RFC 8475






Internet Engineering Task Force (IETF)                        J. Linkova
Request for Comments: 8475                                        Google
Category: Informational                                       M. Stucchi
ISSN: 2070-1721                                                 RIPE NCC
                                                            October 2018


   Using Conditional Router Advertisements for Enterprise Multihoming

Abstract



   This document discusses the most common scenarios of connecting an
   enterprise network to multiple ISPs using an address space assigned
   by an ISP and how the approach proposed in "Enterprise Multihoming
   using Provider-Assigned Addresses without Network Prefix Translation:
   Requirements and Solution" could be applied in those scenarios.  The
   problem of enterprise multihoming without address translation of any
   form has not been solved yet as it requires both the network to
   select the correct egress ISP based on the packet source address and
   hosts to select the correct source address based on the desired
   egress ISP for that traffic.  The aforementioned document proposes a
   solution to this problem by introducing a new routing functionality
   (Source Address Dependent Routing) to solve the uplink selection
   issue.  It also proposes using Router Advertisements to influence the
   host source address selection.  It focuses on solving the general
   problem and covering various complex use cases, and this document
   adopts its proposed approach to provide a solution for a limited
   number of common use cases.  In particular, the focus of this
   document is on scenarios in which an enterprise network has two
   Internet uplinks used either in primary/backup mode or simultaneously
   and hosts in that network might not yet properly support multihoming
   as described in RFC 8028.

Status of This Memo



   This document is not an Internet Standards Track specification; it is
   published for informational purposes.

   This document is a product of the Internet Engineering Task Force
   (IETF).  It represents the consensus of the IETF community.  It has
   received public review and has been approved for publication by the
   Internet Engineering Steering Group (IESG).  Not all documents
   approved by the IESG are candidates for any level of Internet
   Standard; see Section 2 of RFC 7841.

   Information about the current status of this document, any errata,
   and how to provide feedback on it may be obtained at
   https://www.rfc-editor.org/info/rfc8475.



Linkova & Stucchi             Informational                     [Page 1]

RFC 8475                     Conditional RAs                October 2018


Copyright Notice



   Copyright (c) 2018 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 Simplified BSD License text as described in Section 4.e of
   the Trust Legal Provisions and are provided without warranty as
   described in the Simplified BSD License.

Table of Contents



   1.  Introduction  . . . . . . . . . . . . . . . . . . . . . . . .   3
     1.1.  Requirements Language . . . . . . . . . . . . . . . . . .   4
   2.  Common Enterprise Multihoming Scenarios . . . . . . . . . . .   4
     2.1.  Two ISP Uplinks, Primary and Backup . . . . . . . . . . .   4
     2.2.  Two ISP Uplinks, Used for Load-Balancing  . . . . . . . .   5
   3.  Conditional Router Advertisements . . . . . . . . . . . . . .   5
     3.1.  Solution Overview . . . . . . . . . . . . . . . . . . . .   5
       3.1.1.  Uplink Selection  . . . . . . . . . . . . . . . . . .   5
       3.1.2.  Source Address Selection and Conditional RAs  . . . .   5
     3.2.  Example Scenarios . . . . . . . . . . . . . . . . . . . .   8
       3.2.1.  Single Router, Primary/Backup Uplinks . . . . . . . .   8
       3.2.2.  Two Routers, Primary/Backup Uplinks . . . . . . . . .   9
       3.2.3.  Single Router, Load-Balancing between Uplinks . . . .  12
       3.2.4.  Two Routers, Load-Balancing between Uplinks . . . . .  12
       3.2.5.  Topologies with Dedicated Border Routers  . . . . . .  13
       3.2.6.  Intrasite Communication during Simultaneous Uplinks
               Outage  . . . . . . . . . . . . . . . . . . . . . . .  15
       3.2.7.  Uplink Damping  . . . . . . . . . . . . . . . . . . .  15
       3.2.8.  Routing Packets When the Corresponding Uplink Is
               Unavailable . . . . . . . . . . . . . . . . . . . . .  16
     3.3.  Solution Limitations  . . . . . . . . . . . . . . . . . .  16
       3.3.1.  Connections Preservation  . . . . . . . . . . . . . .  17
   4.  IANA Considerations . . . . . . . . . . . . . . . . . . . . .  17
   5.  Security Considerations . . . . . . . . . . . . . . . . . . .  18
     5.1.  Privacy Considerations  . . . . . . . . . . . . . . . . .  18
   6.  References  . . . . . . . . . . . . . . . . . . . . . . . . .  18
     6.1.  Normative References  . . . . . . . . . . . . . . . . . .  18
     6.2.  Informative References  . . . . . . . . . . . . . . . . .  20
   Acknowledgements  . . . . . . . . . . . . . . . . . . . . . . . .  20
   Authors' Addresses  . . . . . . . . . . . . . . . . . . . . . . .  21




