528464,448 L680,688"464,512" fill="none" stroke="black"/> <path d="M680,864472,72 L680,912"472,128" fill="none" stroke="black"/> <path d="M736,80528,336 L736,160"528,432" fill="none" stroke="black"/> <path d="M736,208528,480 L736,336"528,504" fill="none" stroke="black"/> <path d="M736,440536,64 L736,592"536,160" fill="none" stroke="black"/> <path d="M736,640536,208 L736,688"536,264" fill="none" stroke="black"/> <path d="M760,832536,576 L760,864"536,640" fill="none" stroke="black"/> <path d="M808,432544,704 L808,592"544,752" fill="none" stroke="black"/> <path d="M808,640552,512 L808,680"552,576" fill="none" stroke="black"/> <path d="M824,32560,16 L824,208"560,64" fill="none" stroke="black"/> <path d="M824,256568,112 L824,328"568,224" fill="none" stroke="black"/> <path d="M848,128568,272 L848,288"568,336" fill="none" stroke="black"/> <path d="M848,33624,16 L848,432"560,16" fill="none" stroke="black"/> <path d="M848,68824,64 L848,768"560,64" fill="none" stroke="black"/> <path d="M848,8648,112 L848,912"568,112" fill="none" stroke="black"/> <path d="M256,168,224 L656,16"568,224" fill="none" stroke="black"/> <path d="M656,3224,272 L824,32"280,272" fill="none" stroke="black"/> <path d="M48,48320,272 L240,48" fill="none" stroke="black"/> <path d="M 168,80 L 256,80" fill="none" stroke="black"/> <path d="M 664,80 L 736,80" fill="none" stroke="black"/> <path d="M 256,96 L 656,96" fill="none" stroke="black"/> <path d="M 8,128 L 848,128" fill="none" stroke="black"/> <path d="M 8,288 L 848,288"568,272" fill="none" stroke="black"/> <path d="M 24,336 L432,336" fill="none" stroke="black"/> <path d="M 480,336 L 848,336" fill="none" stroke="black"/> <path d="M 24,432 L 432,432"280,336" fill="none" stroke="black"/> <path d="M480,432320,336 L848,432"568,336" fill="none" stroke="black"/> <path d="M192,528160,384 L696,528"448,384" fill="none" stroke="black"/> <path d="M320,576184,416 L616,576"424,416" fill="none" stroke="black"/> <path d="M320,656184,480 L616,656"424,480" fill="none" stroke="black"/> <path d="M536,688272,512 L848,688"552,512" fill="none" stroke="black"/> <path d="M8,736272,576 L440,736"552,576" fill="none" stroke="black"/> <path d="M536,76824,608 L848,768"352,608" fill="none" stroke="black"/> <path d="M24,81640,672 L216,816"176,672" fill="none" stroke="black"/> <path d="M232,816192,672 L424,816"328,672" fill="none" stroke="black"/> <path d="M544,832368,672 L784,832"544,672" fill="none" stroke="black"/> <path d="M480,864400,704 L656,864"544,704" fill="none" stroke="black"/> <path d="M680,86440,736 L848,864"176,736" fill="none" stroke="black"/> <path d="M24,880192,736 L216,880"328,736" fill="none" stroke="black"/> <path d="M232,880400,752 L424,880"544,752" fill="none" stroke="black"/> <path d="M8,91224,768 L440,912" fill="none" stroke="black"/> <path d="M 480,912 L 656,912" fill="none" stroke="black"/> <path d="M 680,912 L 848,912"352,768" fill="none" stroke="black"/> <polygon class="arrowhead"points="832,328 820,322.4 820,333.6"points="544,640 532,634.4 532,645.6" fill="black"transform="rotate(90,824,328)"/>transform="rotate(90,536,640)"/> <polygon class="arrowhead"points="832,208 820,202.4 820,213.6"points="544,264 532,258.4 532,269.6" fill="black"transform="rotate(90,824,208)"/>transform="rotate(90,536,264)"/> <polygon class="arrowhead"points="816,680 804,674.4 804,685.6"points="544,160 532,154.4 532,165.6" fill="black"transform="rotate(90,808,680)"/>transform="rotate(90,536,160)"/> <polygon class="arrowhead"points="816,592 804,586.4 804,597.6"points="536,504 524,498.4 524,509.6" fill="black"transform="rotate(90,808,592)"/>transform="rotate(90,528,504)"/> <polygon class="arrowhead"points="744,640 732,634.4 732,645.6"points="536,432 524,426.4 524,437.6" fill="black"transform="rotate(270,736,640)"/>transform="rotate(90,528,432)"/> <polygon class="arrowhead"points="744,440 732,434.4 732,445.6"points="480,72 468,66.4 468,77.6" fill="black"transform="rotate(270,736,440)"/>transform="rotate(270,472,72)"/> <polygon class="arrowhead"points="744,208 732,202.4 732,213.6"points="472,448 460,442.4 460,453.6" fill="black"transform="rotate(270,736,208)"/>transform="rotate(270,464,448)"/> <polygon class="arrowhead"points="672,80 660,74.4 660,85.6"points="472,344 460,338.4 460,349.6" fill="black"transform="rotate(180,664,80)"/>transform="rotate(270,464,344)"/> <polygon class="arrowhead"points="648,800 636,794.4 636,805.6"points="472,160 460,154.4 460,165.6" fill="black"transform="rotate(90,640,800)"/>transform="rotate(270,464,160)"/> <polygon class="arrowhead"points="528,328 516,322.4 516,333.6"points="352,264 340,258.4 340,269.6" fill="black"transform="rotate(90,520,328)"/>transform="rotate(90,344,264)"/> <polygon class="arrowhead"points="360,328 348,322.4 348,333.6"points="248,264 236,258.4 236,269.6" fill="black"transform="rotate(90,352,328)"/>transform="rotate(90,240,264)"/> <polygon class="arrowhead"points="248,48 236,42.4 236,53.6"points="136,600 124,594.4 124,605.6" fill="black"transform="rotate(0,240,48)"/>transform="rotate(90,128,600)"/> <polygon class="arrowhead"points="176,728 164,722.4 164,733.6"points="136,496 124,490.4 124,501.6" fill="black"transform="rotate(90,168,728)"/>transform="rotate(90,128,496)"/> <polygon class="arrowhead"points="176,592 164,586.4 164,597.6"points="136,264 124,258.4 124,269.6" fill="black"transform="rotate(90,168,592)"/>transform="rotate(90,128,264)"/> <polygon class="arrowhead"points="176,328 164,322.4 164,333.6"points="136,160 124,154.4 124,165.6" fill="black"transform="rotate(90,168,328)"/>transform="rotate(90,128,160)"/> <polygon class="arrowhead"points="176,192 164,186.4 164,197.6"points="80,72 68,66.4 68,77.6" fill="black"transform="rotate(90,168,192)"/>transform="rotate(270,72,72)"/> <polygon class="arrowhead"points="56,640 44,634.4 44,645.6"points="72,160 60,154.4 60,165.6" fill="black"transform="rotate(270,48,640)"/>transform="rotate(270,64,160)"/> <polygon class="arrowhead"points="56,440 44,434.4 44,445.6"points="64,464 52,458.4 52,469.6" fill="black"transform="rotate(270,48,440)"/>transform="rotate(270,56,464)"/> <polygon class="arrowhead"points="56,224 44,218.4 44,229.6"points="64,344 52,338.4 52,349.6" fill="black"transform="rotate(270,48,224)"/>transform="rotate(270,56,344)"/> <g class="text"> <textx="452"x="292" y="36">DHCPv6 Servers</text> <textx="448" y="68">Pool 2001:db8:ddd0::/48x="288" y="52">Pool 3fff:0:d::/48 forclients</text> <text x="452" y="84">on 2001:db8:c001::/64clients on 2001:db8:ff::/64 link</text> <textx="452" y="164">IPv6 Network</text>x="44" y="132">DHCPv6</text> <textx="724" y="180">DHCPv6</text> <text