Skip to main content

IPv6 Prefix Assignment to end-users
draft-palet-v6ops-prefix-assignment-01

Document Type Active Internet-Draft (candidate for v6ops WG)
Author Jordi Palet Martinez
Last updated 2026-04-07 (Latest revision 2026-03-02)
Replaces draft-palet-v6ops-p2p-from-customer-prefix, draft-palet-v6ops-p2p-links, draft-palet-v6ops-point2point
RFC stream Internet Engineering Task Force (IETF)
Intended RFC status (None)
Formats
Additional resources Mailing list discussion
Stream WG state Call For Adoption By WG Issued
Document shepherd (None)
IESG IESG state I-D Exists
Consensus boilerplate Unknown
Telechat date (None)
Responsible AD (None)
Send notices to (None)
draft-palet-v6ops-prefix-assignment-01
v6ops                                                  J. Palet Martinez
Internet-Draft                                          The IPv6 Company
Intended status: Best Current Practice                      2 March 2026
Expires: 3 September 2026

                  IPv6 Prefix Assignment to end-users
                 draft-palet-v6ops-prefix-assignment-01

Abstract

   This document describes different alternatives and best current
   practices for assignment of IPv6 prefixes for end-user broadband
   networks, including considerations about point-to-point links, their
   size, numbering choices, pool choices, customer prefix assignment
   size and persistance of those assignments.

Status of This Memo

   This Internet-Draft is submitted in full conformance with the
   provisions of BCP 78 and BCP 79.

   Internet-Drafts are working documents of the Internet Engineering
   Task Force (IETF).  Note that other groups may also distribute
   working documents as Internet-Drafts.  The list of current Internet-
   Drafts is at https://datatracker.ietf.org/drafts/current/.

   Internet-Drafts are draft documents valid for a maximum of six months
   and may be updated, replaced, or obsoleted by other documents at any
   time.  It is inappropriate to use Internet-Drafts as reference
   material or to cite them other than as "work in progress."

   This Internet-Draft will expire on 3 September 2026.

Copyright Notice

   Copyright (c) 2026 IETF Trust and the persons identified as the
   document authors.  All rights reserved.

   This document is subject to BCP 78 and the IETF Trust's Legal
   Provisions Relating to IETF Documents (https://trustee.ietf.org/
   license-info) in effect on the date of publication of this document.
   Please review these documents carefully, as they describe your rights
   and restrictions with respect to this document.  Code Components
   extracted from this document must include Revised BSD License text as
   described in Section 4.e of the Trust Legal Provisions and are
   provided without warranty as described in the Revised BSD License.

Palet Martinez          Expires 3 September 2026                [Page 1]
Internet-Draft           IPv6 Prefix Assignment               March 2026

Table of Contents

   1.  Introduction  . . . . . . . . . . . . . . . . . . . . . . . .   3
   2.  Requirements Language . . . . . . . . . . . . . . . . . . . .   3
   3.  Point-to-Point Links Considerations . . . . . . . . . . . . .   4
     3.1.  The Ping-Pong Problem in Point-to-Point Links . . . . . .   4
     3.2.  Prefix Size Choices . . . . . . . . . . . . . . . . . . .   4
       3.2.1.  Rationale for using /64 . . . . . . . . . . . . . . .   4
       3.2.2.  Rationale for using /127  . . . . . . . . . . . . . .   5
       3.2.3.  Rationale for using /126 and Other Options  . . . . .   6
       3.2.4.  A Possible Middle-Term Choice . . . . . . . . . . . .   6
     3.3.  Numbering Choices . . . . . . . . . . . . . . . . . . . .   6
       3.3.1.  GUA (Global Unicast Addresses)  . . . . . . . . . . .   6
       3.3.2.  ULA (Unique Local Addresses)  . . . . . . . . . . . .   7
       3.3.3.  Link-Local Addresses Only . . . . . . . . . . . . . .   7
     3.4.  Prefix Pool Considerations  . . . . . . . . . . . . . . .   8
       3.4.1.  /64 GUA from Dedicated Pool for point-to-point
               links . . . . . . . . . . . . . . . . . . . . . . . .   9
       3.4.2.  /64 GUA from Customer Prefix for point-to-point
               links . . . . . . . . . . . . . . . . . . . . . . . .   9
         3.4.2.1.  Routing Aggregation of the Point-to-Point
                 Links . . . . . . . . . . . . . . . . . . . . . . .   9
         3.4.2.2.  DHCPv6 Considerations . . . . . . . . . . . . . .  11
         3.4.2.3.  Router Considerations . . . . . . . . . . . . . .  11
       3.4.3.  Other Choices . . . . . . . . . . . . . . . . . . . .  11
       3.4.4.  Numbering Interfaces  . . . . . . . . . . . . . . . .  11
   4.  Prefix Assignment to End-Users  . . . . . . . . . . . . . . .  12
     4.1.  /48 for every end-user  . . . . . . . . . . . . . . . . .  13
     4.2.  /48 for business customers and /52 or /56 for residential
           customers . . . . . . . . . . . . . . . . . . . . . . . .  14
     4.3.  Prefixes longer than /56  . . . . . . . . . . . . . . . .  16
     4.4.  Considerations for Cellular Operators . . . . . . . . . .  17
   5.  Assigned Prefix Persistence Considerations  . . . . . . . . .  18
     5.1.  Non-persistent assignments perceived as 'easier'  . . . .  19
     5.2.  Non-persistent assignments considered harmful . . . . . .  20
     5.3.  Persistent assignments are the best current practice  . .  23
   6.  Best Common Practices Summary for IPv6 Prefix Assignments . .  24
   7.  Security Considerations . . . . . . . . . . . . . . . . . . .  25
   8.  IANA Considerations . . . . . . . . . . . . . . . . . . . . .  25
   9.  Acknowledgements  . . . . . . . . . . . . . . . . . . . . . .  25
   10. References  . . . . . . . . . . . . . . . . . . . . . . . . .  25
     10.1.  Normative References . . . . . . . . . . . . . . . . . .  25
     10.2.  Informative References . . . . . . . . . . . . . . . . .  28
   Author's Address  . . . . . . . . . . . . . . . . . . . . . . . .  29

Palet Martinez          Expires 3 September 2026                [Page 2]
Internet-Draft           IPv6 Prefix Assignment               March 2026

1.  Introduction

   When deploying IPv6 in broadband networks to end-users (both
   residential and enterprise), there are different alternatives to
   tackle the most common open questions:

   *  What size for the point-to-point links?

   *  How to number the point-to-point links?

   *  What pool to use for the point-to-point links?

   *  What prefix size should be assigned to customers?

   *  Should the customers prefixes be persistent?

   From an operational perspective, each choice may have different
   advantages or disadvantages that need to be taken in consideration
   under the scope of each specific network architecture design, as
   already described in the RIPE BCOP 690 document [RIPE-690] that has
   been extensively used across many years.

   For example, [RFC6164] describes using /127 prefixes for inter-router
   point-to-point links, using two different address pools, one for
   numbering the point-to-point links and another one for delegating the
   prefixes at the end of the point-to-point link.  However, this
   doesn't exclude other choices.

   The proposed approaches are suitable for those point-to-point links
   connecting ISP to customers, but not limited to those cases, and in
   fact, all them are being used by a relevant number of networks
   worldwide, in several different scenarios (service providers,
   enterprise networks, etc.).

   This document describes the best current practices for each of those
   different aspects.

   Note that across this document, we can use interchangeably end-user,
   customer, subscriber or end-site, as what it actually matters is what
   is connected behind an individual link.

2.  Requirements Language

   The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
   "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and
   "OPTIONAL" in this document are to be interpreted as described in BCP
   14 [RFC2119] [RFC8174] when, and only when, they appear in all
   capitals, as shown here.

Palet Martinez          Expires 3 September 2026                [Page 3]
Internet-Draft           IPv6 Prefix Assignment               March 2026

3.  Point-to-Point Links Considerations

3.1.  The Ping-Pong Problem in Point-to-Point Links

   Some point-to-point links may present the ping-pong problem (a
   forwarding loop).  The fundamental root cause of this problem is an
   IPv6 implementation not performing full Neighbor Discovery (NS/NA) on
   addresses that the prefix says could exist on the link.