Linkova & Stucchi             Informational                     [Page 2]

RFC 8475                     Conditional RAs                October 2018


1.  Introduction



   Multihoming is an obvious requirement for many enterprise networks to
   ensure the desired level of network reliability.  However, using more
   than one ISP (and address space assigned by those ISPs) introduces
   the problem of assigning IP addresses to hosts.  In IPv4, there is no
   choice but using address space [RFC1918] and NAT [RFC3022] at the
   network edge [RFC4116].  Using Provider Independent (PI) address
   space is not always an option, since it requires running BGP between
   the enterprise network and the ISPs.  The administrative overhead of
   obtaining and managing PI address space can also be a concern.  As
   IPv6 hosts can, by design, have multiple addresses of the global
   scope [RFC4291], multihoming using provider addresses looks even
   easier for IPv6: each ISP assigns an IPv6 block (usually /48), and
   hosts in the enterprise network have addresses assigned from each ISP
   block.  However, using IPv6 provider-assigned (PA) blocks in a
   multihoming scenario introduces some challenges, including, but not
   limited to:

   o  Selecting the correct uplink based on the packet source address;

   o  Signaling to hosts that some source addresses should or should not
      be used (e.g., an uplink to the ISP went down or became available
      again).

   [PROVIDER-ASSIGNED] discusses these and other related challenges in
   detail in relation to the general multihoming scenario for enterprise
   networks.  It proposes a solution that relies heavily on Rule 5.5 of
   the default address selection algorithm [RFC6724].  Rule 5.5 makes
   hosts prefer source addresses in a prefix advertised by the next hop
   and, therefore, is very useful in multihomed scenarios when different
   routers may advertise different prefixes.  While [RFC6724] defines
   Rule 5.5 as optional, the recent [RFC8028] recommends that multihomed
   hosts SHOULD support it.  Unfortunately, that rule has not been
   widely implemented at the time of writing.  Therefore, network
   administrators in enterprise networks can't yet assume that all
   devices in their network support Rule 5.5, especially in the quite
   common BYOD ("Bring Your Own Device") scenario.  However, while it
   does not seem feasible to solve all the possible multihoming
   scenarios without relying on Rule 5.5, it is possible to provide IPv6
   multihoming using PA address space for the most common use cases.
   This document discusses how the general approach described in
   [PROVIDER-ASSIGNED] can be applied to solve multihoming scenarios
   when:

   o  An enterprise network has two or more ISP uplinks;





Linkova & Stucchi             Informational                     [Page 3]

RFC 8475                     Conditional RAs                October 2018


   o  Those uplinks are used for Internet access in active/backup or
      load-sharing mode without any sophisticated traffic engineering
      requirements;

   o  Each ISP assigns the network a subnet from its own PA address
      space; and

   o  Hosts in the enterprise network are not expected to support Rule
      5.5 of the default address selection algorithm [RFC6724].

1.1.  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.

2.  Common Enterprise Multihoming Scenarios



2.1.  Two ISP Uplinks, Primary and Backup



   This scenario has the following key characteristics:

   o  The enterprise network uses uplinks to two (or more) ISPs for
      Internet access;

   o  Each ISP assigns IPv6 PA address space for the network;

   o  Uplink(s) to one ISP is a primary (preferred) one.  All other
      uplinks are backup and are not expected to be used while the
      primary one is operational;

   o  If the primary uplink is operational, all Internet traffic should
      flow via that uplink;

   o  When the primary uplink fails, the Internet traffic needs to flow
      via the backup uplinks;

   o  Recovery of the primary uplink needs to trigger the traffic
      switchover from the backup uplinks back to the primary one;

   o  Hosts in the enterprise network are not expected to support Rule
      5.5 of the default address selection algorithm [RFC6724].







Linkova & Stucchi             Informational                     [Page 4]

RFC 8475                     Conditional RAs                October 2018


2.2.  Two ISP Uplinks, Used for Load-Balancing



   This scenario has the following key characteristics:

   o  The enterprise network is using uplinks to two (or more) ISPs for
      Internet access;

   o  Each ISP assigns an IPv6 PA address space;

   o  All the uplinks may be used simultaneously, with the traffic flows
      being randomly (not necessarily equally) distributed between them;

   o  Hosts in the enterprise network are not expected to support Rule
      5.5 of the default address selection algorithm [RFC6724].

3.  Conditional Router Advertisements



3.1.  Solution Overview



3.1.1.  Uplink Selection



   As discussed in [PROVIDER-ASSIGNED], one of the two main problems to
   be solved in the enterprise multihoming scenario is the problem of
   the next-hop (uplink) selection based on the packet source address.
   For example, if the enterprise network has two uplinks, to ISP_A and
   ISP_B, and hosts have addresses from subnet_A and subnet_B (belonging
   to ISP_A and ISP_B, respectively), then packets sourced from subnet_A
   must be sent to the ISP_A uplink while packets sourced from subnet_B
   must be sent to the ISP_B uplink.  Sending packets with source
   addresses belonging to one ISP address space to another ISP might
   cause those packets to be filtered out if those ISPs or their uplinks
   implement antispoofing ingress filtering [RFC2827][RFC3704].