x="76" y="196">DHCPv6</text>x="308" y="132">IPv6 Network</text> <textx="720" y="196">Relay-Forward</text>x="436" y="132">DHCPv6</text> <textx="72" y="212">Relay-Forward</text>x="64" y="148">Relay-Forward</text> <textx="180" y="228">DHCPv6</text>x="464" y="148">Relay-Forward</text> <textx="444" y="228">Routex="288" y="164">Route for2001:db8:ddd0::/48</text>3fff:0:d::/48</text> <textx="796" y="228">DHCPv6</text>x="124" y="180">DHCPv6</text> <textx="184" y="244">Relay-Reply</text>x="532" y="180">DHCPv6</text> <textx="800" y="244">Relay-Reply</text>x="128" y="196">Relay-Reply</text> <textx="272" y="356">First-hopx="520" y="196">Relay-Reply</text> <text x="152" y="292">First-hop router/DHCPv6 relay</text> <textx="648" y="356">First-hopx="448" y="292">First-hop Router/DHCPv6 relay</text> <textx="220" y="388">Route 2001:db8:ddd0:1::/64x="152" y="308">3fff:0:d:1::/64 ->fe80::aaaa</text>fe80::aa</text> <textx="596" y="388">Route 2001:db8:ddd0:1::/64</text> <text x="768" y="388">-> fe80::aaaa</text> <text x="220" y="420">Route 2001:db8:ddd0:2::/64x="440" y="308">3fff:0:d:1::/64 ->fe80::cccc</text>fe80::aa</text> <textx="596" y="420">Route 2001:db8:ddd0:2::/64</text> <text x="768" y="420">-> fe80::cccc</text>x="152" y="324">3fff:0:d:2::/64 -> fe80::cc</text> <textx="388" y="516">Sharedx="440" y="324">3fff:0:d:2::/64 -> fe80::cc</text> <text x="308" y="356">Shared IPv6 link</text> <textx="540" y="516">2001:db8:c001::/64</text>x="308" y="372">2001:db8:ff::/64</text> <textx="456" y="596">legacy client B</text>x="476" y="420">DHCPv6</text> <text x="52"y="612">DHCPv6</text> <text x="164" y="612">DHCPv6</text> <text x="452" y="612">no DHCPv6-PD support</text>y="436">DHCPv6</text> <textx="740" y="612">DHCPv6</text>x="288" y="436">Client B (no DHCPv6-PD)</text> <textx="812" y="612">DHCPv6</text>x="480" y="436">Request</text> <text x="56"y="628">Request</text>y="452">Request</text> <text x="292" y="452">link-local address fe80::b</text> <textx="160" y="628">Reply</text>x="532" y="452">DHCPv6</text> <textx="464" y="628">link-localx="304" y="468">global addressfe80::bbbb</text>2001:db8:ff::b</text> <textx="744" y="628">Request</text>x="528" y="468">Reply</text> <textx="808" y="628">Reply</text>x="132" y="516">DHCPv6</text> <textx="468" y="644">global address 2001:db8:c001::bbbb</text>x="128" y="532">Reply</text> <textx="628" y="708">Clientx="372" y="532">Client C</text> <textx="664" y="724">link-localx="392" y="548">link-local addressfe80::cccc</text>fe80::cc</text> <textx="696" y="740">delegatedx="412" y="564">delegated prefix2001:db8:ddd0:2::/64</text>3fff:0:d:2::/64</text> <text x="500" y="612">Router</text> <textx="220" y="756">Clientx="172" y="628">Client A</text> <textx="212" y="772">link-localx="472" y="628">Advertisement</text> <text x="172" y="644">link-local address:fe80::aaaa</text>fe80::aa</text> <textx="228" y="788">delegatedx="468" y="644">containing PIO</text> <text x="184" y="660">delegated prefix:2001:db8:ddd0:1::/64</text>3fff:0:d:1::/64</text> <textx="732" y="788">Router Advertisement</text> <text x="748" y="804">PIO 2001:db8:ddd0:2::/64</text>x="472" y="660">3fff:0:d:2::/64</text> <text x="108"y="836">virtualy="692">virtual system</text> <textx="316" y="836">virtualx="260" y="692">virtual system</text> <textx="116" y="852">2001:db8:ddd0:1::f00</text> <text x="320" y="852">2001:db8:ddd0:1::2345</text> <text x="120" y="868">2001:db8:ddd0:1::cafe</text>x="108" y="708">3fff:0:d:1::de</text> <textx="316" y="868">2001:db8:ddd0:1::abc</text>x="260" y="708">3fff:0:d:1::ad</text> <textx="572" y="884">Tethered device1</text>x="108" y="724">3fff:0:d:1::ca</text> <textx="756" y="884">Tethered device2</text>x="260" y="724">3fff:0:d:1::fe</text> <textx="568" y="900">2001:db8:ddd0:2::5555</text>x="472" y="724">Tethered device</text> <textx="764" y="900">2001:db8:ddd0:2::6666</text>x="468" y="740">3fff:0:d:2::66</text> </g> </svg> </artwork> <artwork type="ascii-art"> <![CDATA[+-----------------------------------------------------------------------------++------------------------------------------------------------------+ | DHCPv6 Servers | | Pool2001:db8:ddd0::/483fff:0:d::/48 for clients on2001:db8:c001::/642001:db8:ff::/64 link |+---------------+-------------------------+-----------------------------+-----++------------+---------------------+----------------------------+--+ ^ | | ^ | | | | | |+-------+------+-------------------------+----------------------+------+-----+ |DHCPv6+-------+------+---------------------+--------------------+-------+---+ | DHCPv6| | IPv6 Network DHCPv6 | | | |Relay-Forward | Relay-Forward | | | ^ v Route for2001:db8:ddd0::/483fff:0:d::/48 ^ v | | | DHCPv6 | | | DHCPv6 | | | Relay-Reply | | | Relay-Reply| | | | | | | | |+-----+-------+-------+-----+--------------------+-----+-------+-------+-----++------+-------+--+----------+------------+------+-------+--------+---+ | | | | | | | | | v | v v | | v+----+---------------+--------------+ +--------------+-------+-------------++----+----------+---------------+ +---------+-------+------------+ | First-hop router/DHCPv6 relay | | First-hop Router/DHCPv6relay |relay| |2001:db8:ddd0:1::/643fff:0:d:1::/64 ->fe80::aaaa|fe80::aa |2001:db8:ddd0:1::/64| 3fff:0:d:1::/64 ->fe80::aaaafe80::aa | |2001:db8:ddd0:2::/643fff:0:d:2::/64 ->fe80::cccc|fe80::cc | |2001:db8:ddd0:2::/643fff:0:d:2::/64 ->fe80::cccfe80::cc |+-----------+------------+----------+ +---------+--------------------+-----++------------+----------+-------+ +--------+----------------+----+ ^ | | Shared IPv6link,link | ^ | | | |2001:db8:c001::/642001:db8:ff::/64 | | | | |--+-------+-----------+-----------+------+----+-----+-----------+---------+-----+- | | | | | | | | | | | |+----------------+--------------++---------------+-------------+ | DHCPv6 | DHCPv6 | ||Legacy| Client B (no DHCPv6-PD)client B| | Request v Request | | |link-local addressfe80::bbbbfe80::b | | ^ DHCPv6 ^ | | |global address2001:db8:c001::b|2001:db8:ff::b| | | Reply | | |+-------------------------------++-----------------------------+ | | | | v | | | v | DHCPv6 |+-------------------+-----+------------++--------------------+--+----------+ | Reply | | Client C | | | | | link-local addressfe80::ccccfe80::cc | | | | | delegated prefix2001:db8:ddd0:2::/64|3fff:0:d:2::/64 | | |+--------------+-+---------------------+ || +------------+-------------------+-+ | v | |+----+-------+----+-----------------------------+| +---+-------------+----------------------+ | RouterAdvertisement| | Client A | | Advertisement |containing PIO| link-local