   IPv6 implementations are assuming that all addresses within the
   prefix must exist at the other end of the point-to-point link, and
   send the traffic straight onto the link.  If the address doesn't
   exist, and there is a covering route back in the other direction, the
   ping-pong problem occurs.

   Full Neighbor Discovery is doing more than just resolving the link-
   layer address of an IPv6 address.  Neighbor Discovery is also
   determining if the address exists.  Even if a point-to-point link
   doesn't have link-layer addresses to resolve, ND determining if an
   address exists on the link is very beneficial because it will prevent
   the ping-pong problem occurring entirely regardless of the IPv6
   prefix length being used on the link.

3.2.  Prefix Size Choices

   [RFC7608] already discusses about the IPv6 prefix length
   recommendations for forwarding, and the need for routing and
   forwarding implementations to ensure that longest-prefix-match works
   on any prefix length up to /128.  So, in this document, we
   concentrate in the most commonly used choices, not excluding other
   options.

3.2.1.  Rationale for using /64

   The IPv6 Addressing Architecture ([RFC4291]) specifies that all the
   Interface Identifiers for all the unicast addresses (except for
   000/3) are required to be 64 bits long and to be constructed in
   Modified EUI-64 format.  This document, has been updated by several
   documents, including "Significance of IPv6 Interface Identifiers"
   ([RFC7136]), but in the end, the IID is still required to be 64 bits
   long even if EUI-64 is not being used.

   The same document also mandates the usage of the predefined subnet-
   router anycast address, which has cleared to zero all the bits that
   do not form the subnet prefix.

Palet Martinez          Expires 3 September 2026                [Page 4]
Internet-Draft           IPv6 Prefix Assignment               March 2026

   Using /64 is the most common scenario and currently the best practice
   by the number of service providers using this approach compared to
   others.

   Using a /64 has the advantage of being future proof and avoids
   renumbering, in the event that new standards take advantage of the 64
   bits for other purposes, or the link becomes a point-to-multipoint,
   or there is a need to use more addresses in the link (e.g.,
   monitoring equipment, managed bridges).

   It has been raised also the issue of some hardware having limitations
   in using prefixes longer then /64, for example using extra hardware
   resources.

   Section 5. of [RFC6164] describes possible issues when using /64 for
   the point-to-point links, such as the ping-pong and the neighbor
   cache exhaustion.  However, it also states that they can be mitigated
   by other means, including the latest ICMPv6 [RFC4443] and ND
   [RFC4861].  Indeed, considering the publication date of that
   document, those issues should not be any longer a concern.  The fact
   is that many operators worldwide, today use /64 without any concerns,
   as vendors have taken the necessary code updates.

   Consequently, we shall conclude that it is a valid and recommended
   approach to use /64 prefixes for the point-to-point links and towards
   that, implementations must follow [RFC4861] not just for "regular"
   point-to-point links, but also for point-to-point tunnels and all
   tunnel/virtual link types.

3.2.2.  Rationale for using /127

   [RFC6164] already do a complete review of reasons why /127 is a good
   approach vs other options.  However, it needs to be considered that
   it was published a number of years ago, and hardware today already
   incorporate mitigations.

   It should be noted that, when using a /127 prefix, configuration of
   each of the addresses within the /127 prefix, at each respective end
   of the link, must be actively validated by the network operator.  A
   missing /127 address from one end of the link, with a local route
   pointing out that end of the link that covers the missing /127
   address, such as a default route, causes a "ping-pong" scenario to
   exist for the missing /127 address.  The link could still be
   successfully carrying transit traffic, and IPv6 will not report any
   errors, because IPv6 doesn't require, neither will check, that all
   interfaces attached to a link have addresses from all prefixes
   assigned to the link, excepting the Link-Local prefix per [RFC4291].

Palet Martinez          Expires 3 September 2026                [Page 5]
Internet-Draft           IPv6 Prefix Assignment               March 2026

   It is a valid approach to use /127 for the point-to-point links,
   however is not future proof neither recommended considering the
   comments from the previous section, and older equipment may not
   support it.

3.2.3.  Rationale for using /126 and Other Options

   /126 was considered by [RFC3627], and despite this document has been
   obsoleted, because was considering /127 as harmful, the
   considerations in Section 4.3 are still valid.

   The same document describes options such as /112 and /120, and all
   those are commonly used in worldwide IPv6 deployments [IPv6-Survey],
   though in a lesser degree than /64 or /127.

   Consequently, we shall conclude that even /126, /120 and /112 are
   valid approaches for the point-to-point links, are not recommened.

3.2.4.  A Possible Middle-Term Choice

   A possible "middle-term" approach, will be to allocate a /64 for each
   point-to-point link, but use just one /127 out of it, making it
   future proof and at the same time avoiding possible issues indicated
   in the previous sections.  This will be also consistent with the
   recommendations of [RFC6583] to prevent ND operational implications.

3.3.  Numbering Choices

   IPv6 provides different unicast addressing scopes which can be
   considered when numbering a point-to-point link.

   It has been reported that certain hardware may consume resources when
   using numbered links.  This is a very specific situation that may
   need to be consider on a case by case basis.

3.3.1.  GUA (Global Unicast Addresses)

   Using Globally Reachable Unicast Addresses is the most common
   approach.  It provides full functionality for both end-points of the
   point-to-point link and consequently, facilitates troubleshooting, so
   it is the recommended approach.

Palet Martinez          Expires 3 September 2026                [Page 6]
Internet-Draft           IPv6 Prefix Assignment               March 2026

3.3.2.  ULA (Unique Local Addresses)

   Some networks use ULAs for numbering the point-to-point links.  This
   approach may cause numerous problems when carrying Internet traffic
   and therefore, is strongly discouraged.  For example, if the CE needs
   to send an ICMPv6 message to a host outside that network (to the
   Internet), the packet with ULA source address will not get thru and
   PMTUD will break, which in turn will completely break that IPv6
   connection when the MTU is not the same for all the path.

   ULAs may be considered as IPv6 private addresses, not intended to be
   used as source or destination addresses across the Internet.  This
   issue also exists in IPv4 when using [RFC1918] addresses on links
   carrying IPv4 Internet traffic.  [RFC6752] discusses this issue for
   IPv4, with much of the discussion applying similarly to IPv6 and
   ULAs.

   However, this approach is valid if, following Section 2.2 of
   [RFC4443], and despite using ULA for the point-to-point link, the
   router is configured with at least one GUA and the source of the
   ICMPv6 messages are always a GUA, per the IPv6 Default Address
   Selection algorithm [RFC6724].  Consequencly, should not be
   recommended and might only be used in cases where the CE is provided
   by the service provider, specifically supports this option, and can't
   be replaced by the end-user.

   A final consideration.  Is not unusual to find ULA in traceroutes at
   some deployments; they are confusing and might even give out
   information that is considered a security risk.

3.3.3.  Link-Local Addresses Only

   Some networks leave the point-to-point links with only Link-Local
   Addresses (LLAs) used at both ends of the link.  This is sometimes
   improperly referred as "unnumbered", because the Link-Local addresses
   are also "numbers".  Furthermore, [RFC4291] requires that all
   interfaces attached to a link have at least a Link-Local "number" or
   address from the Link-Local prefix.

   [RFC7404] (Using Only Link-Local Addressing inside an IPv6 Network)
   discusses pros and cons of this alternative, which in general apply
   for the point-to-point links.

Palet Martinez          Expires 3 September 2026                [Page 7]
Internet-Draft           IPv6 Prefix Assignment               March 2026

   While this choice might work if the point-to-point link is terminated
   in a router, which typically will get configured with a suitable
   routable GUA or ULA, it will not work for devices that can't be
   further configured, for example if they do not support DHCPv6-PD
   [RFC8415].  This is the case for hosts, when the Operating System is
   not expected to be a DHCPv6-PD client and are therefore left without
   any usable GUA to allow traffic forwarding.

