Transmission of IPv6 Packets over AERO Links

The following sections specify the operation of IPv6 over Asymmetric Extended Route Optimization (AERO) links:

AERO Relays relay packets between nodes connected to the same AERO link and also forward packets between the AERO link and the native IPv6 network. The relaying process entails re-encapsulation of IPv6 packets that were received from a first AERO node and are to be forwarded without modification to a second AERO node. AERO Servers configure their AERO interfaces as router interfaces, and provide default routing services to AERO Clients. AERO Servers configure a DHCPv6 relay or server function and facilitate DHCPv6 Prefix Delegation (PD) exchanges. An AERO Server may also act as an AERO Relay. AERO Clients act as requesting routers to receive IPv6 prefixes through a DHCPv6 PD exchange via an AERO Server over the AERO link. (In Server-less environments an alternate AERO link prefix delegation authority may be necessary, but out of scope for this document.) Each AERO Client receives at least a /64 prefix delegation, and may receive even shorter prefixes. AERO Clients that act as routers configure their AERO interfaces as router interfaces and sub-delegate portions of their received prefix delegations to links on EUNs. AERO Clients that act as ordinary hosts configure their AERO interfaces as host interfaces and assign one or more IPv6 addresses taken from their received prefix delegations to the AERO interface but DO NOT assign the delegated prefix itself to the AERO interface. Instead, the host assigns the delegated prefix to a "black hole" route so that unused portions of the prefix are nullified. End system applications on AERO hosts bind directly to the AERO interface, while applications on AERO routers (or IPv6 hosts served by an AERO router) bind to EUN interfaces.

An AERO address is an IPv6 link-local address assigned to an AERO interface and with an IPv6 prefix embedded within the interface identifier. The AERO address is formatted as: fe80::[IPv6 prefix] Each AERO Client configures an AERO address based on the prefix it has received from the AERO link prefix delegation authority (e.g., the DHCPv6 server). The address begins with the prefix fe80::/64 and includes in its interface identifier the base /64 prefix taken from the Client's delegated IPv6 prefix. The base prefix is determined by masking the delegated prefix with the prefix length. For example, if an AERO Client has received the prefix delegation: 2001:db8:1000:2000::/56 it would construct its AERO address as: fe80::2001:db8:1000:2000 The AERO address remains stable as the Client moves between topological locations, i.e., even if its underlying address changes.

AERO interfaces use IPv6-in-IPv6 encapsulation to exchange tunneled packets with AERO neighbors attached to an underlying IPv6 network, and use IPv6-in-IPv4 encapsulation to exchange tunneled packets with AERO neighbors attached to an underlying IPv4 network. AERO interfaces can also use secured tunnel types such as IPsec or TLS in environments where strong authentication and confidentiality are required. When NAT traversal and/or filtering middlebox traversal is necessary, a UDP header is further inserted immediately above the outer IP encapsulation header. Servers assign the link-local address 'fe80::0' to their AERO interface; this provides a handle for Clients to insert into a neighbor cache entry for their current Server. Servers and Relays also configure administratively-assigned link-local addresses on their AERO interfaces to support the operation of the inter-Server/Relay routing system (see: ). Clients initially use a "temporary" IPv6 link-local address in the DHCPv6 PD exchanges used to receive an IPv6 prefix and derive an AERO address. If the Client is pre-provisioned with an IPv6 prefix associated with the AERO service, it SHOULD use the AERO address derived from the prefix as the temporary address. Otherwise, the Client SHOULD use "fe80::1" as the temporary address since this address will not conflict with any valid AERO addresses and will thus not be used in any AERO neighbor discovery messaging. After the Client receives a prefix delegation, it assigns the corresponding AERO address to the AERO interface. DHCPv6 is therefore used to bootstrap the assignment of link-local addresses on the AERO link. AERO interfaces maintain a neighbor cache and use an augmentation of standard unicast IPv6 ND messaging. AERO interfaces use Redirect, Neighbor Solicitation (NS) and Neighbor Advertisement (NA) Router Solicitation (RS) and Router Advertisement (RA) messages the same as for any IPv6 link. Finally, AERO links use link-local-only addressing; hence, Clients MUST ignore any Prefix Information Options (PIOs) they may receive in RA messages. AERO Redirect messages include a TLLAO the same as for any IPv6 link. The TLLAO includes the link-layer address for the target node, which is formed from the concatenation of the 2-octet UDP port number used by the target when it sends UDP-encapsulated packets over the AERO interface (or 0 when the target does not use UDP encapsulation) followed by the 16-octet IP address. The TLLAO format is shown in :