address:fe80::aaaafe80::aa | | containing PIO v2001:db8:ddd0:2::/64| delegated prefix:2001:db8:ddd0:1::/643fff:0:d:1::/64 | | 3fff:0:d:2::/64 |+---------------------+ +------------------++----------------+ +----------------+ |-+-----------+----------+---------+----------- | | virtual system | | virtual system | | | | |2001:db8:ddd0:1::f00| |2001:db8:ddd0:1::2|3fff:0:d:1::de | |+---------+-----------+3fff:0:d:1::ad | |2001:db8:ddd0:1::caf| |2001:db8:ddd0:1::a|+------+----------+ | | 3fff:0:d:1::ca | | 3fff:0:d:1::fe | | | Tethered device | |+---------------------+ +------------------++----------------+ +----------------+ | ||2001:db8:ddd0:2::6666|3fff:0:d:2::66 | |+---------------------+ +-----------------------------------------------+| +-----------------+ +----------------------------------------+ ]]> </artwork> </artset> </figure> </section> <section> <name>Applicability and Limitations</name><t> Delegating<t>Delegating a unique prefix per client provides all the benefits of both SLAAC and DHCPv6 address allocation, but at the cost of greateraddress spaceaddress-space usage. This design would substantially benefit some networks (see <xreftarget="benefits"/>),target="benefits"/>) in which the additional cost of an additional service(DHCPv6 prefix delegation)(such as DHCPv6 Prefix Delegation) andallocatingallocation of a larger amount of address space can easily be justified. Examples of such networks include but are not limitedto: </t>to:</t> <ul><li> Large-scale<li>Large-scale networks where even3-5three to five addresses per client might introduce scalabilityissues. </li> <li> Networksissues.</li> <li>Networks with a high number of virtual hosts, so physical devices require multipleaddresses. </li> <li> Managedaddresses.</li> <li>Managed networks where extensive troubleshooting, device traffic logging, or forensics might berequired. </li>required.</li> </ul><t> In<t>In smaller networks, such as home networks or smallenterprises,enterprises with smaller address space and a lower number of clients, SLAAC is a simpler and often preferredoption. </t>option.</t> </section> <section> <name>Routing and Addressing Considerations</name> <section> <name>Prefix Pool Allocation</name> <t>One simple deployment model is to assign a dedicated prefix pool to each link. The prefixes from each link's pool are only issued to requesting clients on thelink, andlink; if clients move to anotherlinklink, they will obtain a prefix from the pool associated with the new link (see <xref target="mobility"/>).</t> <t>This is very similar to how address pools are allocated when using DHCP to assign individual addresses (e.g., DHCPv4 or DHCPv6 IA_NA), where each link has a dedicated pool of addresses, and clients on the link obtain addresses from the pool. In this model, the network can route the entire pool to the link's first-hop routers, and the routers do not need to advertise individual delegated prefixes into the network's dynamic routing protocol.</t> <t>Other deployment models, such as prefix pools shared over multiple links or routers, arepossible,possible but are not described in this document.</t> </section> <section><name> First-Hop<name>First-Hop RouterRequirements </name> <t> InRequirements</name> <t>In large networks, DHCPv6 servers are usuallycentralized,centralized and reached via DHCPv6 relays co-located with the first-hop routers. To delegate IPv6 prefixes to clients, the first hop routers need to implement DHCPv6 relay functions and meet the requirements defined in <xref target="RFC8987"/>. In particular, per <xref target="RFC8987" sectionFormat="of" section="4.2"/>, the first-hop router must maintain a local routing table that contains all prefixes delegated to clients.</t> <t>With the first-hop routers performing DHCPv6 relay functions, the proposed design neither requires any subsequent relays in the path nor introduces any requirements (e.g., snooping)tofor such subsequent relays, if they aredeployed. </t> <t> Todeployed.</t> <t>To ensure that routes to the delegated prefixes are preserved even if a relay is rebooted or replaced, the operatorMUST<bcp14>MUST</bcp14> ensure that all relays in the network infrastructure support DHCPv6 Bulk Leasequery as defined in <xref target="RFC5460"/>. WhileSection 4.3 of<xreftarget="RFC8987"/>target="RFC8987" section="4.3" sectionFormat="of"/> lists keeping active prefix delegations in persistent storage as an alternative to DHCPv6 Bulk Leasequery, relying on persistent storage has the following drawbacks: </t> <ul><li> In<li>In a network with multiple relays, network state can change significantly while the relay is rebooting (new prefixesdelegated,might be delegated or some prefixesexpiringmight be expiring, etc). </li><li> Persistent<li>Persistent storage might not be preserved if the router is physicallyreplaced. </li>replaced.</li> </ul> <t>Another mechanism for first-hop routers to obtain information about delegated prefixes is by using Active Leasequery <xref target="RFC7653"/>, though this is not yet widely supported.</t> </section> <section anchor="mult_relays"><name> Topologies<name>Topologies with Multiple First-HopRouters </name> <t> InRouters</name> <t>In a topology with redundant first-hop routers, all the routers need to relay DHCPv6 traffic, install the delegated prefixes into their routing tables and, if needed, advertise those prefixes to the network.</t> <t>If the first-hop routers obtain information about delegated prefixes by snooping DHCPv6 Reply messages sent by the server, then all the first-hop routers must be able to snoop these messages. This is possible if the client multicasts the DHCPv6 messages it sends to the server. The server will receive one copy of the client message through each first-hop relay, and will reply unicast to each of them via the relay (or chain of relays) from which it received the message. Thus, all first-hop relays will be able to snoop the replies. Per <xref target="RFC8415" sectionFormat="of" section="14"/>, clients always use multicast unless the server uses the Server Unicast option to explicitly allow unicast communication (<xref target="RFC8415" section="21.12" sectionFormat="comma"/>). Therefore, in topologies with multiple first-hop routers, the DHCPv6 serversMUST<bcp14>MUST</bcp14> be configured not to use the Server Unicast option. It should be noted that <xref target="I-D.ietf-dhc-rfc8415bis"/> deprecates the Server Unicast option precisely because it is not compatible with topologies with multiple first-hoprelays. </t>relays.