   In the case of a router, the route for the assigned prefix is pointed
   towards the link-local address on the router WAN port and the default
   route on the router is pointed towards the link-local address on the
   upstream network equipment port.

   This choice seems easier to implement, compared the previous ones,
   but it also brings some drawbacks, such as difficulties with
   troubleshooting and monitoring.  For example, link local addresses do
   not appear in traceroute, so it makes more difficult to locate the
   exact point of failure.

   It is more useful in scenarios where it is known that only a router
   will be attached to the point-to-point link, and where the configured
   address of the router is known.  Non-routers connecting to a network,
   which can't initiate DHCPv6-PD might experience problems and will
   stay unnumbered upon connection, if a /64 prefix is not used to
   number the link.  This may be also the case for routers, which will
   not be able to complete the DHCPv6-PD in unnumbered links.

   As in the case of ULAs, is not unusual to find LL addresses in
   traceroutes at some deployments; they are confusing and might even
   give out information that is considered a security risk.

   The considerations indicated in the previous section, regarding not
   using ULA as source address of ICMPv6 messages, and instead ensuring
   there is at least one GUA configured for that, also apply if link-
   locals are used for the point-to-point link.  So once more, this
   approach is not recommended.

3.4.  Prefix Pool Considerations

   The logic choice seems to use a dedicated pool of IPv6 addresses, as
   this is the way we are "used to" with IPv4.  Actually, this is done
   often by means of different IPv6 pools at every PoP in a service
   provider network.

   However, this is not neccesarily the best choice.  The fact that the
   default IPv6 link size is /64 and commonly multiple /64s are assigned
   to a single customer, provides an interesting alternative approach
   for combining best practices described in the precedent sections.

Palet Martinez          Expires 3 September 2026                [Page 8]
Internet-Draft           IPv6 Prefix Assignment               March 2026

3.4.1.  /64 GUA from Dedicated Pool for point-to-point links

   Using a /64 prefix from a dedicated pool of IPv6 prefixes is very
   common.  A separate block of IPv6 space is allocated for the WAN
   links to the end customer CEs, so that when CE connects to the
   network and performs router discovery, a /64 prefix is used to number
   both ends of the connection.

   In the case of an end-user host (not router) connected using
   technologies such as PPPoE, the setup process is concluded when the
   host is properly IPv6 numbered and can start sending and receiving
   traffic.

   It is more common when a CE is available, as if it is capable of
   issuing a DHCPv6-PD request, the IPv6 prefix delegation process
   starts and an IPv6 prefix is assigned to the CE.

   A possible benefit of using a dedicated IPv6 pool, is that allows
   applying security policies without harming the customers.  This is
   only true if customers always have a CE at their end of the WAN link.

3.4.2.  /64 GUA from Customer Prefix for point-to-point links

   Using a /64 from the customer prefix, in addition to the advantages
   already indicated when using /64, simplifies the addressing plan and
   storage needs for logging the addresses of the end-sites.

   The use of /64 also facilitates an easier way for routing the shorter
   aggregated prefix into the point-to-point link.  Consequently it
   simplifies the "view" of a more unified addressing plan, providing an
   easier path for following up any issue when operating IPv6 networks
   and typically, will have a great impact in saving expensive hardware
   resources (lower usage of TCAM, typically by half).

   This mechanism would not work in broadcast layer two media that rely
   on ND, because it will try ND for all the addresses within the
   shorter prefix that is being routed thru the point-to-point link.

3.4.2.1.  Routing Aggregation of the Point-to-Point Links

   Following this approach and assuming that a shorter prefix is
   typically delegated to a customer, for example a /48, it is possible
   to simplify the routing aggregation of the point-to-point links.
   Towards this, the point-to-point link may be numbered using the first
   /64 of the /48 delegated to the customer.

   Let's see a practical example:

Palet Martinez          Expires 3 September 2026                [Page 9]
Internet-Draft           IPv6 Prefix Assignment               March 2026

   *  A service provider uses the prefix 2001:db8::/32 and is using
      2001:db8:aaaa::/48 for a given customer.

   *  Instead of allocating the point-to-point link from a different
      addressing pool, it may use 2001:db8:aaaa::/64 (which is the first
      /64 subnet from the 2001:db8:aaaa::/48) to number the link.

   *  This means that, in the case the non-EUI-64 approach is used, the
      point-to-point link may be numbered as 2001:db8:aaaa::1/64 for the
      provider side and 2001:db8:aaaa::2/64 for the customer side.

   *  Note that using the first /64 and interface identifiers 1 and 2 is
      a very common practice.  However other values may be chosen
      according to each case specific needs.

   In this way, as the same address pool is being used for both, the
   prefix and the point-to-point link, one of the advantages of this
   approach is to make very easy the recognition of the point-to-point
   link that belongs to a given customer prefix, or in the other way
   around, the recognition of the prefix that is linked by a given
   point-to-point link.

   For example, making a trace-route to debug any issue to a given
   address in the provider network, will show a straight view, and it
   becomes unnecessary one extra step to check a database that correlate
   an address pool for the point-to-point links and the customer
   prefixes, as all they are the same.

   Moreover, it is possible to use the shorter prefix as the provider
   side numbering for the point-to-point link and keep the /64 for the
   customer side.  In our example, it will become:

   *  Point-to-point link at provider side: 2001:db8:aaaa::1/48

   *  Point-to-point link at customer side: 2001:db8:aaaa::2/64

   This provides one additional advantage as in some platforms the
   configuration may be easier saving one step for the route of the
   delegated prefix (no need for two routes to be configured, one for
   the delegated prefix, one for the point-to-point link).  It is
   possible because the longest-prefix-match rule.

   The behavior of this type of configuration has been successfully
   deployed in different operator and enterprise networks, using
   commonly available implementations with different routing protocols,
   including RIP, BGP, IS-IS, OSPF, along static routing, and no
   failures or interoperability issues have been reported.

Palet Martinez          Expires 3 September 2026               [Page 10]
Internet-Draft           IPv6 Prefix Assignment               March 2026

3.4.2.2.  DHCPv6 Considerations

   When using DHCPv6 [RFC8415], the Prefix Exclude Option for
   DHCPv6-based Prefix Delegation ([RFC6603]), allows the usage
   described above.

   Moreover, [RFC3769] has no explicit requirement that avoids the
   approach described in this document.

   Note that if instead of DHCPv6, other configuration means are being
   used, it is also possible to use a /64 from the customer allocated
   prefix for the point-to-point link.

3.4.2.3.  Router Considerations

   This approach is being used by operators in both, residential/SOHO
   and enterprise networks, so the routers at the customer end for those
   networks MUST support [RFC6603] if DHCPv6-PD is used.

   In the case of Customer Edge Routers there is a specific requirement
   ([RFC7084]) WPD-8 (Prefix delegation Requirements), marked as SHOULD
   for [RFC6603].  However, in a scenario where the approach described
   in this document is followed, together with DHCPv6-PD, the CE Router
   MUST support [RFC6603].

3.4.3.  Other Choices

   In PPPoE deployments, IPv6CP agrees on IIDs that are used to generate
   LLAs for the p2p interface.  In that case, as well as in IPoE
   deployments, may then choose to use DHCPv6-NA to get a single /128
   GUA to use in the WAN interface.

   A valid alternative to DHCPv6-NA is using SLAAC to configure the WAN
   interface with a /64 GUA, as previuosly indicated.  In the case of
   IPoE, the route for the Delegated Prefix will use as next-hop the
   LLA.

3.4.4.  Numbering Interfaces

   Often, in point-to-point links, hardware tokens are not available, or
   there is the need to keep certain bits (u, g) cleared, so the links
   can be manually numbered sequentially with most of the bits cleared
   to zero.  This numbering makes as well easier to remember the
   interfaces, which typically will become numbered as 0 (with 63
   leading zero bits) for the provider side and 1 (with 63 leading zero
   bits) for the customer side.