(Note that in the above TLLAO format, the IP address is formed as an IPv4-compatible IPv6 address (see: ) when the encapsulation IP address family is IPv4. Note also that more than one TLLAO option may appear in a Redirect message, e.g., if the target node has multiple link-layer addresses.) AERO NS, NA, RS and RA messages do not include Source/Target Link Layer Address Options (S/TLLAOs). Instead, AERO nodes determine the link-layer addresses of neighbors by examining the link-layer source address of any NS/NA/RS/RA messages they receive and ignore any S/TLLAOs included in these messages. This is vital to the operation of AERO links for which neighbors are separated by Network Address Translators (NATs) (either IPv4 or IPv6) since the source may have no way of knowing what its translated address will be and hence may not be able to supply the correct values in a S/TLLAO. Finally, AERO interface NS/NA messages only update existing neighbor cache entires and do not create new neighbor cache entries, whereas Redirect, RS and RA messages both update and create neighbor cache entries. This represents a departure from the normal operation of IPv6 ND over common link types, but is consistent with the spirit of IPv6 over NBMA links as discussed in . Note however that this restriction may be relaxed and/or redefined on AERO links that participate in a fully distributed mobility management model (i.e., a "Client-only" AERO link) coordinated in a manner outside the scope of this document.

Nodes on AERO interfaces use a simple data origin authentication for encapsulated packets they receive from other nodes. In particular, AERO Clients accept encapsulated packets with a link-layer source address belonging to their current AERO Server. AERO nodes also accept encapsulated packets with a link-layer source address that is correct for the network-layer source address. The AERO node considers the link-layer source address correct for the network-layer source address if there is an IPv6 forwarding table entry that matches the network-layer source address as well as a neighbor cache entry corresponding to the next hop that includes the link-layer address. An exception is that neighbor discovery messages may include a different link-layer address than the one currently in the neighbor cache, and the new link-layer address updates the neighbor cache entry.

Each AERO node maintains a per-AERO interface conceptual neighbor cache that includes an entry for each neighbor it communicates with on the AERO link, the same as for any IPv6 interface (see ). Neighbor cache entries are created and maintained as follows: When an AERO Server relays a DHCPv6 Reply message to an AERO Client, it creates or updates a neighbor cache entry for the Client based on the information in the IA-PD option. When an AERO node receives a valid RS/RA message, it creates or updates a neighbor cache entry the same as described in . When an AERO Client receives a valid Predirect message (See Section 3.10.5) it creates or updates a neighbor cache entry for the Predirect target L3 and L2 addresses, and also creates an IPv6 forwarding table entry for the Predirected (source) prefix. The node then sets an ACCEPT timer and uses this timer to validate any messages received from the Predirected neighbor. When an AERO Client receives a valid Redirect message (see Section 3.10.7) it creates or updates a dynamic neighbor cache entry for the Redirect target L3 and L2 addresses, and also creates an IPv6 forwarding table entry for the Redirected (destination) prefix. The node then sets a FORWARD timer and uses this timer to determine whether packets can be sent directly to the Redirected neighbor. The node also maintains a constant value MAX_RETRY to limit the number of keepalives sent when a neighbor has gone unreachable. It is RECOMMENDED that FORWARD_TIME be set to the default constant value 30 seconds to match the default REACHABLE_TIME value specified for IPv6 neighbor discovery . It is RECOMMENDED that ACCEPT_TIME be set to the default constant value 40 seconds to allow a 10 second window so that the AERO redirection procedure can converge before the ACCEPT_TIME timer decrements below FORWARD_TIME. It is RECOMMENDED that MAX_RETRY be set to 3 the same as described for IPv6 neighbor discovery address resolution in Section 7.3.3 of . Different values for FORWARD_TIME, ACCEPT_TIME, and MAX_RETRY MAY be administratively set, if necessary, to better match the AERO link's performance characteristics; however, if different values are chosen, all nodes on the link MUST consistently configure the same values. ACCEPT_TIME SHOULD further be set to a value that is sufficiently longer than FORWARD_TIME to allow the AERO redirection procedure to converge.