</t> <t>To recover from crashes or reboots, relays can use Bulk Leasequery or Active Leasequery to issue a QUERY_BY_RELAY_ID with the ID(s) of the other relay(s), as configured by the operator. Additionally, some vendors provide vendor-specific mechanisms to synchronize state between DHCP relays.</t> </section> <section><name>On-link<name>On-Link Communication </name><t> For<t>For security reasons, some networks block on-link device-to-device traffic atlayerLayer 2 to prevent communication between clients on the same link. In this case, delegating a prefix to each client doesn't affect traffic flows, as all traffic is sent to the first-hop router anyway. Depending on the network security policy, the router may allow or drop thetraffic. </t> <t> Iftraffic.</t> <t>If the network does allow peer-to-peer communication, the PIO for the on-link prefix is usually advertised with the L-bit set to 1 <xref target="RFC4861"/>. As a result, all addresses from that prefix are considered on-link, and traffic to those destinations is sent directly (not via routers). If such a network delegates prefixes to clients (as described in this document), then each client will consider other client's destination addresses to be off-link, because those addresses are from the delegated prefixes and are no longer within the on-link prefix. When a client sends traffic to another client, packets will initially be sent to the default router. The router will respond with an ICMPv6 redirect message(Section 4.5 of <xref(<xref target="RFC4861"/>).section="4.5" sectionFormat="of"/>). If the client receives and accepts the redirect, then traffic can flow directly from device to device.ThereforeTherefore, the administrator deploying the solution described in this documentSHOULD<bcp14>SHOULD</bcp14> ensure that the first-hop routers can send ICMPv6 redirects (the routers are configured to do so and the security policies permit thosemessages). </t>messages).</t> </section> </section> <section> <name>DHCPv6-PD Server Considerations</name><t> This<t>This document does not introduce any changes to the DHCPv6 protocol itself. However, for the proposed solution to work correctly, the DHCPv6-PD server needs to be configured asfollows: </t>follows:</t> <ul><li> The<li>The serverMUST<bcp14>MUST</bcp14> follow recommendations from <xref target="RFC8168"/>recommendationson processing prefix-lengthhints. </li> <li> Thehints.</li> <li>The serverMUST<bcp14>MUST</bcp14> provide a prefix short enough for the client to extend the network to at least oneinterface,interface and allow nodes on that interface to obtain addresses via SLAAC. The serverMAY<bcp14>MAY</bcp14> provide more address space to clients that ask for it, either by delegating multiple such prefixes, or by delegating a single prefix of a shorter length. It should be noted that <xref target="RFC8168"/> allows the server to provide a prefix shorter than the prefix-length hint value received from theclient. </li> <li> Ifclient.</li> <li>If the server receives the sameSOLICITSolicit message from the same client multiple times through multiple relays, itMUST<bcp14>MUST</bcp14> reply to all of them with the same prefix(es). This ensures that all the relays will correctly configure routes to the delegatedprefixes. </li> <li> Theprefixes.</li> <li>The serverMUST NOT<bcp14>MUST NOT</bcp14> send the Server Unicast option to the client unless the network topology guarantees that no client is connected to a link with multiple relays (see <xreftarget="mult_relays"/>). </li> <li> Intarget="mult_relays"/>).</li> <li>In order to ensure uninterrupted connectivity when a first-hop router crashes or reboots, the serverMUST<bcp14>MUST</bcp14> support Bulk Leasequery or ActiveLeasequery. </li>Leasequery.</li> </ul><t> As<t>As most operators have some experience with IPv4, they can use a similar approach for choosing the pool size and the timers (such asT1/T2T1 and T2 timers). Inparticularparticular, the following factorsshallshould be taken intoaccount: </t>account:</t> <ul><li> the<li>the expected maximum number ofclients; </li> <li>clients;</li> <li>the average duration ofaclientconnection; </li> <li> howconnections;</li> <li>how mobile the clients are (a network where all clients are connected to a single wired VLAN might choose longer timers than a network where clients can switch between multiple wirelessSSIDs); </li> <li> expected level of recurringnetworks);</li> <li>how often clients are expected to reconnect to the network (for example, a corporate authenticated Wi-Fi network might be using longer timers than an open publicWi-Fi). </li>Wi-Fi).</li> </ul><t> DHCPv6<t>DHCPv6 servers that delegate prefixes can interface with Dynamic DNS infrastructure to automatically populate reverseDNS,DNS using wildcard records, similarly to what is described insection 2.5.2 of RFC<xreftarget="RFC8501"/>.target="RFC8501" sectionFormat="of" section="2.2"/>. Networks that also wish to populate forward DNS cannot do so automatically based only on DHCPv6 prefix delegation transactions, but they can do so in other ways, such as by supporting DHCPv6 address registration as described in <xreftarget="I-D.ietf-dhc-addr-notification"/>. </t>target="I-D.ietf-dhc-addr-notification"/>.</t> <t>Some additional recommendations driven by security and privacy considerations are discussed in <xref target="Security"/> and <xref target="privacy"/>.</t> </section> <section> <name>Prefix Length Considerations</name> <t>Delegating a prefix of sufficient size to use SLAAC allows the client to extend the network, providing limitless addresses to IPv6 nodes connected to it (e.g., virtualmachines,machines or tethered devices), because all IPv6 hosts are required to support SLAAC <xref target="RFC8504"/>. Additionally, even clients that support other forms of address assignment require SLAAC for some functions, such as forming dedicated addresses for the use of 464XLAT (seeSection 6.3 of<xreftarget="RFC6877"/>).</t>target="RFC6877" section="6.3" sectionFormat="of"/>).</t> <t>At the time ofwritingwriting, the only prefix size that will allow devices to use SLAAC is 64 bits. Also, as noted in <xref target="RFC7421"/>, using anIIDinterface identifier (IID) shorter than 64 bits and a subnet prefix longer than 64 bits is outside the current IPv6 specifications. Choosing longer prefixes would require the client and any connected system to use other address assignment mechanisms. This would limit the applicability of the proposed solution, as other mechanisms are not currently supported by manyhosts. </t>hosts.