   While some work is being done in the Source Address Dependent Routing
   (SADR) (such as [DESTINATION]), the simplest way to implement the
   desired functionality currently is to apply a policy that selects a
   next hop or an egress interface based on the packet source address.
   Currently, most SMB/Enterprise-grade routers have such functionality
   available.

3.1.2.  Source Address Selection and Conditional RAs



   Another problem to be solved in the multihoming scenario is the
   source address selection on hosts.  In the normal situation (all
   uplinks are up/operational), hosts have multiple global unique
   addresses and can rely on the default address selection algorithm
   [RFC6724] to pick up a source address, while the network is
   responsible for choosing the correct uplink based on the source



Linkova & Stucchi             Informational                     [Page 5]

RFC 8475                     Conditional RAs                October 2018


   address selected by a host, as described in Section 3.1.1.  However,
   some network topology changes (i.e., changing uplink status) might
   affect the global reachability for packets sourced from particular
   prefixes; therefore, such changes have to be signaled back to the
   hosts.  For example:

   o  An uplink to ISP_A went down.  Hosts should not use addresses from
      an ISP_A prefix;

   o  A primary uplink to ISP_A that was not operational has come back
      up.  Hosts should start using the source addresses from an ISP_A
      prefix.

   [PROVIDER-ASSIGNED] provides a detailed explanation of why Stateless
   Address Autoconfiguration (SLAAC) [RFC4862] and Router Advertisements
   (RAs) [RFC4861] are the most suitable mechanisms for signaling
   network topology changes to hosts, thereby influencing the source
   address selection.  Sending an RA to change the preferred lifetime
   for a given prefix provides the following functionality:

   o  Deprecating addresses by sending an RA with preferred_lifetime set
      to 0 in the corresponding Prefix Information option (PIO)
      [RFC4861].  This indicates to hosts that addresses from that
      prefix should not be used;

   o  Making a previously unused (deprecated) prefix usable again by
      sending an RA containing a PIO with nonzero preferred lifetime.
      This indicates to hosts that addresses from that prefix can be
      used again.

   It should be noted that only the preferred lifetime for the affected
   prefix needs to be changed.  As the goal is to influence the source
   address selection algorithm on hosts rather than prevent them from
   forming addresses from a specific prefix, the valid lifetime should
   not be changed.  Actually, changing the valid lifetime would not even
   be possible for unauthenticated RAs (which is the most common
   deployment scenario), because Section 5.5.3 of [RFC4862] prevents
   hosts from setting the valid lifetime for addresses to zero unless
   RAs are authenticated.

   To provide the desired functionality, first-hop routers are required
   to:

   o  Send RAs triggered by defined event policies in response to an
      uplink status change event; and






Linkova & Stucchi             Informational                     [Page 6]

RFC 8475                     Conditional RAs                October 2018


   o  While sending periodic or solicited RAs, set the value in the
      given RA field (e.g., PIO preferred lifetime) based on the uplink
      status.

   The exact definition of the "uplink status" depends on the network
   topology and may include conditions like:

   o  Uplink interface status change;

   o  Presence of a particular route in the routing table;

   o  Presence of a particular route with a particular attribute (next
      hop, tag, etc.) in the routing table;

   o  Protocol adjacency change.

   In some scenarios, when two routers are providing first-hop
   redundancy via Virtual Router Redundancy Protocol (VRRP) [RFC5798],
   the master-backup status can be considered to be a condition for
   sending RAs and changing the preferred lifetime value.  See
   Section 3.2.2 for more details.

   If hosts are provided with the IPv6 addresses of ISP DNS servers via
   a Recursive DNS Server (RDNSS) (see "IPv6 Router Advertisement
   Options for DNS Configuration" [RFC8106]), it might be desirable for
   the conditional RAs to update the Lifetime field of the RDNSS option
   as well.

   The trigger is not only forcing the router to send an unsolicited RA
   to propagate the topology changes to all hosts.  Obviously, the
   values of the RA fields (like PIO Preferred Lifetime or DNS Server
   Lifetime) changed by the particular trigger need to stay the same
   until another event causes the value to be updated.  For example, if
   an ISP_A uplink failure causes the prefix to be deprecated, all
   solicited and unsolicited RAs sent by the router need to have the
   preferred lifetime for that PIO set to 0 until the uplink comes back
   up.

   It should be noted that the proposed solution is quite similar to the
   existing requirement L-13 for IPv6 Customer Edge Routers [RFC7084]
   and the documented behavior of homenet devices [RFC7788].  It is
   using the same mechanism of deprecating a prefix when the
   corresponding uplink is not operational, applying it to an
   enterprise-network scenario.