The AERO link Maximum Transmission Unit (MTU) is 64KB minus the encapsulation overhead for IPv4 and 4GB minus the encapsulation overhead for IPv6 . This is the most that IPv4 and IPv6 (respectively) can convey within the constraints of protocol constants, but actual sizes available for tunneling will frequently be much smaller. The base tunneling specifications for IPv4 and IPv6 typically set a static MTU on the tunnel interface to 1500 bytes minus the encapsulation overhead or smaller still if the tunnel is likely to incur additional encapsulations on the path. This can result in path MTU related black holes when packets that are too large to be accommodated over the AERO link are dropped, but the resulting ICMP Packet Too Big (PTB) messages are lost on the return path. As a result, AERO nodes use the following MTU mitigations to accommodate larger packets. AERO nodes set their AERO interface MTU to the larger of the underlying interface MTU minus the encapsulation overhead, and 1500 bytes. AERO nodes optionally cache other per-neighbor MTU values in the underlying IP path MTU discovery cache initialized to the underlying interface MTU. AERO nodes admit packets that are no larger than 1280 bytes minus the encapsulation overhead (*) as well as packets that are larger than 1500 bytes into the tunnel without fragmentation, i.e., as long as they are no larger than the AERO interface MTU before encapsulation and also no larger than the cached per-neighbor MTU following encapsulation. For IPv4, the node sets the "Don't Fragment" (DF) bit to 0 for packets no larger than 1280 bytes minus the encapsulation overhead (*) and sets the DF bit to 1 for packets larger than 1500 bytes. If a large packet is lost in the path, the node may optionally cache the MTU reported in the resulting PTB message or may ignore the message, e.g., if there is a possibility that the message is spurious. For packets destined to an AERO node that are larger than 1280 bytes minus the encapsulation overhead (*) but no larger than 1500 bytes, the node uses outer IP fragmentation to fragment the packet into two pieces (where the first fragment contains 1024 bytes of the fragmented inner packet) then admits the fragments into the tunnel. If the outer protocol is IPv4, the node admits the packet into the tunnel with DF set to 0 and subject to rate limiting to avoid reassembly errors . For both IPv4 and IPv6, the node also sends a 1500 byte probe message (**) to the neighbor, subject to rate limiting. To construct a probe, the node prepares an ICMPv6 Neighbor Solicitation (NS) message with trailing padding octets added to a length of 1500 bytes but does not include the length of the padding in the IPv6 Payload Length field. The node then encapsulates the NS in the outer encapsulation headers (while including the length of the padding in the outer length fields), sets DF to 1 (for IPv4) and sends the padded NS message to the neighbor. If the neighbor returns an NA message, the node may then send whole packets within this size range and (for IPv4) relax the rate limiting requirement. AERO nodes MUST be capable of reassembling packets up to 1500 bytes plus the encapsulation overhead length. It is therefore RECOMMENDED that AERO nodes be capable of reassembling at least 2KB. (*) Note that if it is known without probing that the minimum Path MTU to an AERO node is MINMTU bytes (where 1280 < MINMTU < 1500) then MINMTU can be used instead of 1280 in the fragmentation threshold considerations listed above. (**) It is RECOMMENDED that no probes smaller than 1500 bytes be used for MTU probing purposes, since smaller probes may be fragmented if there is a nested tunnel somewhere on the path to the neighbor. Probe sizes larger than 1500 bytes MAY be used, but may be unnecessary since original sources are expected to implement when sending large packets.

AERO interfaces encapsulate IPv6 packets according to whether they are entering the AERO interface for the first time or if they are being forwarded out the same AERO interface that they arrived on. This latter form of encapsulation is known as "re-encapsulation". AERO interfaces encapsulate packets per the specifications in except that the interface copies the "Hop Limit", "Traffic Class" and "Congestion Experienced" values in the inner IPv6 header into the corresponding fields in the outer IP header. For packets undergoing re-encapsulation, the AERO interface instead copies the "TTL/Hop Limit", "Type of Service/Traffic Class" and "Congestion Experienced" values in the original outer IP header into the corresponding fields in the new outer IP header (i.e., the values are transferred between outer headers and *not* copied from the inner network layer header). When UDP encapsulation is used, the AERO interface inserts a UDP header immediately above the outer IP header. The AERO interface sets the UDP source port to a constant value that it will use in each successive packet it sends, sets the UDP checksum field to zero (see: ) and sets the UDP length field to the length of the inner packet plus 8 bytes for the UDP header itself. For packets sent via a Server, the AERO interface sets the UDP destination port to 8060 (i.e., the IANA-registerd port number for AERO). For packets sent to a neighboring Client, the AERO interface sets the UDP destination port to the port value stored in the neighbor cache entry for this neighbor. The AERO interface next sets the outer IP protocol number to the appropriate value for the first protocol layer within the encapsulation (e.g., IPv6, UDP, IPsec, etc.). When IPv6 is used as the outer IP protocol, the ITE then sets the flow label value in the outer IPv6 header the same as described in . When IPv4 is used as the outer IP protocol, the AERO interface sets the DF bit as discussed in Section 3.6. AERO interfaces decapsulate packets destined either to the node itself or to a destination reached via an interface other than the receiving AERO interface per the specifications in . When the encapsulated packet includes a UDP header, the AERO interface examines the first octet of data following the UDP header. If the most significant four bits of the first octet encode the value '0110', the inner header is an IPv6 header; otherwise, the packet is discarded. During the decapsulation, the AERO interface records the UDP source port in the neighbor cache entry for this neighbor then discards the UDP header. Note that AERO messaging and addressing can also be used in conjunction with other tunnel types such as IPsec and TLS . In that case, the native encapsulation format of the tunnel is used, and the AERO messaging and addressing mechanisms are applied as a layered extension. All other aspects of AERO neighbor coordination are as-specified in this document.