</t> <t>For the same reasons, a prefix length of /64 or shorter is required to extend the networkusingas described in <xref target="RFC7084"/> (see requirement L-2), and a prefix length of /64 is required to provide global connectivity for stub networks as per <xref target="I-D.ietf-snac-simple"/>.</t><t> Assigning<t>Assigning a prefix of sufficient size to support SLAAC is possible on large networks. In general, any network that numbers clients from an IPv4 prefix of length X (e.g., X=/18,X=/24),X=/24) would require an IPv6 prefix of length X+32 (e.g., X=/40, X=/56) to provide a /64 prefix to every device. As an example, <xref target="RFC7934" section="9.2"/> suggests that even a very large network that assigns every single one of the 16 million IPv4 addresses in 10.0.0.0/8 would only need an IPv6 /40. A /40 prefix is a small amount of address space: there are 32 times more /40s in the current IPv6 unicast range 2000::/3 than there are IPv4 addresses. Existing sites that currently use a /48 prefix cannot support more than 64k clients in this model without renumbering, though many networks of such size haveLIRLocal Internet Registry (LIR) status and can justify bigger addressblocks. </t>blocks.</t> <t>Note that assigning a prefix of sufficient size to support SLAAC does not require that subtended nodes use SLAAC; they can use other address assignment mechanisms as well.</t> </section> <section anchor="mobility"> <name>Client Mobility</name><t> As<t>As perSection 18.2.12 of<xreftarget="RFC8415"/>,target="RFC8415" section="18.2.12" sectionFormat="of"/>, when the client moves to a new link, itMUST<bcp14>MUST</bcp14> initiate a Rebind/Reply message exchange.ThereforeTherefore, when the client moves between network attachmentpointspoints, it would refresh its delegated prefix the same way it refreshes addresses assigned (via SLAAC or DHCPv6 IA_NA) from a shared on-linkprefix: </t>prefix:</t> <ul> <li> When a client moves from between different attachment points on the same link (e.g., roams between two APs while connected to the sameSSIDwireless network or moves between two switchports belonging to the same VLAN), the delegated prefix does not change, and the first-hop routers have a route for the prefix with the nexthop set to the client link-local address on that link. As per requirement S-2(Section 4.3 ofin <xreftarget="RFC8987"/>),target="RFC8987" section="4.3" sectionFormat="of"/>, the DHCPv6-relays (the first-hop routers)MUST<bcp14>MUST</bcp14> retain the route for the delegating prefix until the route is released or removed due to expiring DHCP timers. Therefore, if the client reconnects to the same link, the prefix doesn'tchange. </li> <li> Whenchange.</li> <li>When a client moves to a different link, the DHCPv6 server provides the client with a new prefix, so thebehaviourbehavior is consistent with SLAAC or DHCPv6-assigned addresses, which are also different on the newlink. </li>link.</li> </ul><t> In<t>In theory, DHCPv6 servers can delegate the same prefix to the same client even if the client changes the attachment points. However, while allowing the client to keep the same prefix while roaming between links might provide some benefits for the client, it is not feasible without changing DHCPv6 relaybehaviour:behavior: after the client moves to a new link, the DHCPv6 relays would retain the route pointing to the client's link-local address on the old link for the duration of DHCPv6 timers (see requirement S-2,Section 4.3 of<xreftarget="RFC8987"/>).target="RFC8987" section="4.3" sectionFormat="of"/>). As a result, the first-hop routers would have two routes for the same prefix pointing to different links, causing connectivity issues for theclient. </t>client.</t> <t>It should be noted that addressing clients from a shared on-link prefix also does not allow clients to keep addresses while roaming between links, so the proposed solution is not different in that regard. In addition to that,quite oftendifferent links often have different security policies applied (for example, corporate internalnetwork vsnetworks versus guestnetwork),networks), hence clients on different links need to use differentprefixes. </t>prefixes.</t> </section> <section anchor="savi"> <name>Antispoofing and SAVI Interaction</name><t> Enabling the<t>Enabling unicast Reverse Path Forwarding(uRPF,(uRPF) <xreftarget="RFC3704"/>)target="RFC3704"/> on the first-hop router interfaces towards clients provides the first layer of defense against spoofing. A spoofed packet sent by a malicious client would be dropped by the router unless the spoofed address belongs to a prefix delegated to another client on the same interface. Therefore the malicious client can only spoof addresses already delegated to another client on the same link or anotherclientclient's link-localaddress. </t> <t> Sourceaddress.</t> <t>Source Address Validation Improvement(SAVI,(SAVI) <xreftarget="RFC7039"/>)target="RFC7039"/> provides more reliable protection against address spoofing. Administrators deploying the proposed solution on SAVI-enabled infrastructureSHOULD<bcp14>SHOULD</bcp14> ensure that SAVI perimeter devices support DHCPv6-PD snooping to create the correct binding for the delegated prefixes (see <xref target="RFC7513"/>). Using FCFS SAVI(<xref target="RFC6620"/>) for protecting<xref target="RFC6620"/> to protect link-local addresses andcreatingcreate SAVI bindings for DHCPv6-PD assigned prefixes would preventspoofing. </t> <t> Somespoofing.</t> <t>Some infrastructure devices do not implement SAVI as defined in <xreftarget="RFC7039"/> buttarget="RFC7039"/>; instead, they perform other forms of address tracking and snooping for security or performance improvement purposes (e.g., ND proxy). This is very commonbehaviourbehavior for wireless devices(access(such as access points and controllers). AdministratorsSHOULD<bcp14>SHOULD</bcp14> ensure that such devices are able to snoop DHCPv6-PDpackets,packets so the traffic from the delegated prefixes is notdropped. </t> <t> Itdropped.</t> <t>It should be noted that using DHCPv6-PD makes it harder for an attacker to select the spoofed source address. When all clients are using the same shared link to form addresses, the attacker might learn addresses used by other clients by listening to multicast Neighbor Solicitations andNeighbourNeighbor Advertisements. In DHCPv6-PD environments, however, the attacker can only learn about otherclients'sclients' global addresses by listening to multicast DHCPv6 messages, which are not transmitted so often, and may not be received by the client at all because they are sent to multicast groups that are specific to DHCPv6 servers andrelays. </t>relays.