Network Working Group | F. Templin, Ed. |
Internet-Draft | Boeing Research & Technology |
Obsoletes: rfc6706 (if approved) | July 21, 2014 |
Intended status: Standards Track | |
Expires: January 22, 2015 |
Transmission of IP Packets over AERO Links
draft-templin-aerolink-30.txt
This document specifies the operation of IP over tunnel virtual links using Asymmetric Extended Route Optimization (AERO). Nodes attached to AERO links can exchange packets via trusted intermediate routers that provide forwarding services to reach off-link destinations and redirection services for route optimization. AERO provides an IPv6 link-local address format known as the AERO address that supports operation of the IPv6 Neighbor Discovery (ND) protocol and links IPv6 ND to IP forwarding. Admission control and provisioning are supported by the Dynamic Host Configuration Protocol for IPv6 (DHCPv6), and node mobility is naturally supported through dynamic neighbor cache updates. Although IPv6 ND messaging is used in the control plane, both IPv4 and IPv6 are supported in the data plane.
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 http://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 January 22, 2015.
Copyright (c) 2014 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 (http://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.
This document specifies the operation of IP over tunnel virtual links using Asymmetric Extended Route Optimization (AERO). The AERO link can be used for tunneling to neighboring nodes over either IPv6 or IPv4 networks, i.e., AERO views the IPv6 and IPv4 networks as equivalent links for tunneling. Nodes attached to AERO links can exchange packets via trusted intermediate routers that provide forwarding services to reach off-link destinations and redirection services for route optimization that addresses the requirements outlined in [RFC5522].
AERO provides an IPv6 link-local address format known as the AERO address that supports operation of the IPv6 Neighbor Discovery (ND) [RFC4861] protocol and links IPv6 ND to IP forwarding. Admission control and provisioning are supported by the Dynamic Host Configuration Protocol for IPv6 (DHCPv6) [RFC3315], and node mobility is naturally supported through dynamic neighbor cache updates. Although IPv6 ND message signalling is used in the control plane, either of IPv4 and IPv6 can be used in the data plane. The remainder of this document presents the AERO specification.
The terminology in the normative references applies; the following terms are defined within the scope of this document:
Throughout the document, the simple terms "Client", "Server" and "Relay" refer to "AERO Client", "AERO Server" and "AERO Relay", respectively. Capitalization is used to distinguish these terms from DHCPv6 client/server/relay. This is an important distinction to avoid ambiguity, e.g., an AERO Server also acts as a DHCPv6 relay, an AERO Relay may also act as a DHCPv6 server, etc.
Throughout the document, it is said that an address is "applied" to an AERO interface since the address need not always be "assigned" to the interface in the traditional sense. However, the address must at least be bound to the interface in some fashion for operation of the IPv6 ND protocol.
The terminology of [RFC4861] (including the names of node variables and protocol constants) applies to this document. Also throughout the document, the term "IP" is used to generically refer to either Internet Protocol version (i.e., IPv4 or IPv6).
The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in [RFC2119].
The following sections specify the operation of IP over Asymmetric Extended Route Optimization (AERO) links:
.-(::::::::) .-(:::: IP ::::)-. +-----------+ (:: Internetwork ::) | DHCPv6 | `-(::::::::::::)-' | Server X | `-(::::::)-' +-----------+ | +--------------+ +------+-------+ +--------------+ |AERO Server S1| | AERO Relay R | |AERO Server S2| | (default->R) | |(C->S1; D->S2)| | (default->R) | | Nbr: A | +-------+------+ | Nbr: B | +-------+------+ | +------+-------+ | | | X---+---+-------------------+------------------+---+---X | AERO Link | +-----+--------+ +--------+-----+ |AERO Client A | |AERO Client B | | default->S1 | | default->S2 | +--------------+ +--------------+ .-. .-. ,-( _)-. ,-( _)-. .-(_ IP )-. .-(_ IP )-. (__ EUN ) (__ EUN ) `-(______)-' `-(______)-' | | +--------+ +--------+ | Host C | | Host D | +--------+ +--------+
Figure 1: AERO Link Reference Model
Figure 1 above presents the AERO link reference model. In this model:
In this model, there may be many additional Relays, Servers and Clients. Each Sever peers with each Relay in a dynamic routing protocol session to advertise its list of associated Clients. Each Relay advertises the ASPs for the AERO link into the native IP Internetwork and serves as a gateway between the AERO link and the Internetwork. Clients may associate with only a single Server or with multiple Server, e.g., for fault tolerance and/or load balancing.