depicts the AERO reference operational scenario. The figure shows an AERO Server('A'), two AERO Clients ('B', 'D') and three ordinary IPv6 hosts ('C', 'E', 'F'):

B; E->D) | +--------------+ fe80::0 L2(A) | X-----+-----------+-----------+--------X | AERO Link | L2(B) L2(D) fe80::2001:db8:0:0 fe80::2001:db8:1:0 .-. +--------------+ +--------------+ ,-( _)-. | AERO Client B| | AERO Client D| .-(_ IPv6 )-. | (default->A) | | (default->A) |--(__ EUN ) +--------------+ +--------------+ `-(______)-' 2001:DB8:0::/48 2001:DB8:1::/48 | | 2001:db8:1::1 .-. +-------------+ ,-( _)-. 2001:db8:0::1 | Host E | .-(_ IPv6 )-. +-------------+ +-------------+ (__ EUN )--| Host C | `-(______)-' +-------------+ ]]> In , AERO Server ('A') connects to the AERO link and connects to the IPv6 Internet, either directly or via an AERO Relay (not shown). Server ('A') assigns the address fe80::0 to its AERO interface with link-layer address L2(A). Server ('A') next arranges to add L2(A) to a published list of valid Servers for the AERO link. AERO Client ('B') registers the IPv6 prefix 2001:db8:0::/48 in a DHCPv6 PD exchange via AERO Server ('A') then assigns the address fe80::2001:db8:0:0 to its AERO interface with link-layer address L2(B). Client ('B') configures a default route via the AERO interface with next-hop address fe80::0 and link-layer address L2(A), then sub-delegates the prefix 2001:db8:0::/48 to its attached EUNs. IPv6 host ('C') connects to the EUN, and configures the address 2001:db8:0::1. AERO Client ('D') registers the IPv6 prefix 2001:db8:1::/48 in a DHCPv6 PD exchange via AERO Server ('A') then assigns the address fe80::2001:db8:1:0 to its AERO interface with link-layer address L2(D). Client ('D') configures a default route via the AERO interface with next-hop address fe80::0 and link-layer address L2(A), then sub-delegates the prefix 2001:db8:1::/48 to its attached EUNs. IPv6 host ('E') connects to the EUN, and configures the address 2001:db8:1::1. Finally, IPv6 host ('F') connects to an IPv6 network outside of the AERO link domain. Host ('F') configures its IPv6 interface in a manner specific to its attached IPv6 link, and assigns the address 2001:db8:2::1 to its IPv6 link interface.

AERO Clients observe the IPv6 node requirements defined in . AERO Clients first discover the link-layer address of an AERO Server via static configuration, or through an automated means such as DNS name resolution. In the absence of other information, the Client resolves the Fully-Qualified Domain Name (FQDN) "linkupnetworks.domainname", where "domainname" is the DNS domain appropriate for the Client's attached underlying network. The Client then creates a static neighbor cache entry with fe80::0 as the network-layer address and the discovered address as the link-layer address then creates a static default IPv6 forwarding table entry with fe80::0 as the next hop address. Next, the Client acts as a requesting router to request an IPv6 prefix through DHCPv6 PD via the AERO Server using a temporary link-local address (see: Section 3.3) as the IPv6 source address and fe80::0 as the IPv6 destination address. The Client includes a DHCPv6 Unique Identifier (DUID) in the Client Identifier option of its DHCPv6 messages and includes any additional authenticating information necessary to authenticate itself to the DHCPv6 server. (Note that other DUID formats such as DUID-UUID may also be used.) If the Client is pre-provisioned with an IPv6 prefix associated with the AERO service, it MAY also include the prefix in an IA-PD option in its DHCPv6 Request command to indicate its preferred prefix to the DHCPv6 server. After the Client receives its prefix delegation, it assigns the link-local AERO address taken from the prefix to the AERO interface (see: Section 3.3) and sub-delegates the prefix to nodes and links within its attached EUNs (the AERO link-local address thereafter remains stable as the Client moves). The Client further renews its prefix delegation via standard DHCPv6 procedures by sending DHCPv6 Renew messages with its AERO address as the IPv6 source address, fe80::0 as the IPv6 destination address and the same DUID value in the Client Identifier option. The Client then sends an RS message to the Server to receive an RA message with a default router lifetime and other configuration information. The Client ignores any Prefix Information Options (PIOs) included in the RA message, since the AERO link is link-local-only. The Client further ignores any RS messages it might receive, since only Servers may process RS messages. The Client can then send periodic NS messages to the Server to obtain new NA messages for Neighbor Unreachability Detection (NUD) and to refresh any network state, and can send periodic RS messages to obtain new RA messages in order to update the default router lifetimes. (The Client can alternately use RS messages for both purposes, but NS/NA exchanges are the standard method for performing NUD.) The Client can also forward IPv6 packets destined to networks beyond its local EUNs via the Server as an IPv6 default router. The Server may in turn return a Redirect message informing the Client of a neighbor on the AERO link that is topologically closer to the final destination as specified in . Note that, since the Client's AERO address is configured from the unique prefix delegation it receives via the Server, there is no need for Duplicate Address Detection (DAD) on AERO links. Other nodes maliciously attempting to hijack an authorized Client's AERO address will be denied due to an unacceptable link-layer address and/or security parameters (see: Security Considerations).