Linkova & Stucchi             Informational                     [Page 7]

RFC 8475                     Conditional RAs                October 2018


3.2.  Example Scenarios



   This section illustrates how the conditional RAs solution can be
   applied to the most common enterprise multihoming scenarios,
   described in Section 2.

3.2.1.  Single Router, Primary/Backup Uplinks



                                                              --------
                                             ,-------,       /        \
                   +----+ 2001:db8:1::/48  ,'         ',    :          :
                   |    |-----------------+    ISP_A    +--+:          :
 2001:db8:1:1::/64 |    |                  ',         ,'    :          :
                   |    |                    '-------'      :          :
H1-----------------| R1 |                                   : INTERNET :
                   |    |                    ,-------,      :          :
 2001:db8:2:1::/64 |    | 2001:db8:2::/48  ,'         ',    :          :
                   |    |-----------------+    ISP_B    +--+:          :
                   +----+                  ',         ,'    :          :
                                             '-------'       \        /
                                                              --------

              Figure 1: Single Router, Primary/Backup Uplinks

   Let's look at a simple network topology where a single router acts as
   a border router to terminate two ISP uplinks and as a first-hop
   router for hosts.  Each ISP assigns a /48 to the network, and the
   ISP_A uplink is a primary one, to be used for all Internet traffic,
   while the ISP_B uplink is a backup, to be used only when the primary
   uplink is not operational.

   To ensure that packets with source addresses from ISP_A and ISP_B are
   only routed to ISP_A and ISP_B uplinks, respectively, the network
   administrator needs to configure a policy on R1:

   IF (packet_source_address is in 2001:db8:1::/48)
       and
       (packet_destination_address is not in
       (2001:db8:1::/48 or 2001:db8:2::/48))
       THEN
           default next hop is ISP_A_uplink










Linkova & Stucchi             Informational                     [Page 8]

RFC 8475                     Conditional RAs                October 2018


   IF (packet_source_address is in 2001:db8:2::/48)
       and
       (packet_destination_address is not in
       (2001:db8:1::/48 or 2001:db8:2::/48))
       THEN
           default next hop is ISP_B_uplink

   Under normal circumstances, it is desirable that all traffic be sent
   via the ISP_A uplink; therefore, hosts (the host H1 in the example
   topology figure) should be using source addresses from
   2001:db8:1:1::/64.  When or if the ISP_A uplink fails, hosts should
   stop using the 2001:db8:1:1::/64 prefix and start using
   2001:db8:2:1::/64 until the ISP_A uplink comes back up.  To achieve
   this, the RA configuration on the R1 device for the interface facing
   H1 needs to have the following policy:

   prefix 2001:db8:1:1::/64 {
       IF (ISP_A_uplink is up)
           THEN
               preferred_lifetime = 604800
           ELSE
               preferred_lifetime = 0
   }

   prefix 2001:db8:2:1::/64 {
       IF (ISP_A_Uplink is up)
           THEN
               preferred_lifetime = 0
           ELSE
               preferred_lifetime = 604800
   }

   A similar policy needs to be applied to the RDNSS lifetime if ISP_A
   and ISP_B DNS servers are used.

3.2.2.  Two Routers, Primary/Backup Uplinks



   Let's look at a more complex scenario where two border routers are
   terminating two ISP uplinks (one each), acting as redundant first-hop
   routers for hosts.  The topology is shown in Figure 2.











Linkova & Stucchi             Informational                     [Page 9]

RFC 8475                     Conditional RAs                October 2018


                                                              --------
                                             ,-------,       /        \
  2001:db8:1:1::/64 +----+ 2001:db8:1::/48 ,'         ',    :          :
                   _|    |----------------+    ISP_A    +--+:          :
                  | | R1 |                 ',         ,'    :          :
                  | +----+                   '-------'      :          :
H1----------------|                                         : INTERNET :
                  | +----+                   ,-------,      :          :
                  |_|    | 2001:db8:2::/48 ,'         ',    :          :
                    | R2 |----------------+    ISP_B    +--+:          :
 2001:db8:2:1::/64  +----+                 ',         ,'    :          :
                                             '-------'       \        /
                                                              --------

               Figure 2: Two Routers, Primary/Backup Uplinks

   In this scenario, R1 sends RAs with PIO for 2001:db8:1:1::/64 (ISP_A
   address space), and R2 sends RAs with PIO for 2001:db8:2:1::/64
   (ISP_B address space).  Each router needs to have a forwarding policy
   configured for packets received on its hosts-facing interface:

   IF (packet_source_address is in 2001:db8:1::/48)
       and
       (packet_destination_address is not in
       (2001:db8:1::/48 or 2001:db8:2::/48))
       THEN
           default next hop is ISP_A_uplink