Palet Martinez          Expires 3 September 2026               [Page 11]
Internet-Draft           IPv6 Prefix Assignment               March 2026

   Using interface identifiers as 0 and 1 is not only a very simple
   approach, but also a very common practice.  Other different choices
   can as well be used as required in each case.

   On the other hand, using EUI-64 or pseudo-random IIDs, makes it more
   difficult to remember and handle the interfaces, but provides an
   additional degree of protection against port (actually address)
   scanning as described at [RFC7707].

4.  Prefix Assignment to End-Users

   When deploying IPv6, compared with IPv4, several new aspects need to
   be considered:

   1.   A service provider doesn't have scarcity of IPv6 addresses,
        needs to think big when planning, which in turn will facilitate
        operations and avoid renumbering.

   2.   All the RIR policies permit assignments of a /48 per site and
        allow to obtain the prefix size according to each service
        provider needs, depending on their own justified need based on
        addressing plans.

   3.   To keep addressing plans usable and understandable and to align
        with DNS reverse zone delegations, the size of the delegated
        prefix should align with a nibble boundary.

   4.   IPv6 default prefix lengh is /64 and must not be broken down, ir
        order to avoid all kind of unexpected consecuences and
        interoperability issues.

   5.   There is no NAT in IPv6, as there is no need for it.

   6.   It is expected that each customer is able to have more than
        sufficient subnets as may be needed, not just at the time being,
        but also in the near future.

   7.   Following [RFC7934], [RFC8273], [RFC9762] and [RFC9663], it is a
        good practice to allocate to each end-device no just a single
        IPv6 address per interface, but instead a complete /64 prefix in
        each interface, allowing the seamless support of features such
        as containers or virtual machines, among others.

   8.   As end-user networks keep growing, they become more complex and
        in IPv6 they contain multiple chained stub networks [RFC9818].

Palet Martinez          Expires 3 September 2026               [Page 12]
Internet-Draft           IPv6 Prefix Assignment               March 2026

   9.   As a consecuence, it is expected that each customer gets n x /64
        (not just a single /64), being "n" sufficiently big, for actual
        and future needs.

   10.  In summary, IPv6 is not the same as IPv4 and consequently
        previous IPv4 best practices are not, in general, the best
        advise.

   Considering the above premises, it is very clear that assigning to
   end-users a single /64 must be avoided at all means.

   The next possible choice will be a /60, which leaves only 16 /64s
   available (and one of them may be used for the point-to-point link).
   This seems clearly insuficiente for enterprises even for small/medium
   ones, and in general, considering the actual trends of deployment of
   smart homes, this will be become too small in the very near future,
   and should be avoided as well.

4.1.  /48 for every end-user

   The existing RIR policies for IPv6 allocation, allow to obtain by
   default at least a /32 basically without any justification.  In some
   cases, such as in RIPE NCC, even a /29 without any justification.  Is
   is also possible in RIPE NCC to create as many LIRs as needed for an
   organizacion, and every LIR could get a /29 without any
   justification).  A /32 contains 65.535 /48s, so even assuming that
   some of those /48s are lost because the network infrastructure,
   hierarchy, routing and so on, we can calculate a worst case scenario
   where a minimum of 50.000 /48s will become available for customers.

   A bigger service provider will need a bigger prefix, for example a
   provider with 100.000 customers will need a /31 instead of a /32, a
   provider with 200.000 customers will need a /30, and so on.  It
   should be clear that, as indicated already, all the RIR policies are
   consistent with this, which means that if you have more customers and
   you decide to provide a /48 for each end-user or end-site, you are
   able to obtain whatever prefix size you need.

   Based on that, here is no any practical reason to not assign /48s for
   every end-site, which in existing studies shows that is what is being
   done by a majority of service providers worldwide.

   This has several advantages worth to mention:

   1.  It match with the minimum IPv6 prefix routable in Internet.

   2.  It match with the ULA prefix size.

Palet Martinez          Expires 3 September 2026               [Page 13]
Internet-Draft           IPv6 Prefix Assignment               March 2026

   3.  It macth with the prefix size used by transition mechanisms.

   4.  It match with the standard direct assignment from RIRs to end-
       sites (Provider Indepentent, also called portable addresses).

   Keeping this matching facilitates possible network changes,
   exchanging service providers or moving to Provider Independent
   addresses, all that with a direct mapping of existing addressing
   plans of the end-site that obtained a prefix delegation.  At the same
   time, from the perspective of the service provider having a flat
   addressing plan, with the same delegated prefix size for all the
   customers it is clearly a big advantage, in terms of operations,
   monitoring, avoiding mistakes and much less call-centre time to sort
   out problems.

   An example of this, if we consider a service provider that got
   2001:db8::/32 and for a given PoP is delegating prefixes out of
   2001:db8:1000::/36, the 4.096 customers in this PoP will get:

   *  2001:db8:1000::/48

   *  2001:db8:1001::/48

   *  2001:db8:1002::/48

   *  2001:db8:1003::/48

   *  ...

   *  2001:db8:1fff::/48

   In this case, the end-users are able to use for their internal
   addressing plans the bits marked as "X" in the addressing space of
   2001:db8:1000:XXXX::/48.

   In summary, a /48 is a very common practice for prefix size
   assignments to end-users, being followed by many service providers
   worldwide.

4.2.  /48 for business customers and /52 or /56 for residential
      customers

   Some operators prefer, even if this increase their OPEX, to
   differentiate between business and residential customers.  This
   rationale is understood to be mainly coming from sales and marketing
   departments where they wish to create some distinction in services
   between different types of customer.  This method can be considered
   as pragmatic, future-proof and has nearly no downsides.

Palet Martinez          Expires 3 September 2026               [Page 14]
Internet-Draft           IPv6 Prefix Assignment               March 2026

   However it has negative implications for the customers, such as the
   lost of "matches" indicated in the previous section, which means that
   if a customer already has an addressing plan, they will be forced to
   redo it and renumber in case of a different prefix size.  Instead,
   using the same prefix size, allow them to keep the addressing plan
   and just replace some of the prefix bits rather than have a complete
   numbering plan change.  Note that in many cases, SMEs may be using
   residential services.

   For more advanced customers or SMEs that manually configure their
   servers and infrastructure, local addressing is a bit harder for /52
   or /56 prefixes, as you are unable to use all 4 digits between the
   double colons in the address notation.  This means that there is a
   higher risk of making mistakes whilst making the addressing plans and
   sub-assign addresses to devices outside the assigned prefix.  This
   should not be a problem in simple cases where a CE assigns all
   prefixes, but it might cause annoyances for those advanced users that
   may need to offer content and services from their own networks.

   A /56 only provides 256 subnets to customers, while a /52 is
   providing 4.096 subnets.  Even if 256 seems to be sufficient for most
   of the cases, if you start considering that many devices will be
   requiring a /64 per interface, then it quickly becomes too short.  So
   if the reason for not providing /48 to each end-site is the
   differentiation among business and residential customers, the choice
   of /52 seems much safer in terms of future proof and maintains that
   differentiation factor.

   Using the same example as in the previous section, if we consider a
   service provider that got 2001:db8::/32 and for a given PoP is
   delegating prefixes out of 2001:db8:1000::/36, then there will be
   space for 65.536 customers:

   *  2001:db8:1000:0000::/52

   *  2001:db8:1000:1000::/52

   *  2001:db8:1000:2000::/52

   *  2001:db8:1000:3000::/52

   *  ...

   *  2001:db8:1fff:f000::/52

   In this case, the end-users are able to use for their internal
   addressing plans the bits marked as "X" in the addressing space of
   2001:db8:1000:0XXX::/52.

Palet Martinez          Expires 3 September 2026               [Page 15]
Internet-Draft           IPv6 Prefix Assignment               March 2026

   Once more, using the same example, considering a service provider
   that got 2001:db8::/32 and for a given PoP is delegating prefixes out
   of 2001:db8:1000::/36, the 1.048.576 customers in this PoP will get:

   *  2001:db8:1000:0000::/56

   *  2001:db8:1000:0100::/56

   *  2001:db8:1000:0200::/56

   *  2001:db8:1000:0300::/56

   *  ...