AERO Servers observe the IPv6 router requirements defined in and further configure a DHCPv6 relay function on their AERO links. When the AERO Server relays a Client's DHCPv6 PD messages to the DHCPv6 server, it wraps each message in a "Relay-forward" message per and includes a DHCPv6 Interface Identifier option that encodes a value that identifies the AERO link to the DHCPv6 server. The AERO Server then includes the Client's link-layer address in a Client Link Layer Address Option (CLLAO) with the link-layer address format shown in , i.e., a 2-octet UDP port number followed by a 16-octet IP address. The Server sets the CLLAO 'option-length' field to 20 (2 plus the length of the link-layer address) and sets the 'link-layer type' field to TBD (see: IANA Considerations). The Server finally includes a DHCPv6 Echo Request Option (ERO) that encodes the option code for the CLLAO in a 'requested-option-code-n' field. The CLLAO information will therefore subsequently be echoed back in the DHCPv6 Server's "Relay-reply" message. When the DHCPv6 server issues the IPv6 prefix delegation in a "Relay-reply" message via the AERO Server (acting as a DHCPv6 relay), the AERO Server obtains the Client's link-layer address from the echoed CLLAO option and obtains the Client's delegated prefix from the included IA_PD option. The Server then creates a static neighbor cache entry for the Client's AERO address (see: Section 3.3) with the Client's link-layer address as the link-layer address for the neighbor cache entry. The Server also configures an IPv6 forwarding table entry that lists the Client's AERO address as the next hop toward the delegated IPv6 prefix with a lifetime derived from the DHCPv6 lease lifetime. The AERO Server finally injects the Client's prefix as an IPv6 route into the inter-Server/Relay routing system (see: ) then relays the DHCPv6 message to the Client while using fe80::0 as the IPv6 source address, the link-local address found in the "peer address" field of the Relay-reply message as the IPv6 destination address, and the Client's link-layer address as the destination link-layer address. Servers respond to RS/NS messages from Clients on their AERO interfaces by returning an RA/NA message. When the Server receives an RS/NS message, it updates the neighbor cache entry using the network layer source address as the neighbor's network layer address and using the link-layer source address of the RS/NS message as the neighbor's link-layer address. The Server SHOULD NOT include PIOs in any RA messages it sends to Clients, since the Client will ignore any such options. Servers ignore any RA messages they may receive from a Client. Servers MAY examine RA messages they may receive from other Servers for consistency verification purposes. When the Server forwards a packet via the same AERO interface on which it arrived, it initiates an AERO route optimization procedure as specified in .

describes the AERO reference operational scenario. We now discuss the operation and protocol details of AERO Redirection with respect to this reference scenario.