   IF (packet_source_address is in 2001:db8:2::/48)
       and
       (packet_destination_address is not in
       (2001:db8:1::/48 or 2001:db8:2::/48))
       THEN
           default next hop is ISP_B_uplink

   In this case, there is more than one way to ensure that hosts are
   selecting the correct source address based on the uplink status.  If
   VRRP is used to provide first-hop redundancy, and the master router
   is the one with the active uplink, then the simplest way is to use
   the VRRP mastership as a condition for RA.  So, if ISP_A is the
   primary uplink, the routers R1 and R2 need to be configured in the
   following way:

   R1 is the VRRP master by default (when the ISP_A uplink is up).  If
   the ISP_A uplink is down, then R1 becomes a backup (the VRRP
   interface-status tracking is expected to be used to automatically





Linkova & Stucchi             Informational                    [Page 10]

RFC 8475                     Conditional RAs                October 2018


   modify the VRRP priorities and trigger the mastership switchover).
   RAs on R1's interface facing H1 needs to have the following policy
   applied:

   prefix 2001:db8:1:1::/64 {
       IF (vrrp_master)
           THEN
               preferred_lifetime = 604800
           ELSE
               preferred_lifetime = 0
   }

   R2 is VRRP backup by default.  RA on R2's interface facing H1 needs
   to have the following policy applied:

   prefix 2001:db8:2:1::/64 {
       IF(vrrp_master)
           THEN
               preferred_lifetime = 604800
           ELSE
               preferred_lifetime = 0
   }

   If VRRP is not used or interface status tracking is not used for
   mastership switchover, then each router needs to be able to detect
   the uplink failure/recovery on the neighboring router, so that RAs
   with updated preferred lifetime values are triggered.  Depending on
   the network setup, various triggers can be used, such as a route to
   the uplink interface subnet or a default route received from the
   uplink.  The obvious drawback of using the routing table to trigger
   the conditional RAs is that some additional configuration is
   required.  For example, if a route to the prefix assigned to the ISP
   uplink is used as a trigger, then the conditional RA policy would
   have the following logic:

   R1:

   prefix 2001:db8:1:1::/64 {
       IF (ISP_A_uplink is up)
           THEN
               preferred_lifetime = 604800
           ELSE
              preferred_lifetime = 0
   }







Linkova & Stucchi             Informational                    [Page 11]

RFC 8475                     Conditional RAs                October 2018


   R2:

   prefix 2001:db8:2:1::/64 {
       IF (ISP_A_uplink_route is present)
           THEN
               preferred_lifetime = 0
           ELSE
               preferred_lifetime = 604800
   }

3.2.3.  Single Router, Load-Balancing between Uplinks



   Let's look at the example topology shown in Figure 1, but with both
   uplinks used simultaneously.  In this case, R1 would send RAs
   containing PIOs for both prefixes, 2001:db8:1:1::/64 and
   2001:db8:2:1::/64, changing the preferred lifetime based on
   particular uplink availability.  If the interface status is used as
   an uplink availability indicator, then the policy logic would look
   like the following:

   prefix 2001:db8:1:1::/64 {
       IF (ISP_A_uplink is up)
           THEN
               preferred_lifetime  = 604800
           ELSE
               preferred_lifetime = 0
   }
   prefix 2001:db8:2:1::/64 {
       IF (ISP_B_uplink is up)
           THEN
               preferred_lifetime  = 604800
           ELSE
               preferred_lifetime = 0
   }

   R1 needs a forwarding policy to be applied to forward packets to the
   correct uplink based on the source address, similar to the policy
   described in Section 3.2.1.

3.2.4.  Two Routers, Load-Balancing between Uplinks



   In this scenario, the example topology is similar to the one shown in
   Figure 2, but both uplinks can be used at the same time.  This means
   that both R1 and R2 need to have the corresponding forwarding policy
   to forward packets based on their source addresses.






Linkova & Stucchi             Informational                    [Page 12]

RFC 8475                     Conditional RAs                October 2018


   Each router would send RAs with PIO for the corresponding prefix,
   setting preferred_lifetime to a nonzero value when the ISP uplink is
   up and deprecating the prefix by setting preferred_lifetime to 0 in
   the case of uplink failure.  The uplink recovery would trigger
   another RA with a nonzero preferred lifetime to make the addresses
   from the prefix preferred again.  The example RA policy on R1 and R2
   would look like:

   R1:
   prefix 2001:db8:1:1::/64 {
       IF (ISP_A_uplink is up)
           THEN
               preferred_lifetime  = 604800
           ELSE
               preferred_lifetime = 0
   }

   R2:

   prefix 2001:db8:2:1::/64 {
       IF (ISP_B_uplink is up)
           THEN
               preferred_lifetime  = 604800
           ELSE
               preferred_lifetime = 0
   }

3.2.5.  Topologies with Dedicated Border Routers



   For simplicity, all topologies above show the ISP uplinks terminated
   on the first-hop routers.  Obviously, the proposed approach can be
   used in more complex topologies when dedicated devices are used for
   terminating ISP uplinks.  In that case, VRRP mastership or interface
   status cannot be used as a trigger for conditional RAs.  Route
   presence as described in Section 3.2.2 should be used instead.
