The DHCPv6 server is authoritative for the management of the AERO link's AERO Service Prefixes (ASPs). The DHCPv6 server is therefore critical infrastructure for the AERO link, but need not otherwise participate as an AERO node. AERO Servers communicate with the DHCPv6 server either via the AERO link itself or via a different IPv6 link.
AERO Relays relay packets between nodes connected to the same AERO link and also forward packets between the AERO link and the native Internetwork. The relaying process entails re-encapsulation of IP packets that were received from a first AERO node and are to be forwarded without modification to a second AERO node. AERO Relays present the AERO link to the native Internetwork as a set of one or more ASPs.
AERO Servers provide default routing services to AERO Clients. AERO Servers configure a DHCPv6 relay function and facilitate Prefix Delegation (PD) exchanges between AERO Clients and the DHCPv6 server. Each delegated prefix becomes an AERO Client Prefix (ACP) taken from an ASP.
AERO Clients act as requesting routers to receive ACPs through DHCPv6 PD exchanges via AERO Servers over the AERO link. (Each Client MAY associate with a single Server or with multiple Servers.) Each IPv6 AERO Client receives at least a /64 IPv6 ACP, and may receive even shorter prefixes. Similarly, each IPv4 AERO Client receives at least a /32 IPv4 ACP (i.e., a singleton IPv4 address), and may receive even shorter prefixes.
AERO Clients that act as routers sub-delegate portions of their ACPs to links on EUNs. End system applications on AERO Clients that act as routers bind to EUN interfaces (i.e., and not the AERO interface).
AERO Clients that act as ordinary hosts assign one or more IP addresses from their ACPs to the AERO interface but DO NOT assign the ACP itself to the AERO interface. Instead, the Client assigns the ACP to a "black hole" route so that unused portions of the prefix are nullified. End system applications on AERO Clients that act as hosts bind directly to the AERO interface.
An AERO address is an IPv6 link-local address with an embedded ACP and applied to a Client's AERO interface. The AERO address is formed as follows:
For IPv6, the AERO address begins with the prefix fe80::/64 and includes in its interface identifier the base prefix taken from the Client's IPv6 ACP. The base prefix is determined by masking the ACP with the prefix length. For example, if the AERO Client receives the IPv6 ACP:
it constructs its AERO address as:
[RFC4291] that includes the base prefix taken from the Client's IPv4 ACP. For example, if the AERO Client receives the IPv4 ACP:
For IPv4, the AERO address is formed as an IPv4-mapped IPv6 address
it constructs its AERO address as:
The AERO address remains stable as the Client moves between topological locations, i.e., even if its link-layer addresses change.
NOTE: In some cases, prospective neighbors may not have a priori knowledge of the Client's ACP length and may therefore send initial IPv6 ND messages with an AERO destination address that matches the ACP but does not correspond to the base prefix. In that case, the Client MUST accept the address as equivalent to the base address, but then use the base address as the source address of any IPv6 ND message replies. For example, if the Client receives the IPv6 ACP 2001:db8:1000:2000::/56 then subsequently receives an IPv6 ND message with destination address fe80::2001:db8:1000:2001, it accepts the message but uses fe80::2001:db8:1000:2000 as the source address of any IPv6 ND replies.
AERO interfaces use IP-in-IPv6 encapsulation [RFC2473] to exchange tunneled packets with AERO neighbors attached to an underlying IPv6 network, and use IP-in-IPv4 encapsulation [RFC2003][RFC4213] to exchange tunneled packets with AERO neighbors attached to an underlying IPv4 network. AERO interfaces can also coordinate secured tunnel types such as IPsec [RFC4301] or TLS [RFC5246]. When Network Address Translator (NAT) traversal and/or filtering middlebox traversal may be necessary, a UDP header is further inserted immediately above the IP encapsulation header.
AERO interfaces maintain a neighbor cache, and AERO Clients and Servers use an adaptation of standard unicast IPv6 ND messaging. AERO interfaces use unicast Neighbor Solicitation (NS), Neighbor Advertisement (NA), Router Solicitation (RS) and Router Advertisement (RA) messages the same as for any IPv6 link. AERO interfaces use two redirection message types -- the first known as a Predirect message and the second being the standard Redirect message (see Section 3.9). AERO links further use link-local-only addressing; hence, AERO nodes ignore any Prefix Information Options (PIOs) they may receive in RA messages.