   *  2001:db8:1fff:ff00::/56

   In this case, the end-users are able to use for their internal
   addressing plans the bits marked as "X" in the addressing space of
   2001:db8:1000:00XX::/56.

   An alternative is to reserve a /48 for residential customers, but
   actually assign them just the first /52 or /56.  If subsequently
   required, they can then be upgraded to the required prefix size
   without the need to renumber, or the spare prefixes can be used for
   new customers in the same PoP.  Note that this only makes sense in
   case of IPv6 exhaustion, as obtaining a bigger allocation from the
   relevant RIR seems feasible accordig to existing policies and the
   actual trends on policy development.

4.3.  Prefixes longer than /56

   It is strongly discouraged to assign prefixes longer than /56 unless
   there are very strong and unsolvable technical reasons for doing
   this.

   There are enough IPv6 addresses to delegate end-users a /48, so a /52
   or /56 already represents a sizeable restriction Just to give an idea
   about order of magnitude, there are 1.048.576 /52s in a /32.  There
   is no need to delegate fewer addresses than that, so if the service
   provider IPv6 allocation is insufficient to provide a /52 or /56 to
   each end-site, explain to the relevant RIR that your initial
   allocation was too small and that you require a larger allocation
   based on your IPv6 implementation plan.

Palet Martinez          Expires 3 September 2026               [Page 16]
Internet-Draft           IPv6 Prefix Assignment               March 2026

   Assigning a /64 or longer prefix does not conform to IPv6 standards
   and will break functionality in customer LANs.  With a single /64,
   the end customer CE will have just one possible network on the LAN
   side and it will not be possible to subnet, assign VLANs, alternative
   SSIDs, or have several chained routers in the same customer network
   as indicated by [RFC9818], etc.

   Some CEs use a /64 for the loopback interface and some may have
   multiple LAN segments predefined (for example Guest WiFi network and
   wired LAN), so as soon as there is more than one LAN segment behind
   the CE, exceptions will have to be added to the ISP provisioning
   system that will greatly complicate management.  It is not possible
   to assign less than a /64 to each LAN port/segment/subnet/VLAN due to
   IPv6 protocol requirements (SLAAC, ND, etc..) that reserve the last
   64 bits of the IPv6 addresses for the hosts.

   Recent documents such as [RFC7934], [RFC8273], [RFC9762] and
   [RFC9663] show that it is beccoming more convenient to assign /64 to
   each host interface (e.g. for security, isolation of customers, or
   having many virtual machines or containers on a single host), again
   explaining the need for delegating many /64s to each customer.

   A growing number of operators are also using prefix colouring to
   deploy products with distinct Service Level Agreements (e.g. voice,
   data, video) and this requires at least a unique /64 for each product
   or service.  If combined with home networking technologies, the
   number of prefixes increases quite quickly and delegation of multiple
   IPv6 prefixes at the same time will occur, so assignments of less
   than /56 will probably require renumbering in the near future.

   The Broadband Forum also recommends a similar approach in their
   [TR-177] document: "A delegated prefix for use within the home
   network (mandatory).  The Broadband Forum suggests a size for the
   delegated prefix of least a /60 for home network or SOHO
   environments, with a recommended prefix length of /56.  The delegated
   prefix may be extended to a /48 for larger organizations.".

4.4.  Considerations for Cellular Operators

   There is a clear exception to the rules described above when
   assigning prefixes in a cellular network.  In this case, a /64 will
   need to be provided for each PDP context for cellular phones, whereas
   for celullar modems/routers, i.e. in the case of broadband by means
   of cellular access, it will still be necessary to choose a /48, /52
   or /56 in accordance with the aforementioned considerations.

Palet Martinez          Expires 3 September 2026               [Page 17]
Internet-Draft           IPv6 Prefix Assignment               March 2026

5.  Assigned Prefix Persistence Considerations

   Because the scarcity of IPv4 addresses, the most common practice for
   service providers was to assign temporary or "dynamic" leases of
   addresses to the CEs.  We could not actually say this was a real
   "best practice", but instead the only choice to share addresses among
   the majority of the customers, and then use exceptions in
   provisioning systems for "static" leases of individual addresses,
   avoiding manual configurations.  We could also say that it is a best
   practice for end-site IPv4 routed prefixes to be assigned in such way
   that makes them permanent.

   In the end, this created some wording confusion, because "dynamic" is
   associated to DHCP, which is "the protocol" typically being used for
   those leases, even if they are "static" (vs "manual" configuration).
   This is the reason, to avoid wording confusion, is preferable to use
   a differenciated terminology: persistent vs non-persistent instead of
   dynamic/static.

   Persistent prefix assignment means that a prefix is assigned to a
   customer and that prefix will always remain the same regardless of
   how many times the customer (re)connects or renews the lease.
   Typically this is achieved by means of an AAA mechanism, which may be
   dependent on DHCPv6 and/or other provisioning systems.  It is
   “persistent prefix” for the same customer in the same link/location
   regardless of the provisioning system being used.

   DHCPv6-PD [RFC8415] is commonly used for non-persistent assignments,
   in the same way it is common for IPv4, where a bigger pool of IPv6
   space is assigned to each termination point (PoP with BNG, BRAS,
   etc.), and so as customers connect, they are randomly assigned
   prefixes from this when they (re)connect or the lease time expires.
   They may get the same prefix, but typically it will be different
   (non-persistent) unless the lease time is so high that it effectively
   become “persistent”. In this case, a customer will get the same
   prefix even if they switch off their router for several days, weeks
   or months, depending on the lease time.

   Often persistency is viewed as against privacy or even security,
   however, a persistent prefix (specially in the case of IPv6
   considering the number of individual addresses per prefix) is not
   pointing towards a specific device and even less to an individual.
   Actually even with non-persistent prefixes the situation is the same
   and is only a matter of timeframe.  The only difference is in case of
   using that "persistency" for "attacking" a specific individual as the
   attacker will have more time than if the service provider renew the
   prefix for example every 24 hours; should then we suggest that the
   renewal should be done every hour or every 30 minutes?  What about

Palet Martinez          Expires 3 September 2026               [Page 18]
Internet-Draft           IPv6 Prefix Assignment               March 2026

   the damage created to existing sessions with each renewal?. Operating
   systems already deploy mitigations such as privacy extensions
   [RFC4941].

5.1.  Non-persistent assignments perceived as 'easier'

   When faced with the task of how to deploy IPv6, it is easy to go for
   an option that you're used to (in IPv4), initially requires less
   effort, time and energy.  However, this will usually create more
   problems later on.  As we indicaded before, IPv6 is not the same as
   IPv4, and best current practices in IPv4 are not neccesarily best
   current practices in IPv6, most of the time in the other way around,
   they may somehow ruin your deployment efforts.

   The easiest method is to assign prefixes from a pool of IPv6 prefixes
   to termination points (BNG, BRAS or other equipment, depending on the
   type of access network) and let the provisioning system decide what
   customers will get when they connect and ask for a prefix delegation
   over DHCPv6-PD.  As already indicated above, this practice is carried
   over from the IPv4 world where addresses are commonly assigned
   "dynamically" (non-persistent) and where CEs perform NAT to conserve
   IPv4 space.  Bear in mind that end-sites with an IPv4 subnet behind
   their CE never got “non-persistent” assigned IPv4 prefixes as this
   would require reconfiguration of all hosts on their network every few
   hours or days.  Instead they are typically using private IPv4
   addresses as per [RFC1918], which get translated by the NAT to a
   single public IPv4 address or a small block of them.  By contrast,
   because there is no scarcity of IPv6 addresses, NAT is not needed and
   every IPv6 end-site can get a sufficiently big IPv6 prefix so it is
   unnecessary to apply this extra complexity of the “IPv4 model” to
   IPv6.