Linkova & Stucchi             Informational                    [Page 13]

RFC 8475                     Conditional RAs                October 2018


   Let's look at the example topology shown in Figure 3:

                                2001:db8:1::/48              --------
    2001:db8:1:1::/64                     ,-------,        ,'        ',
              +----+  +---+  +----+     ,'         ',     :            :
             _|    |--|   |--| R3 |----+    ISP_A    +---+:            :
            | | R1 |  |   |  +----+     ',         ,'     :            :
            | +----+  |   |               '-------'       :            :
  H1--------|         |LAN|                               :  INTERNET  :
            | +----+  |   |               ,-------,       :            :
            |_|    |  |   |  +----+     ,'         ',     :            :
              | R2 |--|   |--| R4 |----+    ISP_B    +---+:            :
              +----+  +---+  +----+     ',         ,'     :            :
  2001:db8:2:1::/64                       '-------'        ',        ,'
                                2001:db8:2::/48              --------

                    Figure 3: Dedicated Border Routers

   For example, if ISP_A is a primary uplink and ISP_B is a backup, then
   the following policy might be used to achieve the desired behavior
   (H1 is using ISP_A address space, 2001:db8:1:1::/64, while the ISP_A
   uplink is up and only using the ISP_B 2001:db8:2:1::/64 prefix if the
   uplink is non-operational):

   R1 and R2 policy:

   prefix 2001:db8:1:1::/64 {
       IF (ISP_A_uplink_route is present)
           THEN
               preferred_lifetime = 604800
           ELSE
               preferred_lifetime = 0
   }

   prefix 2001:db8:2:1::/64 {
       IF (ISP_A_uplink_route is present)
           THEN
               preferred_lifetime = 0
           ELSE
               preferred_lifetime = 604800
   }










Linkova & Stucchi             Informational                    [Page 14]

RFC 8475                     Conditional RAs                October 2018


   For the load-balancing case, the policy would look slightly
   different: each prefix has a nonzero preferred_lifetime only if the
   corresponding ISP uplink route is present:

   prefix 2001:db8:1:1::/64 {
       IF (ISP_A_uplink_route is present)
           THEN
               preferred_lifetime = 604800
           ELSE
               preferred_lifetime = 0
   }

   prefix 2001:db8:2:1::/64 {
       IF (ISP_B_uplink_route is present)
           THEN
               preferred_lifetime = 604800
           ELSE
               preferred_lifetime = 0
   }

3.2.6.  Intrasite Communication during Simultaneous Uplinks Outage



   Prefix deprecation as a result of an uplink status change might lead
   to a situation in which all global prefixes are deprecated (all ISP
   uplinks are not operational for some reason).  Even when there is no
   Internet connectivity, it might be still desirable to have intrasite
   IPv6 connectivity (especially when the network in question is an
   IPv6-only one).  However, while an address is in a deprecated state,
   its use is discouraged, but not strictly forbidden [RFC4862].  In
   such a scenario, all IPv6 source addresses in the candidate set
   [RFC6724] are deprecated, which means that they still can be used (as
   there are no preferred addresses available), and the source address
   selection algorithm can pick up one of them, allowing intrasite
   communication.  However, some operating systems might just fall back
   to IPv4 if the network interface has no preferred IPv6 global
   addresses.  Therefore, if intrasite connectivity is vital during
   simultaneous outages of multiple uplinks, administrators might
   consider using Unique Local Addresses (ULAs) [RFC4193] or
   provisioning additional backup uplinks to protect the network from
   double-failure cases.

3.2.7.  Uplink Damping



   If an actively used uplink (a primary one or one used in a load-
   balancing scenario) starts flapping, it might lead to the undesirable
   situation of flapping addresses on hosts: every time the uplink goes
   up, hosts receive an RA with a nonzero preferred PIO lifetime, and
   every time the uplink goes down, all addresses in the affected prefix



Linkova & Stucchi             Informational                    [Page 15]

RFC 8475                     Conditional RAs                October 2018


   become deprecated.  This would, undoubtedly, negatively impact the
   user experience, not to mention the impact of spikes of duplicate
   address detection traffic every time an uplink comes back up.
   Therefore, it's recommended that router vendors implement some form
   of damping policy for conditional RAs and either postpone sending an
   RA with a nonzero lifetime for a PIO when the uplink comes up for a
   number of seconds or (even) introduce accumulated penalties/
   exponential backoff algorithm for such delays.  (In the case of
   multiple simultaneous uplink failure, when all but one of the uplinks
   are down and the last remaining one is flapping, it might result in
   all addresses being deprecated for a while after the flapping uplink
   recovers.)