With reference to , when the IPv6 source host ('C') sends a packet to an IPv6 destination host ('E'), the packet is first forwarded via the EUN to AERO Client ('B'). Client ('B') then forwards the packet over its AERO interface to AERO Server ('A'), which then re-encapsulates and forwards the packet to AERO Client ('D'), where the packet is finally forwarded to the IPv6 destination host ('E'). When Server ('A') re-encapsulates and forwards the packet back out on its advertising AERO interface, it must arrange to redirect Client ('B') toward Client ('D') as a better next-hop node on the AERO link that is closer to the final destination. However, this redirection process applied to AERO interfaces must be more carefully orchestrated than on ordinary links since the parties may be separated by potentially many underlying network routing hops. Consider a first alternative in which Server ('A') informs Client ('B') only and does not inform Client ('D') (i.e., "classical redirection"). In that case, Client ('D') has no way of knowing that Client ('B') is authorized to forward packets from the claimed source address, and it may simply elect to drop the packets. Also, Client ('B') has no way of knowing whether Client ('D') is performing some form of source address filtering that would reject packets arriving from a node other than a trusted default router, nor whether Client ('D') is even reachable via a direct path that does not involve Server ('A'). Consider a second alternative in which Server ('A') informs both Client ('B') and Client ('D') separately, via independent redirection control messages (i.e., "augmented redirection"). In that case, if Client ('B') receives the redirection control message but Client ('D') does not, subsequent packets sent by Client ('B') could be dropped due to filtering since Client ('D') would not have a route to verify the claimed source address. Also, if Client ('D') receives the redirection control message but Client ('B') does not, subsequent packets sent in the reverse direction by Client ('D') would be lost. Since both of these alternatives have shortcomings, a new redirection technique (i.e., "AERO redirection") is needed.

Again, with reference to , when source host ('C') sends a packet to destination host ('E'), the packet is first forwarded over the source host's attached EUN to Client ('B'), which then forwards the packet via its AERO interface to Server ('A'). Server ('A') then re-encapsulates and forwards the packet out the same AERO interface toward Client ('D') and also sends an AERO "Predirect" message forward to Client ('D') as specified in . The Predirect message includes Client ('B')'s network- and link-layer addresses as well as information that Client ('D') can use to determine the IPv6 prefix used by Client ('B') . After Client ('D') receives the Predirect message, it process the message and returns an AERO Redirect message destined for Client ('B') via Server ('A') as specified in . During the process, Client ('D') also creates or updates a dynamic neighbor cache entry for Client ('B'), and creates an IPv6 forwarding table entry for Client ('B')'s IPv6 prefix. When Server ('A') receives the Redirect message, it re-encapsulates the message and forwards it on to Client ('B') as specified in . The message includes Client ('D')'s network- and link-layer addresses as well as information that Client ('B') can use to determine the IPv6 prefix used by Client ('D'). After Client ('B') receives the Redirect message, it processes the message as specified in . During the process, Client ('B') also creates or updates a dynamic neighbor cache entry for Client ('D'), and creates an IPv6 forwarding table entry for Client ('D')'s IPv6 prefix. Following the above Predirect/Redirect message exchange, forwarding of packets from Client ('B') to Client ('D') without involving Server ('A) as an intermediary is enabled. The mechanisms that support this exchange are specified in the following sections.

AERO Redirect/Predirect messages use the same format as for ICMPv6 Redirect messages depicted in Section 4.5 of , but also include a new "Prefix Length" field taken from the low-order 8 bits of the Redirect message Reserved field (valid values for the Prefix Length field are 0 through 64). The Redirect/Predirect messages are formatted as shown in :

When an AERO Server forwards a packet out the same AERO interface that it arrived on, the Server sends a Predirect message forward toward the AERO Client nearest the destination instead of sending a Redirect message back to AERO Client nearest the source. In the reference operational scenario, when Server ('A') forwards a packet sent by Client ('B') toward Client ('D'), it also sends a Predirect message forward toward Client ('D'), subject to rate limiting (see Section 8.2 of ). Server ('A') prepares the Predirect message as follows: the link-layer source address is set to 'L2(A)' (i.e., the underlying address of Server ('A')). the link-layer destination address is set to 'L2(D)' (i.e., the underlying address of Client ('D')). the network-layer source address is set to fe80::0 (i.e., the link-local address of Server ('A')). the network-layer destination address is set to fe80::2001:db8:1:0 (i.e., the AERO address of Client ('D')). the Type is set to 137. the Code is set to 1 to indicate "Predirect". the Prefix Length is set to the length of the prefix to be applied to Target address. the Target Address is set to fe80::2001:db8:0::0 (i.e., the AERO address of Client ('B')). the Destination Address is set to the IPv6 source address of the packet that triggered the Predirection event. the message includes one or more TLLAOs set to 'L2(B)' and any other underlying address(es) of Client ('B'). the message includes a Redirected Header Option (RHO) that contains the originating packet truncated to ensure that at least the network-layer header is included but the size of the message does not exceed 1280 bytes. Server ('A') then sends the message forward to Client ('D').