Palet Martinez          Expires 3 September 2026               [Page 19]
Internet-Draft           IPv6 Prefix Assignment               March 2026

   Non-persistent prefix assignment also appears initially easier as it
   facilitates aggregation of internal routing tables according to end
   customer connection termination points.  Every termination point has
   its own pool of IPv6 prefixes that are nicely aggregable, whilst it
   may appear that with persistent IPv6 prefix assignments it is
   necessary to discover which customer is terminated at which
   termination point, group them into larger IPv6 pools, and then update
   our database accordingly.  However, this is resolved by the
   provisioning system doing it in the other way around: from an AAA
   database, which is typically tied to an IPAM for the initial
   assignment to each new customer.  In this way, when a new end-site
   joins a network for the first time, it is provisioned with a prefix
   from the pool assigned to the PoP where it actually connects (so it
   is already agregatted in that PoP), and this information is recorded
   in the AAA for future connections.  Note also that in many networks,
   depending on their size there may be a single PoP, so this is a
   lesser issue.

   Some operators may wonder what happens when a customer moves to
   another city or neighbourhood.  Moving to a different location
   typically means all the equipment is switched off whilst transporting
   them.  If services must remain on-line, clearly there has been some
   previous planning, services have been replicated in the new end-site,
   and probably there was a need to reconfigure them, including an
   actual renumbering with the new assigned prefix.  Since it will be
   necessary to provision the new connection elsewhere, then it is
   perfectly fine to also change the prefix of the customer to match the
   criteria for that aggregation point.  This can be avoided if the new
   end-site link is in the same “aggregation region/point” and the
   customer can be off-line for a certain period of time while moving
   the equipment from one location to another.  If the customer
   specifically requests that the same IPv6 prefix moves with them
   across different aggregation points, service providers may need
   special internal routing table changes and the service provider may
   decide to offer that service at an extra cost or not offering it at
   all.  However, that should be an uncommon scenario.

5.2.  Non-persistent assignments considered harmful

   Taking a scenario where there is a connected CE with a dynamically
   assigned prefix (e.g. 2001:db8:aaaa::/48).  The WAN link may use the
   first /64 segment (2001:db8:aaaa:0::/64) and the CE may sub-assign
   local loopback addresses out of the first /64 segment
   (2001:db8:aaaa:1::/64) and sub-assign 2001:db8:aaaa:2::/64 to the
   first LAN interface (probably bridged with the main WiFi SSID).  At
   the same time, there may be a guest WiFi SSID (2001:db8:aaaa:3::/64)
   and even a specific WiFi SSID for IoT devices (2001:db8:aaaa:4::/64).
   As there are some smart devices running containers of VMs in the

Palet Martinez          Expires 3 September 2026               [Page 20]
Internet-Draft           IPv6 Prefix Assignment               March 2026

   network (even exposing services to Internet), there may be some
   additional /64s sub-assigned to their interfaces and moreover, some
   shorter prefixes, for example /60s for other parts of the network.

   So, to make it simple, let's consider that the hosts on the default/
   main wired/WiFi network autoconfigure IPv6 addresses out of
   2001:db8:aaaa:2::/64 segment and start communicating with the rest of
   the Internet.

   Now all of a sudden there is a power outage (which is very common in
   many regions/countries in the world, even “highly-developed”
   countries) or the CE freezes and reboots and the connection has to be
   established again.  This time a new IPv6 prefix is assigned:
   2001:db8:bbbb::/48.  If the CE knows that the delegated prefix has
   changed, it should send out RA packets with a prefix valid lifetime
   of 0 to tell all devices that the old addresses are no longer valid.
   However, the CE rarely knows that before the reboot there was a
   different prefix on the network, and the packets to revoke the old
   IPv6 addresses do not get sent.  In this case, multiple IPv6
   addresses from completely different assigned prefixes end up on the
   same network interface, some of which will no longer work and may
   imply increase the number of claims to the service provide help desk.
   This gets even more complicated if there are multiple consecutive
   power outages (very common as well) affecting the CE.

   Just consider the added complexity of the other prefixes being sub-
   assigned across the different segments on the network, either using
   SLAAC or DHCPv6-PD.  All them keep trying connecting to Internet
   using the old prefix.

   Different OS vendors treat this scenario differently, but there often
   ends up being a wrong source IPv6 address for sending packets through
   the CE out to the Internet.  As the network operator's equipment
   deleted the previous delegated prefix route back to the CE, any
   return packets never reach the originating device and IPv6
   connectivity will be broken until the old IPv6 addresses time-out and
   are automatically removed from the interfaces.  If another
   reconnection to the ISP is required in the interim, there will be a
   third set of IPv6 addresses from yet another assigned prefix, and
   this will cause even more confusion.

   Several big content providers and device vendors are measuring “IPv6
   brokenness” in operator networks by matching the SYN, SYN ACK and ACK
   packets received/sent to/from a single source.  If they see a SYN and
   send back SYN ACK, but never receive ACK in a timely manner, they
   slowly stop serving AAAA records for the ASNs where they see this, as
   they prefer a stable IPv4 connectivity even if it presents a higher
   latency (due to NATs/CGNs), versus a bad IPv6 network.  It is a

Palet Martinez          Expires 3 September 2026               [Page 21]
Internet-Draft           IPv6 Prefix Assignment               March 2026

   matter of those services or device vendors offering "best quality of
   service" to their customers.  So, assigning IPv6 non-persistent
   prefixes, can imply that some content providers or device vendors
   start ignoring your IPv6 traffic quite quickly and force your
   customers back to IPv4.  And these are usually destinations (big
   content providers) where you will be sending a significant amount of
   traffic.  Clearly this has implications in networks using CGN, as the
   usage of CGN will significantly increase.

   Using non-persistent prefixes also means that it is necessary to have
   a logging system that registers which WAN and LAN prefixes are being
   used at each moment by each customer site, which additional storage
   capacity, in order to comply with legal/regulatory requirements.

   Finally, if users have services exposed to Internet, such as web,
   email or VPNs, they typically need to manually configure the
   addresses of those, so using non-persistent prefixes is not an option
   in this case.  With the growth of broadband capacities (such as
   FTTH), it is becoming more and more common that end-sites run servers
   or services on their LANs, including the likes of IP cameras or
   security systems in a home.  Persistent prefixes allows FQDNs to be
   assigned to individual addresses of those prefixes.  In fact,
   persistent-prefixes offers the service providers the oportunity to
   sell/resell value added services to customers, including managed DNS
   services, without the complexity of developing systems that
   continuously keep track of the prefixes assigned to each customer.

   Using non-persistent prefix delegation, enforces the service
   providers to verify that the CEs used by the customers will not have
   the problems described above.  Commonly this can not be ensured, so
   the current best practice is to avoid using non-persistent prefixes.

   It also needs to be realised that non-persistent prefixes have other
   implications and additional OPEX, because it is not only the CEs that
   need to be numbered, but all the devices behind them in the customer
   LANs as well.  Different implementations can have very different
   behaviors that may affect the number of support calls to service
   provider help-desk every time a prefix changes.

   Last, but not least, exactly the same as with IPv4, when the end-site
   IPv6 assigned prefix is renumbered, even if correctly done by the
   service provider (instead of a power outage or CE reboot), many
   running applications will not survive, creating problems to users and
   applications, which in turn can imply subsequent help-desk calls and
   even possible legal or consumer claims.

Palet Martinez          Expires 3 September 2026               [Page 22]
Internet-Draft           IPv6 Prefix Assignment               March 2026

   This has been extensively described in [RFC8978] and a solution
   proposed by [RFC9096].  The standard updates required for a
   definitive solution is work in progress in
   [I-D.ietf-6man-slaac-renum].  Note that even when this document is
   approved, it doesn't resolve many issues, such as those related to
   manual configuration changes required in the network, for instance,
   DNS, devices renumbering, servers, containers or VM renumbering.  In
   addition to that, it may take many years to get implemented in CEs
   and we know that many of them remain operating in the networks
   without firmware updates for many years, so is not a practical
   solucion to "wait".

5.3.  Persistent assignments are the best current practice

   As already mentioned, a best practice in IPv4 was to assign
   persistent IPv4 prefixes whilst giving an end-user a routed prefix,
   and same is true for IPv6 where persistent prefix assignment is
   strongly recommended.  It is comparable to assign IPv4 prefixes to
   assigning IPv6 ones.  It is not comparable to assign IPv4 addresses
   to IPv6 prefixes.