3.2.8.  Routing Packets When the Corresponding Uplink Is Unavailable



   Deprecating IPv6 addresses by setting the preferred lifetime to 0
   discourages but does not strictly forbid its usage in new
   communications.  A deprecated address may still be used for existing
   connections [RFC4862].  Therefore, when an ISP uplink goes down, the
   corresponding border router might still receive packets with source
   addresses belonging to that ISP address space while there is no
   available uplink to send those packets to.

   The expected router behavior would depend on the uplink selection
   mechanism.  For example, if some form of SADR is used, then such
   packets will be dropped as there is no route to the destination.  If
   policy-based routing is used to set a next hop, then the behavior
   would be implementation dependent and may vary from dropping the
   packets to forwarding them based on the routing table entries.  It
   should be noted that there is no return path to the packet source (as
   the ISP uplink is not operational).  Therefore, even if the outgoing
   packets are sent to another ISP, the return traffic might not be
   delivered.

3.3.  Solution Limitations



   It should be noted that the proposed approach is not a "silver
   bullet" for all possible multihoming scenarios.  It would work very
   well for networks with relatively simple topologies and
   straightforward routing policies.  The more complex the network
   topology and the corresponding routing policies, the more
   configuration would be required to implement the solution.

   Another limitation is related to the load-balancing between the
   uplinks.  In the scenario in which both uplinks are active, hosts
   would select the source prefix using the Default Address Selection
   algorithm [RFC6724]; therefore, the load between two uplinks most
   likely would not be evenly distributed.  (However, the proposed



Linkova & Stucchi             Informational                    [Page 16]

RFC 8475                     Conditional RAs                October 2018


   mechanism does allow a creative way of controlling uplinks load in
   software-defined networks where controllers might selectively
   deprecate prefixes on some hosts but not others to move egress
   traffic between uplinks).  Also, the prefix selection does not take
   into account any other properties of uplinks (such as latency), so
   egress traffic might not be sent to the nearest uplink if the
   corresponding prefix is selected as a source.  In general, if not all
   uplinks are equal, and some uplinks are expected to be preferred over
   others, then the network administrator should ensure that prefixes
   from non-preferred ISP(s) are kept deprecated (so primary/backup
   setup is used).

3.3.1.  Connections Preservation



   The proposed solution is not designed to preserve connection state
   after an uplink failure.  If all uplinks to an ISP go down, all
   sessions to/from addresses from that ISP address space are
   interrupted as there is no egress path for those packets and there is
   no return path from the Internet to the corresponding prefix.  In
   this regard, it is similar to IPv4 multihoming using NAT, where an
   uplink failure and failover to another uplink means that a public
   IPv4 address changes and all existing connections are interrupted.

   However, an uplink recovery does not necessarily lead to connections
   interruption.  In the load-sharing/balancing scenario, an uplink
   recovery does not affect any existing connections at all.  In the
   active/backup topology, when the primary uplink recovers from the
   failure and the backup prefix is deprecated, the existing sessions
   (established to/from the backup ISP addresses) can be preserved if
   the routers are configured as described in Section 3.2.1 and send
   packets with the backup ISP source addresses to the backup uplink,
   even when the primary one is operational.  As a result, the primary
   uplink recovery makes the usage of the backup ISP addresses
   discouraged but still possible.

   It should be noted that in IPv4 multihoming with NAT, when the egress
   interface is chosen without taking packet source address into account
   (as internal hosts usually have addresses from [RFC1918] space),
   sessions might not be preserved after an uplink recovery unless
   packet forwarding is integrated with existing NAT sessions tracking.

4.  IANA Considerations



   This document has no IANA actions.







Linkova & Stucchi             Informational                    [Page 17]

RFC 8475                     Conditional RAs                October 2018


5.  Security Considerations



   This memo introduces no new security considerations.  It relies on
   RAs [RFC4861] and the SLAAC [RFC4862] mechanism and inherits their
   security properties.  If an attacker is able to send a rogue RA, they
   could deprecate IPv6 addresses on hosts or influence source-address-
   selection processes on hosts.

   The potential attack vectors include, but are not limited to:

   o  An attacker sends a rogue RA deprecating IPv6 addresses on hosts;

   o  An attacker sends a rogue RA making addresses preferred while the
      corresponding ISP uplink is not operational;

   o  An attacker sends a rogue RA making addresses preferred for a
      backup ISP, steering traffic to an undesirable (e.g., more
      expensive) uplink.

   Therefore, the network administrators SHOULD secure RAs, e.g., by
   deploying an RA guard [RFC6105].

5.1.  Privacy Considerations



   This memo introduces no new privacy considerations.