</t> </section> <section> <name>Migration Strategies and Co-existence with SLAAC Using PrefixesFromfrom the PIO</name><t> It<t>It would be beneficial for the network to explicitly indicate its support of DHCPv6-PD for connectedclients. </t>clients.</t> <ul><li> In<li>In small networks (e.g., home networks), where the number of clients is not too high, the number of available prefixes becomes a limiting factor. If every phone or laptop in a home network were to request a unique prefix suitable for SLAAC, the home network might run out of prefixes, if the prefix allocated to theCPECustomer Premises Equipment (CPE) by its ISP is toosmall (e.g.,long. For example, if an ISP delegates a /60,itthe CPE would only be able to delegate15fifteen /64 prefixes toclients).clients. So while the enterprise network administrator might want all phones in the network to request a prefix, it would be highly undesirable for the same phone to request a prefix when connecting to a homenetwork. </li> <li> Whennetwork.</li> <li>When the network supports both a unique prefix per client and a PIO with A=1 as address assignment methods, it's highly desirable for the client NOT to use the PIO prefix to form global addresses and instead only use the prefix delegated via DHCPv6-PD. Starting both SLAAC using the PIO prefix andDHCPv6-PDDHCPv6-PD, and then deprecatingthatthe SLAAC addresses after receiving a delegated prefix would be very disruptive for applications. If the client continues to use addresses formed from the PIOprefixprefix, it would not only undermine the benefits of the proposed solution (see <xref target="benefits"/>), but it would also introduce complexity and unpredictability in the source address selection. Therefore, the client needs to know what address assignment method to use and whether or not to use the prefix inPIO or not,the PIO, if the network provides the PIO withAthe 'A' flagset. </li>set.</li> </ul><t> The<t>The deployment model described in this document does not require the network to signal support of DHCPv6-PD: for example, devices acting as<xref target="RFC7084"/>compatible routers <xref target="RFC7084"/> will be able to receive prefixes via DHCPv6-PD even without suchsignalling.signaling. Also, some clients may decide to startDHCPv6-PD,DHCPv6-PD and acquireprefixes,prefixes if they detect that the network does not provide addresses via SLAAC. To fully achieve the benefits described in this section, <xreftarget="I-D.collink-6man-pio-pflag"/>target="I-D.ietf-6man-pio-pflag"/> defines a new PIO flag to signal that DHCPv6-PD is the preferred method of obtainingprefixes. </t>prefixes.</t> </section> <section anchor="benefits"> <name>Benefits</name><t> The<t>The proposed solution provides the followingbenefits: </t>benefits:</t> <ul><li> Network<li>Network device resources (e.g., memory) need to scale to the number of devices, not the number of IPv6 addresses. The first-hop routers have a single route per device pointing to the device's link-local address. This can potentially enable hardware costsavings,savings; forexampleexample, if hardware such as wireless LAN controllers is limited to supporting only a specific number of client addresses, or in VXLAN deployments where each client address consumes one routing tableentry. </li> <li> Theentry.</li> <li>The cost of having multiple addresses is offloaded to the clients. Hosts are free to create and use as many addresses as they need without imposing any additional costs onto thenetwork. </li> <li> Ifnetwork.</li> <li>If all clients connected to the given link support this mode of operation and can generate addresses from the delegated prefixes, there is no reason to advertise a common prefix assigned to that link in the PIO with the 'A' flag set.ThereforeTherefore, it is possible to remove the global shared prefix from that link and the router interface completely, so no global addresses are on-link for the link. This would lead to reducing the attack surface for Neighbor Discovery attacks described in <xreftarget="RFC6583"/>. </li> <li> DHCPv6-PDtarget="RFC6583"/>.</li> <li>DHCPv6-PD logs andfirst-hop routersrouting tables obtained from first-hop routers provide complete information on IPv6 to MAC mapping, which can be used for forensics and troubleshooting. Such information is much less dynamic than the NDcache and thereforecache; therefore, it's much easier for an operator to collect and processit. </li> <li> Ait.</li> <li>A dedicated prefix per client allows the network administrator to createper-devicesecurity policies(ACLs)per device (such as ACLs) even if the client is using temporary addresses. This mitigates one of the issues described in <xreftarget="I-D.ietf-opsec-ipv6-addressing"/>. </li> <li> Fatetarget="I-D.ietf-opsec-ipv6-addressing"/>.</li> <li>Fate sharing: all global addresses used by a given client are routed as a single prefix. Either all of them work ornot,none of them work, which makes failures easier to diagnose andmitigate. </li> <li> Lowermitigate.</li> <li>Lower level of multicast traffic: less Neighbor Discovery(<xref target="RFC4861"/>)<xref target="RFC4861"/> multicast packets, asthere are only clients link-local addressesthe routers need toresolve.resolve only the clients' link-local addresses. Also, there is noDADDuplicate Address Detection (DAD) traffic except for the clients' link-localaddresses. </li> <li> Abilityaddresses.</li> <li>Ability to extend the network transparently. If the network delegates to the client a prefix of sufficient size to support SLAAC, the client cantoprovide connectivity to other hosts, as is possible in IPv4 with NAT (e.g., by acting as an IPv6CECustomer Edge (CE) router as described in <xreftarget="RFC7084"/>). </li>target="RFC7084"/>).</li> </ul> </section> <section anchor="privacy"> <name>Privacy Considerations</name><t> If<t>If an eavesdropper or information collector is aware that a given client is using the proposed mechanism, then they may be able to track the client based on its prefix. The privacy implications of this are equivalent to the privacy implications of networks using stateful DHCPv6 address assignment: in both cases, the IPv6 addresses are determined by the server, either because the server assigns a full 128-bit address in a shared prefix, or because the server determines what prefix is delegated to the client. Administrators deploying the proposed mechanism can use similar methods to mitigate the impact as the ones used today in networks that use stateful DHCPv6 address assignment.