   If required for provisioning, the connection between a network and
   the end-site CE can be non-persistent using /64 (or even /127 if the
   equipment supports it), but choosing persistent IPv6 assigned
   prefixes for end-user LANs will avoid a lot of difficulties
   experienced by some earlier IPv6 network operators.

   This is especially the case where there are non-residential customers
   as well, which typically will use residential links, as this means
   you can have a single provisioning system and it is unnecessary to
   maintain a logging system, increasing also the security of the
   network because the undubitable identificacion of end-sites and
   disallowing connections for end-sites that have not the proper
   authorization in the AAA.  This is because each end-site always have
   the same prefix(es) and can be identified in the AAA or provisioning
   system, reducing complexity and making things cheaper.  A bit of
   initial thought, planning and consideration for future needs can
   therefore save a great deal of time and energy when IPv6 has been
   deployed.

   End-site users will establish a reputation for their prefix over time
   based on things like compromised devices and malware within their
   networks.  If there isn't persistence then users will pick up tainted
   reputation from past users of that space.  Operators should publish a
   feed of their prefix lenghts as per [I-D.ietf-opsawg-prefix-lengths],
   so that IP reputation systems can take these into account.  This
   specially matters if parts of an AS are using widely differing
   lenghts (e.g., /48 in some and /52 in others).

Palet Martinez          Expires 3 September 2026               [Page 23]
Internet-Draft           IPv6 Prefix Assignment               March 2026

   From the economic/business perspective, another important
   consideration with persistent prefixes is that it is possible to
   ‘name' value added services using DNS (e.g.
   camera1.username.ispname.com) that can result in considerable new
   income streams.  Trying to deploy new services or applications with
   non-persistent prefixes is always more difficult and costly, and will
   increase time spent on operations, management and troubleshooting.
   This is just one example to make clear that persistent prefixes
   enable innovation, new services and applications, which in turn means
   new business possibilities.

   Further to the above, it should be noted that persistent assignments
   of prefixes to customers may increase the incentive for customers to
   remain with that ISP, and result in decreased customer churn.

   Finally, if the use of persistent prefixes is not feasible for
   unavoidable technical reasons, using non-persistent but long-lived
   assignments must be considered, so the lease time is as long as the
   provisioning system allows.  Note that there are several choices to
   keep end-sites correctly correlated to the AAA information for the
   correct prefix provisioning, and not just differentiated user/
   password (which in many service providers is the same for all the
   customers to simplify the CE provisioning).  For example, using UUIDs
   or DUIDs [RFC6355]/[RFC8415].  Even in the case where DHCPv6 relay
   are required, Relay Agent Remote-ID (37) can be used as per
   [RFC4649].

   It should be remarked that persistence of prefixes shoud be expected
   only when customers are connected to the same aggregation point.
   Outside of that, customers moving to different locations, has
   different routing and consequently business considerations.

6.  Best Common Practices Summary for IPv6 Prefix Assignments

   We can summarize the best common practices for IPv6 prefix
   assignments to end-sites as follows:

   1.  The prefix assigned to end-site should be a GUA /48, using
       DHCPv6-PD [RFC8415].

   2.  The prefix assigned to end-site should be persistent.

   3.  The prefix for the WAN point-to-point link should be a GUA /64, a
       GUA /127 or Link Local Addresses.

   4.  The prefix for the WAN point-to-point link could be the first /64
       from the assigned end-site prefix, using the Prefix Exclude
       Option for DHCPv6-based Prefix Delegation ([RFC6603]).

Palet Martinez          Expires 3 September 2026               [Page 24]
Internet-Draft           IPv6 Prefix Assignment               March 2026

   5.  The assigned prefix must not be longer than /56.

   6.  Operators should publish a feed of their prefix lenghts as per
       [I-D.ietf-opsawg-prefix-lengths], so that IP reputation systems
       can take these into account.

7.  Security Considerations

   This document does not have any new specific security considerations.

8.  IANA Considerations

   This document does not have any new specific IANA considerations.

9.  Acknowledgements

   The author would like to acknowledge the inputs of Brian Carpenter,
   Daryll Swer, Erik Nygren, Richard Patterson, Michael Richardson, Mark
   Smith, Lorenzo Colitti, Jen Linkova, Tom Hill, Nick Buraglio, Ron
   Bonica, Eric Vyncke and Mikael Abrahamsson.

   Acknowledge is also due to my co-authors of RIPE-690 ("Best Current
   Operational Practice for Operators: IPv6 prefix assignment for end-
   users - persistent vs non-persistent, and what size to choose",
   https://www.ripe.net/publications/docs/ripe-690) and global
   community, which provided valuable inputs that were key for this
   document.

   Acknowledgement to co-authors, Cesar Olvera and Miguel Angel Díaz, of
   a previous related document (draft-palet-v6ops-point2point, 2006), as
   well as inputs for that version from Alain Durand, Chip Popoviciu,
   Daniel Roesen, Fred Baker, Gert Doering, Olaf Bonness, Ole Troan,
   Pekka Savola and Vincent Jardin, are also granted.

10.  References

10.1.  Normative References

   [I-D.ietf-6man-slaac-renum]
              Gont, F., Zorz, J., Patterson, R., and J. Linkova,
              "Improving the Robustness of Stateless Address
              Autoconfiguration (SLAAC) to Flash Renumbering Events",
              Work in Progress, Internet-Draft, draft-ietf-6man-slaac-
              renum-13, 15 February 2026,
              <https://datatracker.ietf.org/doc/html/draft-ietf-6man-
              slaac-renum-13>.

Palet Martinez          Expires 3 September 2026               [Page 25]
Internet-Draft           IPv6 Prefix Assignment               March 2026

   [I-D.ietf-opsawg-prefix-lengths]
              Gasser, O., Bush, R., Candela, M., and R. Housley,
              "Publishing End-Site Prefix Lengths", Work in Progress,
              Internet-Draft, draft-ietf-opsawg-prefix-lengths-14, 17
              December 2025, <https://datatracker.ietf.org/doc/html/
              draft-ietf-opsawg-prefix-lengths-14>.

   [RFC2119]  Bradner, S., "Key words for use in RFCs to Indicate
              Requirement Levels", BCP 14, RFC 2119,
              DOI 10.17487/RFC2119, March 1997,
              <https://www.rfc-editor.org/info/rfc2119>.

   [RFC3769]  Miyakawa, S. and R. Droms, "Requirements for IPv6 Prefix
              Delegation", RFC 3769, DOI 10.17487/RFC3769, June 2004,
              <https://www.rfc-editor.org/info/rfc3769>.

   [RFC4291]  Hinden, R. and S. Deering, "IP Version 6 Addressing
              Architecture", RFC 4291, DOI 10.17487/RFC4291, February
              2006, <https://www.rfc-editor.org/info/rfc4291>.

   [RFC4443]  Conta, A., Deering, S., and M. Gupta, Ed., "Internet
              Control Message Protocol (ICMPv6) for the Internet
              Protocol Version 6 (IPv6) Specification", STD 89,
              RFC 4443, DOI 10.17487/RFC4443, March 2006,
              <https://www.rfc-editor.org/info/rfc4443>.

   [RFC4649]  Volz, B., "Dynamic Host Configuration Protocol for IPv6
              (DHCPv6) Relay Agent Remote-ID Option", RFC 4649,
              DOI 10.17487/RFC4649, August 2006,
              <https://www.rfc-editor.org/info/rfc4649>.

   [RFC4861]  Narten, T., Nordmark, E., Simpson, W., and H. Soliman,
              "Neighbor Discovery for IP version 6 (IPv6)", RFC 4861,
              DOI 10.17487/RFC4861, September 2007,
              <https://www.rfc-editor.org/info/rfc4861>.

   [RFC4941]  Narten, T., Draves, R., and S. Krishnan, "Privacy
              Extensions for Stateless Address Autoconfiguration in
              IPv6", RFC 4941, DOI 10.17487/RFC4941, September 2007,
              <https://www.rfc-editor.org/info/rfc4941>.