6.  References



6.1.  Normative References



   [RFC1918]  Rekhter, Y., Moskowitz, B., Karrenberg, D., de Groot, G.,
              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>.

   [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>.

   [RFC2827]  Ferguson, P. and D. Senie, "Network Ingress Filtering:
              Defeating Denial of Service Attacks which employ IP Source
              Address Spoofing", BCP 38, RFC 2827, DOI 10.17487/RFC2827,
              May 2000, <https://www.rfc-editor.org/info/rfc2827>.







Linkova & Stucchi             Informational                    [Page 18]

RFC 8475                     Conditional RAs                October 2018


   [RFC3022]  Srisuresh, P. and K. Egevang, "Traditional IP Network
              Address Translator (Traditional NAT)", RFC 3022,
              DOI 10.17487/RFC3022, January 2001,
              <https://www.rfc-editor.org/info/rfc3022>.

   [RFC3704]  Baker, F. and P. Savola, "Ingress Filtering for Multihomed
              Networks", BCP 84, RFC 3704, DOI 10.17487/RFC3704, March
              2004, <https://www.rfc-editor.org/info/rfc3704>.

   [RFC4116]  Abley, J., Lindqvist, K., Davies, E., Black, B., and V.
              Gill, "IPv4 Multihoming Practices and Limitations",
              RFC 4116, DOI 10.17487/RFC4116, July 2005,
              <https://www.rfc-editor.org/info/rfc4116>.

   [RFC4193]  Hinden, R. and B. Haberman, "Unique Local IPv6 Unicast
              Addresses", RFC 4193, DOI 10.17487/RFC4193, October 2005,
              <https://www.rfc-editor.org/info/rfc4193>.

   [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>.

   [RFC4862]  Thomson, S., Narten, T., and T. Jinmei, "IPv6 Stateless
              Address Autoconfiguration", RFC 4862,
              DOI 10.17487/RFC4862, September 2007,
              <https://www.rfc-editor.org/info/rfc4862>.

   [RFC6105]  Levy-Abegnoli, E., Van de Velde, G., Popoviciu, C., and J.
              Mohacsi, "IPv6 Router Advertisement Guard", RFC 6105,
              DOI 10.17487/RFC6105, February 2011,
              <https://www.rfc-editor.org/info/rfc6105>.

   [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>.

   [RFC8028]  Baker, F. and B. Carpenter, "First-Hop Router Selection by
              Hosts in a Multi-Prefix Network", RFC 8028,
              DOI 10.17487/RFC8028, November 2016,
              <https://www.rfc-editor.org/info/rfc8028>.

   [RFC8106]  Jeong, J., Park, S., Beloeil, L., and S. Madanapalli,
              "IPv6 Router Advertisement Options for DNS Configuration",
              RFC 8106, DOI 10.17487/RFC8106, March 2017,
              <https://www.rfc-editor.org/info/rfc8106>.





Linkova & Stucchi             Informational                    [Page 19]

RFC 8475                     Conditional RAs                October 2018


   [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>.

6.2.  Informative References



   [DESTINATION]
              Lamparter, D. and A. Smirnov, "Destination/Source
              Routing", Work in Progress,
              draft-ietf-rtgwg-dst-src-routing-06, October 2017.

   [PROVIDER-ASSIGNED]
              Baker, F., Bowers, C., and J. Linkova, "Enterprise
              Multihoming using Provider-Assigned Addresses without
              Network Prefix Translation: Requirements and Solution",
              Work in Progress,
              draft-ietf-rtgwg-enterprise-pa-multihoming-07, June 2018.

   [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>.

   [RFC5798]  Nadas, S., Ed., "Virtual Router Redundancy Protocol (VRRP)
              Version 3 for IPv4 and IPv6", RFC 5798,
              DOI 10.17487/RFC5798, March 2010,
              <https://www.rfc-editor.org/info/rfc5798>.

   [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>.

   [RFC7788]  Stenberg, M., Barth, S., and P. Pfister, "Home Networking
              Control Protocol", RFC 7788, DOI 10.17487/RFC7788, April
              2016, <https://www.rfc-editor.org/info/rfc7788>.

Acknowledgements



   Thanks to the following people (in alphabetical order) for their
   review and feedback: Mikael Abrahamsson, Lorenzo Colitti, Marcus
   Keane, Erik Kline, David Lamparter, Dusan Mudric, Erik Nordmark, and
   Dave Thaler.








Linkova & Stucchi             Informational                    [Page 20]

RFC 8475                     Conditional RAs                October 2018


Authors' Addresses



   Jen Linkova
   Google
   Mountain View, California  94043
   United States of America

   Email: furry@google.com


   Massimiliano Stucchi
   RIPE NCC
   Stationsplein, 11
   Amsterdam  1012 AB
   The Netherlands

   Email: mstucchi@ripe.net


































Linkova & Stucchi             Informational                    [Page 21]