</t> <t>Except for networks (such as datacenter networks) where hosts do not need temporary addresses(<xref target="RFC8981"/>),<xref target="RFC8981"/>, the networkSHOULD:</t><bcp14>SHOULD</bcp14>:</t> <ul> <li>Ensure that when a client requests a prefix, the prefix is randomly assigned and not allocated deterministically.</li> <li>Use short prefix lifetimes (e.g.,hours),hours) to ensure that when a client disconnects and reconnects it gets a different prefix.</li> <li>Allow the client to have more than one prefix at the same time. This allows the client to rotate prefixes using a mechanism similar to temporary addresses, but that operates on prefixes instead of on individual addresses. In thiscasecase, the prefix's lifetimeMUST<bcp14>MUST</bcp14> be short enough to allow the client to use a reasonable rotation interval without using too much address space. For example, if every 24 hours thetheclient asks for a new prefix and stops renewing the old prefix, and the Valid Lifetime of delegated prefixes is one hour, then the client will consume two prefixes for one hour out of 24 hours, and thus willon averageconsume just under 1.05prefixes.</li>prefixes on average.</li> </ul> </section> <section anchor="IANA"><!-- All drafts are required to have an IANA considerations section. See RFC 8126 for a guide.--><name>IANA Considerations</name> <t>Thismemo includesdocument has norequest to IANA.</t>IANA actions.</t> </section> <section anchor="Security"><!-- All drafts are required to have a security considerations section. See RFC 3552 for a guide. --><name>Security Considerations</name><t> A<t>A malicious (or just misbehaving) client might attempt to exhaust the DHCPv6-PD pool by sending a large number of requests with differingDUIDs.DHCP Unique Identifiers (DUIDs). To prevent a misbehaving client from denying service to other clients, the DHCPv6 server or relayMUST<bcp14>MUST</bcp14> support limiting the number of prefixes delegated to a given client at any given time.</t> <t>Networks can protect against malicious clients by authenticating devices using tokens that cannot be spoofed (e.g., 802.1x authentication) and limiting the number of link-local addresses or MAC addresses that each client is allowed to use. Note that this is not a new issue, as the same attack might be implemented using DHCPv4 or DHCPv6 IA_NA requests; in particular, while it is unlikely for clients to be able to exhaust an IA_NA address pool, clients using IA_NA can exhaust other resources such as DHCPv6 and routing infrastructure resources such as server RAM, ND cache entries,TCAMTernary Content-Addressable Memory (TCAM) entries, SAVI entries,etc. </t> <t> Aetc.</t> <t>A malicious client might request a prefix and then release it very quickly, causing routing convergence events on the relays. The impact of this attack can be reduced if the network rate-limits the amount of broadcast and multicast messages from theclient. </t> <t> Delegatingclient.</t> <t>Delegating the same prefix for the same client introduces privacy concerns. The proposed mitigation is discussed in <xreftarget="privacy"/>. </t> <t> Spoofingtarget="privacy"/>.</t> <t>Spoofing scenarios and prevention mechanisms are discussed in <xreftarget="savi"/>. </t>target="savi"/>.</t> </section><!-- NOTE: The Acknowledgements and Contributors sections are at the end of this template --></middle> <back> <displayreference target="I-D.ietf-6man-pio-pflag" to="PIO-PFLAG"/> <displayreference target="I-D.ietf-dhc-rfc8415bis" to="RFC8415bis"/> <displayreference target="I-D.ietf-dhc-addr-notification" to="ADDR-NOTIFICATION"/> <displayreference target="I-D.ietf-opsec-ipv6-addressing" to="IPv6-ADDRESS"/> <displayreference target="I-D.ietf-snac-simple" to="SNAC-SIMPLE"/> <references> <name>References</name> <references> <name>Normative References</name> <xi:include href="https://www.rfc-editor.org/refs/bibxml/reference.RFC.2119.xml"/> <xi:include href="https://www.rfc-editor.org/refs/bibxml/reference.RFC.4193.xml"/> <xi:include href="https://www.rfc-editor.org/refs/bibxml/reference.RFC.7084.xml"/> <xi:include href="https://www.rfc-editor.org/refs/bibxml/reference.RFC.5460.xml"/> <xi:include href="https://www.rfc-editor.org/refs/bibxml/reference.RFC.6620.xml"/> <xi:include href="https://www.rfc-editor.org/refs/bibxml/reference.RFC.6877.xml"/> <xi:include href="https://www.rfc-editor.org/refs/bibxml/reference.RFC.8168.xml"/> <xi:include href="https://www.rfc-editor.org/refs/bibxml/reference.RFC.8174.xml"/> <xi:include href="https://www.rfc-editor.org/refs/bibxml/reference.RFC.8273.xml"/> <xi:include href="https://www.rfc-editor.org/refs/bibxml/reference.RFC.8415.xml"/> <xi:include href="https://www.rfc-editor.org/refs/bibxml/reference.RFC.8981.xml"/> <xi:include href="https://www.rfc-editor.org/refs/bibxml/reference.RFC.8987.xml"/><!-- The recommended and simplest way to include a well known reference --></references> <references> <name>Informative References</name> <xi:include href="https://www.rfc-editor.org/refs/bibxml/reference.RFC.3704.xml"/> <xi:include href="https://www.rfc-editor.org/refs/bibxml/reference.RFC.4861.xml"/> <xi:include href="https://www.rfc-editor.org/refs/bibxml/reference.RFC.4862.xml"/> <xi:include href="https://www.rfc-editor.org/refs/bibxml/reference.RFC.6459.xml"/> <xi:include href="https://www.rfc-editor.org/refs/bibxml/reference.RFC.6583.xml"/> <xi:include href="https://www.rfc-editor.org/refs/bibxml/reference.RFC.7039.xml"/> <xi:include href="https://www.rfc-editor.org/refs/bibxml/reference.RFC.7278.xml"/> <xi:include href="https://www.rfc-editor.org/refs/bibxml/reference.RFC.7348.xml"/> <xi:include href="https://www.rfc-editor.org/refs/bibxml/reference.RFC.7421.xml"/> <xi:include href="https://www.rfc-editor.org/refs/bibxml/reference.RFC.7513.xml"/> <xi:include href="https://www.rfc-editor.org/refs/bibxml/reference.RFC.7653.xml"/> <xi:include href="https://www.rfc-editor.org/refs/bibxml/reference.RFC.7934.xml"/> <xi:include href="https://www.rfc-editor.org/refs/bibxml/reference.RFC.8200.xml"/> <xi:include href="https://www.rfc-editor.org/refs/bibxml/reference.RFC.8501.xml"/> <xi:include href="https://www.rfc-editor.org/refs/bibxml/reference.RFC.8504.xml"/> <!-- [I-D.ietf-6man-pio-pflag] IESG state: I-D Exists as of 07/09/24 --> <xi:includehref="https://bib.ietf.org/public/rfc/bibxml3/reference.I-D.collink-6man-pio-pflag.xml"/>href="https://bib.ietf.org/public/rfc/bibxml3/reference.I-D.ietf-6man-pio-pflag.xml"/> <!