   [RFC6355]  Narten, T. and J. Johnson, "Definition of the UUID-Based
              DHCPv6 Unique Identifier (DUID-UUID)", RFC 6355,
              DOI 10.17487/RFC6355, August 2011,
              <https://www.rfc-editor.org/info/rfc6355>.

Palet Martinez          Expires 3 September 2026               [Page 26]
Internet-Draft           IPv6 Prefix Assignment               March 2026

   [RFC6603]  Korhonen, J., Ed., Savolainen, T., Krishnan, S., and O.
              Troan, "Prefix Exclude Option for DHCPv6-based Prefix
              Delegation", RFC 6603, DOI 10.17487/RFC6603, May 2012,
              <https://www.rfc-editor.org/info/rfc6603>.

   [RFC6724]  Thaler, D., Ed., Draves, R., Matsumoto, A., and T. Chown,
              "Default Address Selection for Internet Protocol Version 6
              (IPv6)", RFC 6724, DOI 10.17487/RFC6724, September 2012,
              <https://www.rfc-editor.org/info/rfc6724>.

   [RFC7084]  Singh, H., Beebee, W., Donley, C., and B. Stark, "Basic
              Requirements for IPv6 Customer Edge Routers", RFC 7084,
              DOI 10.17487/RFC7084, November 2013,
              <https://www.rfc-editor.org/info/rfc7084>.

   [RFC7136]  Carpenter, B. and S. Jiang, "Significance of IPv6
              Interface Identifiers", RFC 7136, DOI 10.17487/RFC7136,
              February 2014, <https://www.rfc-editor.org/info/rfc7136>.

   [RFC7934]  Colitti, L., Cerf, V., Cheshire, S., and D. Schinazi,
              "Host Address Availability Recommendations", BCP 204,
              RFC 7934, DOI 10.17487/RFC7934, July 2016,
              <https://www.rfc-editor.org/info/rfc7934>.

   [RFC8174]  Leiba, B., "Ambiguity of Uppercase vs Lowercase in RFC
              2119 Key Words", BCP 14, RFC 8174, DOI 10.17487/RFC8174,
              May 2017, <https://www.rfc-editor.org/info/rfc8174>.

   [RFC8273]  Brzozowski, J. and G. Van de Velde, "Unique IPv6 Prefix
              per Host", RFC 8273, DOI 10.17487/RFC8273, December 2017,
              <https://www.rfc-editor.org/info/rfc8273>.

   [RFC8415]  Mrugalski, T., Siodelski, M., Volz, B., Yourtchenko, A.,
              Richardson, M., Jiang, S., Lemon, T., and T. Winters,
              "Dynamic Host Configuration Protocol for IPv6 (DHCPv6)",
              RFC 8415, DOI 10.17487/RFC8415, November 2018,
              <https://www.rfc-editor.org/info/rfc8415>.

   [RFC9096]  Gont, F., Žorž, J., Patterson, R., and B. Volz, "Improving
              the Reaction of Customer Edge Routers to IPv6 Renumbering
              Events", BCP 234, RFC 9096, DOI 10.17487/RFC9096, August
              2021, <https://www.rfc-editor.org/info/rfc9096>.

   [RFC9663]  Colitti, L., Linkova, J., Ed., and X. Ma, Ed., "Using
              DHCPv6 Prefix Delegation (DHCPv6-PD) to Allocate Unique
              IPv6 Prefixes per Client in Large Broadcast Networks",
              RFC 9663, DOI 10.17487/RFC9663, October 2024,
              <https://www.rfc-editor.org/info/rfc9663>.

Palet Martinez          Expires 3 September 2026               [Page 27]
Internet-Draft           IPv6 Prefix Assignment               March 2026

   [RFC9762]  Colitti, L., Linkova, J., Ma, X., Ed., and D. Lamparter,
              "Using Router Advertisements to Signal the Availability of
              DHCPv6 Prefix Delegation to Clients", RFC 9762,
              DOI 10.17487/RFC9762, June 2025,
              <https://www.rfc-editor.org/info/rfc9762>.

   [RFC9818]  Winters, T., "DHCPv6 Prefix Delegation on IPv6 Customer
              Edge (CE) Routers in LANs", RFC 9818,
              DOI 10.17487/RFC9818, July 2025,
              <https://www.rfc-editor.org/info/rfc9818>.

10.2.  Informative References

   [IPv6-Survey]
              Palet Martinez, J., "IPv6 Deployment Survey (Residential/
              Household Services)", January 2018,
              <https://indico.uknof.org.uk/event/41/contribution/5/
              material/slides/0.pdf>.

   [RFC1918]  Rekhter, Y., Moskowitz, B., Karrenberg, D., de Groot, G.
              J., and E. Lear, "Address Allocation for Private
              Internets", BCP 5, RFC 1918, DOI 10.17487/RFC1918,
              February 1996, <https://www.rfc-editor.org/info/rfc1918>.

   [RFC3627]  Savola, P., "Use of /127 Prefix Length Between Routers
              Considered Harmful", RFC 3627, DOI 10.17487/RFC3627,
              September 2003, <https://www.rfc-editor.org/info/rfc3627>.

   [RFC6164]  Kohno, M., Nitzan, B., Bush, R., Matsuzaki, Y., Colitti,
              L., and T. Narten, "Using 127-Bit IPv6 Prefixes on Inter-
              Router Links", RFC 6164, DOI 10.17487/RFC6164, April 2011,
              <https://www.rfc-editor.org/info/rfc6164>.

   [RFC6583]  Gashinsky, I., Jaeggli, J., and W. Kumari, "Operational
              Neighbor Discovery Problems", RFC 6583,
              DOI 10.17487/RFC6583, March 2012,
              <https://www.rfc-editor.org/info/rfc6583>.

   [RFC6752]  Kirkham, A., "Issues with Private IP Addressing in the
              Internet", RFC 6752, DOI 10.17487/RFC6752, September 2012,
              <https://www.rfc-editor.org/info/rfc6752>.

   [RFC7404]  Behringer, M. and E. Vyncke, "Using Only Link-Local
              Addressing inside an IPv6 Network", RFC 7404,
              DOI 10.17487/RFC7404, November 2014,
              <https://www.rfc-editor.org/info/rfc7404>.

Palet Martinez          Expires 3 September 2026               [Page 28]
Internet-Draft           IPv6 Prefix Assignment               March 2026

   [RFC7608]  Boucadair, M., Petrescu, A., and F. Baker, "IPv6 Prefix
              Length Recommendation for Forwarding", BCP 198, RFC 7608,
              DOI 10.17487/RFC7608, July 2015,
              <https://www.rfc-editor.org/info/rfc7608>.

   [RFC7707]  Gont, F. and T. Chown, "Network Reconnaissance in IPv6
              Networks", RFC 7707, DOI 10.17487/RFC7707, March 2016,
              <https://www.rfc-editor.org/info/rfc7707>.

   [RFC8978]  Gont, F., Žorž, J., and R. Patterson, "Reaction of IPv6
              Stateless Address Autoconfiguration (SLAAC) to Flash-
              Renumbering Events", RFC 8978, DOI 10.17487/RFC8978, March
              2021, <https://www.rfc-editor.org/info/rfc8978>.

   [RIPE-690] "Best Current Operational Practice for Operators: IPv6
              prefix assignment for end-users - persistent vs non-
              persistent, and what size to choose", October 2017,
              <https://www.ripe.net/publications/docs/ripe-690/>.

   [TR-177]   The Broadband Forum, "TR-177 - IPv6 in the context of TR-
              101", November 2017,
              <https://www.broadband-forum.org/pdfs/tr-177-1-0-1.pdf>.

Author's Address

   Jordi Palet Martinez
   The IPv6 Company
   Molino de la Navata, 75
   28420 La Navata - Galapagar Madrid
   Spain
   Email: jordi.palet@theipv6company.com
   URI:   http://www.theipv6company.com/

Palet Martinez          Expires 3 September 2026               [Page 29]