When Client ('D') receives a Predirect message, it accepts the message only if it has a link-layer source address of the Server, i.e. 'L2(A)'. Client ('D') further accepts the message only if it is willing to serve as a redirection target. Next, Client ('D') validates the message according to the ICMPv6 Redirect message validation rules in Section 8.1 of . In the reference operational scenario, when the Client ('D') receives a valid Predirect message, it either creates or updates a dynamic neighbor cache entry that stores the Target Address of the message as the network-layer address of Client ('B') and stores the link-layer address(es) found in the TLLAO(s) as the link-layer address(es) of Client ('B'). Client ('D') then applies the Prefix Length to the Interface Identifier portion of the Target Address and records the resulting IPv6 prefix in its IPv6 forwarding table. After processing the message, Client ('D') prepares a Redirect message response as follows: the link-layer source address is set to 'L2(D)' (i.e., the link-layer address of Client ('D')). the link-layer destination address is set to 'L2(A)' (i.e., the link-layer address of Server ('A')). the network-layer source address is set to 'L3(D)' (i.e., the AERO address of Client ('D')). the network-layer destination address is set to 'L3(B)' (i.e., the AERO address of Client ('B')). the Type is set to 137. the Code is set to 0 to indicate "Redirect". the Prefix Length is set to the length of the prefix to be applied to the Target and Destination address. the Target Address is set to fe80::2001:db8:1::1 (i.e., the AERO address of Client ('D')). the Destination Address is set to the IPv6 destination address of the packet that triggered the Redirection event. the message includes one or more TLLAOs set to 'L2(D)' and any other underlying address(es) of Client ('D'). the message includes as much of the RHO copied from the corresponding AERO Predirect message as possible such that at least the network-layer header is included but the size of the message does not exceed 1280 bytes. After Client ('D') prepares the Redirect message, it sends the message to Server ('A').

When Server ('A') receives a Redirect message, it accepts the message only if it has a neighbor cache entry that associates the message's link-layer source address with the network-layer source address. Next, Server ('A') validates the message according to the ICMPv6 Redirect message validation rules in Section 8.1 of . Following validation, Server ('A') re-encapsulates the Redirect then relays the re-encapsulated Redirect on to Client ('B') as follows. In the reference operational scenario, Server ('A') receives the Redirect message from Client ('D') and prepares to re-encapsulate and forward the message to Client ('B'). Server ('A') first verifies that Client ('D') is authorized to use the Prefix Length in the Redirect message when applied to the AERO address in the network-layer source of the Redirect message, and discards the message if verification fails. Otherwise, Server ('A') re-encapsulates the message by changing the link-layer source address of the message to 'L2(A)', changing the network-layer source address of the message to fe80::0, and changing the link-layer destination address to 'L2(B)' . Server ('A') finally relays the re-encapsulated message to the ingress node ('B') without decrementing the network-layer IPv6 header Hop Limit field. While not shown in , AERO Relays relay Redirect and Predirect messages in exactly this same fashion described above. See in Appendix A for an extension of the reference operational scenario that includes Relays.

When Client ('B') receives the Redirect message, it accepts the message only if it has a link-layer source address of the Server, i.e. 'L2(A)'. Next, Client ('B') validates the message according to the ICMPv6 Redirect message validation rules in Section 8.1 of . Following validation, Client ('B') then processes the message as follows. In the reference operational scenario, when Client ('B') receives the Redirect message, it either creates or updates a dynamic neighbor cache entry that stores the Target Address of the message as the network-layer address of Client ('D') and stores the link-layer address(es) found in the TLLAO(s) as the link-layer address(es) of Client ('D'). Client ('B') then applies the Prefix Length to the Interface Identifier portion of the Target Address and records the resulting IPv6 prefix in its IPv6 forwarding table. Now, Client ('B') has an IPv6 forwarding table entry for Client('D')'s prefix, and Client ('D') has an IPv6 forwarding table entry for Client ('B')'s prefix. Thereafter, the clients may exchange ordinary network-layer data packets directly without forwarding through Server ('A').