-- [I-D.ietf-dhc-rfc8415bis] IESG state: I-D Exists as of 07/09/24 --> <xi:include href="https://bib.ietf.org/public/rfc/bibxml3/reference.I-D.ietf-dhc-rfc8415bis.xml"/><xi:include href="https://bib.ietf.org/public/rfc/bibxml3/reference.I-D.ietf-dhc-addr-notification"/><!-- [I-D.ietf-dhc-addr-notification] IESG state: In RFC Ed Queue as of 07/09/24 --> <reference anchor="I-D.ietf-dhc-addr-notification" target="https://datatracker.ietf.org/doc/html/draft-ietf-dhc-addr-notification-13"> <front> <title>Registering Self-generated IPv6 Addresses using DHCPv6</title> <author fullname="Warren Kumari" initials="W." surname="Kumari"> <organization>Google, LLC</organization> </author> <author fullname="Suresh Krishnan" initials="S." surname="Krishnan"> <organization>Cisco Systems, Inc.</organization> </author> <author fullname="Rajiv Asati" initials="R." surname="Asati"> <organization>Independent</organization> </author> <author fullname="Lorenzo Colitti" initials="L." surname="Colitti"> <organization>Google, LLC</organization> </author> <author fullname="Jen Linkova" initials="J." surname="Linkova"> <organization>Google, LLC</organization> </author> <author fullname="Sheng Jiang" initials="S." surname="Jiang"> <organization>Beijing University of Posts and Telecommunications</organization> </author> <date day="16" month="May" year="2024"/> </front> <seriesInfo name="Internet-Draft" value="draft-ietf-dhc-addr-notification-13"/> </reference> <!-- [I-D.ietf-opsec-ipv6-addressing] IESG state: Expired as of 04/08/24 --> <xi:include href="https://bib.ietf.org/public/rfc/bibxml3/reference.I-D.ietf-opsec-ipv6-addressing.xml"/> <!-- [I-D.ietf-snac-simple] IESG state: I-D Exists as of 06/20/24 --> <xi:include href="https://bib.ietf.org/public/rfc/bibxml3/reference.I-D.ietf-snac-simple.xml"/> </references> </references> <sectionanchor="appendix" title="Appendix: Multipleanchor="appendix"> <name>Multiple AddressesConsiderations">Considerations</name> <t>While a typical IPv4 host normally has only one IPv4 address per interface, an IPv6 device almost always has multiple addresses assigned to its interface. At the very least, a host can be expected to have one link-local address, one temporary address, and, in most cases, one stable global address. On a network providing NAT64 service, an IPv6-only host running the 464XLAT customer-side translator(CLAT,(CLAT) <xreftarget="RFC6877"/>)target="RFC6877"/> would use a dedicated 464XLAT address, configured via SLAAC (seeSection 6.3 of<xreftarget="RFC6877"/>),target="RFC6877" section="6.3" sectionFormat="of"/>), which brings the total number of addresses to4.four. Other common scenarios where the number of addresses per host interface might increasesignificantly,significantly include but are not limitedto: </t>to:</t> <ul><li> Devices<li>Devices runningcontainers/namespaces:containers or namespaces: eachcontainer/namespacecontainer or namespace would have multiple addresses as described above. As a result, a device running just a few containers in a bridge mode can easily have 20 or more IPv6 addresses on the givenlink. </li> <li> Networkslink.</li> <li>Networks assigning multiple prefixes to a given link: multihomed networks, networks usingULAUnique Local IPv6 Unicast Addresses (ULA, <xreftarget="RFC4193"/>target="RFC4193"/>) and non-ULA prefixes together, ornetworknetworks performing a graceful renumbering from one prefix toanother. </li>another.</li> </ul><t> <xref<t><xref target="RFC7934"/> discusses this aspect and explicitly states that IPv6 deploymentsSHOULD NOT<bcp14>SHOULD NOT</bcp14> limit the number of IPv6 addresses a host can have. However, it has beenbeenobserved that networks often do limit the number of on-link addresses per device, likely in an attempt to protect network resources and prevent DoSattacks. </t> <t> Theattacks.</t> <t>The most common scenario of network-imposed limitations isNeighbor Discovery (ND)ND proxy. Many enterprise-scale wireless solutions implement ND proxy to reduce the amount of broadcast and multicast downstream (AP to clients) traffic and provide SAVI functions. To perform ND proxy, a device usually maintains atable,table containing IPv6 and MAC addresses of connected clients. At least some implementations have hardcoded limits on how many IPv6 addresses per single MAC such a table can contain. When the limit isexceededexceeded, thebehaviourbehavior isimplementation-dependent.implementation dependent. Some vendors just fail to install an N+1 address to the table. Others delete the oldest entry for this MAC and replace it with the new address. In any case, the affected addresses lose network connectivity without receiving any implicit signal, with traffic being silentlydropped. </t>dropped.</t> </section> <section anchor="Acknowledgements" numbered="false"><!-- [REPLACE/DELETE] an Acknowledgements section is optional --><name>Acknowledgements</name> <t>Thanks toHarald Alvestrand, Nick Buraglio, Brian Carpenter, Tim Chown, Roman Danyliw, Gert Doering, David Farmer, Fernando Gont, Joel Halpern, Nick Hilliard, Bob Hinden, Martin Hunek, Erik Kline, Warren Kumari, David Lamparter, Andrew McGregor, Tomek Mrugalski, Alexandre Petrescu, Jurgen Schonwalder, Pascal Thubert, Ole Troan, Eric Vyncke, Eduard Vasilenko, Timothy Winters, Chongfeng Xie, Peter Yee<contact fullname="Harald Alvestrand"/>, <contact fullname="Nick Buraglio"/>, <contact fullname="Brian Carpenter"/>, <contact fullname="Tim Chown"/>, <contact fullname="Roman Danyliw"/>, <contact fullname="Gert Doering"/>, <contact fullname="David Farmer"/>, <contact fullname="Fernando Gont"/>, <contact fullname="Joel Halpern"/>, <contact fullname="Nick Hilliard"/>, <contact fullname="Bob Hinden"/>, <contact fullname="Martin Hunek"/>, <contact fullname="Erik Kline"/>, <contact fullname="Warren Kumari"/>, <contact fullname="David Lamparter"/>, <contact fullname="Andrew McGregor"/>, <contact fullname="Tomek Mrugalski"/>, <contact fullname="Alexandre Petrescu"/>, <contact fullname="Jurgen Schonwalder"/>, <contact fullname="Pascal Thubert"/>, <contact fullname="Ole Troan"/>, <contact fullname="Eric Vyncke"/>, <contact fullname="Eduard Vasilenko"/>, <contact fullname="Timothy Winters"/>, <contact fullname="Chongfeng Xie"/>, and <contact fullname="Peter Yee"/> for the discussions, their input, and all contributions.</t></section> <section anchor="Contributors" numbered="false"> <!-- [REPLACE/DELETE] a Contributors section is optional --> <name>Contributors</name></section> </back> </rfc>