Each AERO node uses its delegated prefix to create an AERO address (see: Section 3.3). It can then send unicast NS messages to elicit NA messages from other AERO nodes the same as described for Neighbor Unreachability Detection (NUD) in except that the messages do not include S/TLLAOs. When an AERO node sends an NS/NA message, it MUST use its AERO address as the IPv6 source address and the AERO address of the neighbor as the IPv6 destination address. When an AERO node receives an NS/NA message, it accepts the message if it has a neighbor cache entry for the neighbor; otherwise, it ignores the message. When a source Client is redirected to a target Client it SHOULD test the direct path to the target by sending an initial NS message to elicit a solicited NA response. While testing the path, the source Client SHOULD continue sending packets via the Server until target Client reachability has been confirmed. The source Client SHOULD thereafter continue to test the direct path to the target Client (see Section 7.3 of ) in order to keep dynamic neighbor cache entries alive. In particular, the source Client sends NS messages to the target Client subject to rate limiting in order to receive solicited NA messages. If at any time the direct path appears to be failing, the source Client can resume sending packets via the Server which may or may not result in a new redirection event. When a target Client receives an NS message from a source Client, it resets the ACCEPT_TIME timer if a neighbor cache entry exists; otherwise, it discards the NS message. When a source Client receives a solicited NA message form a target Client, it resets the FORWARD_TIME timer if a neighbor cache entry exists; otherwise, it discards the NA message. When both the FORWARD_TIME and ACCEPT_TIME timers on a dynamic neighbor cache entry expire, the Client deletes both the neighbor cache entry and the corresponding IPv6 forwarding table entry. If the source Client is unable to elicit an NA response from the target Client after MAX_RETRY attempts, it SHOULD consider the direct path unusable for forwarding purposes. Otherwise, the source Client may continue to send packets directly to the target Client and SHOULD thereafter process any link-layer errors as a hint that the direct path to the target Client has either failed or has become intermittent.

When a Client needs to change its link-layer address (e.g., due to a mobility event, due to a change in underlying network interface, etc.), it sends an immediate NS message forward to any of its correspondents (including the Server and any other Clients) which then discover the new link-layer address. If two Clients change their link-layer addresses simultaneously, the NS/NA messages may be lost. In that case, the Clients SHOULD delete their respective dynamic neighbor cache and IPv6 forwarding table entries, and allow packets to again flow through their Server(s) which MAY result in a new AERO redirection exchange. When a Client needs to change to a new AERO Server, it issues a new DHCPv6 Request message via the new AERO Server as the DHCPv6 relay. The new AERO Server then relays the message to the DHCPv6 server and processes the resulting exchange the same as described in Section 3.9.2. After the Client receives the resulting DHCPv6 Reply message, it sends an RS message to the new Server to receive a new RA message and update its neighbor cache entry for fe80::0. After conducting the DHCPv6 exchange via the new AERO Server, the Client then sends a "terminating NS" message to the old AERO Server. The terminating NS message is prepared exactly the same as for an ordinary NS message, except that the Code field contains the value '1'. When the old Server receives the terminating NS message, it withdraws the IPv6 route from the routing system and deletes the neighbor cache entry and IPv6 forwarding table entry for the Client. The old Server then returns an NA message which the Client can use to verify that the termination signal has been processed. (Note that the old Server can impose a small delay before deleting the neighbor cache and IPv6 forwarding table entries so that any packets already in the system can still be delivered to the Client.) An alternative to sending a "terminating NS" message would be for the Client to somehow perform a DHCPv6 exchange with the DHCPv6 relay function on the old AERO Server but without involving the DHCPv6 server. This would be insecure because the Client only has a DHCPv6 security association with the DHCPv6 server and not the DHCPv6 relay. Conversely, the AERO Client and Server already require an authentic exchange of IPv6 Neighbor Discovery messages.

A source Client may connect only to an IPvX underlying network, while the target Client connects only to an IPvY underlying network. In that case, the source Client has no means for reaching the target directly (since they connect to underlying networks of different IP protocol versions) and so must ignore any Redirects and continue to send packets via the Server.

When the underlying network does not support multicast, AERO nodes map IPv6 link-scoped multicast addresses (including "All_DHCP_Relay_Agents_and_Servers") to the underlying IP address of the current AERO Server. When the underlying network supports multicast, AERO nodes use the multicast address mapping specification found in for IPv4 underlying networks and use a direct multicast mapping for IPv6 underlying networks. (In the latter case, "direct multicast mapping" means that if the IPv6 multicast destination address of the inner packet is "M", then the IPv6 multicast destination address of the encapsulating header is also "M".)

In some AERO link scenarios, there may be no Server on the link and/or no need for Clients to use a Server as an intermediary trust anchor. In that case, Clients can then establish dynamic neighbor cache entries and IPv6 forwarding table entries by performing direct Client-to-Client Predirect/Redirect exchanges, and some other form of trust basis must be applied so that each Client can verify that the prospective neighbor is authorized to use its claimed prefix. When there is no Server on the link, Clients must arrange to receive prefix delegations and publish the delegations via a secure alternate prefix delegation authority through some means outside the scope of this document.

IPv6 hosts serviced by an AERO Client can reach IPv4-only services via a NAT64 gateway within the IPv6 network. AERO nodes can use the Default Address Selection Policy with DHCPv6 option the same as on any IPv6 link. All other (non-multicast) functions that operate over ordinary IPv6 links operate in the same fashion over AERO links.