AERO interface Redirect, Predirect and unsolicited NA messages include Target Link-Layer Address Options (TLLAOs) formatted as shown in Figure 2:
0 1 2 3 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Type = 2 | Length = 3 | Reserved | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Link ID | Preference | UDP Port Number (or 0) | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | | +-- --+ | | +-- IP Address --+ | | +-- --+ | | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
Figure 2: AERO Target Link-Layer Address Option (TLLAO) Format
In this format, Link ID is an integer value between 0 and 255 corresponding to an underlying interface of the target node, and Preference is an integer value between 0 and 255 indicating the node's preference for this underlying interface, with 0 being highest preference and 255 being lowest. UDP Port Number and IP Address are set to the addresses used by the target node when it sends encapsulated packets over the underlying interface. When no UDP encapsulation is used, UDP Port Number is set to 0. When the encapsulation IP address family is IPv4, IP Address is formed as an IPv4-mapped IPv6 address [RFC4291].
When a Relay enables an AERO interface, it applies an administratively assigned link-local address fe80::ID to the interface for communicating with Servers on the link. Each fe80::ID address MUST be unique among all Relays and Servers on the link, and MUST NOT collide with any potential AERO addresses, e.g., the addresses could be assigned as fe80::1, fe80::2, fe80::3, etc. The Relay also maintains an IP forwarding table entry for each Client-Server association and maintains a neighbor cache entry for each Server on the link. Relays do not require the use of IPv6 ND messaging for reachability determination since Relays and Servers engage in a dynamic routing protocol over the AERO interface. At a minimum, however, Relays respond to NS messages by returning an NA.
When a Server enables an AERO interface, it applies the address fe80:: to the interface as a link-local Subnet Router Anycast address, and also applies an administratively assigned link-local address fe80::ID to support the operation of the IPv6 ND protocol and to communicate with Relays on the link. The Server maintains a neighbor cache entry for each Relay on the link, and also creates per-Client neighbor cache entries whenever it discovers a new Client. At a minimum, when the Server receives an NS/RS messages on the AERO interface it returns an NA/RA message. When the Server receives an NS/NA, it also update timers in existing neighbor cache entries but does not create new neighbor cache entries nor update cached link-layer addresses. Servers also engage in a dynamic routing protocol with all Relays on the link. Finally, the Server provides a simple conduit between Clients and Relays, or between Clients and other Clients. Therefore, packets enter the Server's AERO interface from the link layer and are forwarded back out the link layer without ever leaving the AERO interface and therefore without ever disturbing the network layer.
When a Client enables an AERO interface, it invokes prefix delegation to receive an ACP. Next, it applies the corresponding AERO address to the AERO interface, i.e., the prefix delegation bootstraps the provisioning of a unique link-local address. The Client maintains a neighbor cache entry for each of its Servers and each of its active peer Clients. When the Client receives Redirect/Predirect messages on the AERO interface it updates or creates neighbor cache entries, including link-layer address information. Unsolicited NA messages update the cached link-layer address for the neighbor Client (e.g., following a link-layer address change due to node mobility) but do not create new neighbor cache entries. RA messages as well as NS/NA messages used for Neighbor Unreachability Detection (NUD) update timers in existing neighbor cache entires but do not update link-layer addresses nor create new neighbor cache entries. Redirect, Predirect and unsolicited NA messages SHOULD include a Timestamp option (see Section 5.3 of [RFC3971]) that other AERO nodes can use to verify the message time of origin. Predirect, NS and RS messages SHOULD include a Nonce option (see Section 5.3 of [RFC3971]) that recipients echo back in corresponding responses. Finally, the Client need not maintain any IP forwarding table entries for neighboring Clients. Instead, it can set a single "route-to-interface" default route in the IP forwarding table pointing to the AERO interface, and all forwarding decisions can be made within the AERO interface based on neighbor cache entries.
AERO interfaces may be configured over multiple underlying interfaces. For example, common mobile handheld devices have both wireless local area network ("WLAN") and cellular wireless links. These links are typically used "one at a time" with low-cost WLAN preferred and highly-available cellular wireless as a standby. In a more complex example, aircraft frequently have many wireless data link types (e.g. satellite-based, terrestrial, air-to-air directional, etc.) with diverse performance and cost properties.
If a Client's multiple underlying interfaces are used "one at a time" (i.e., all other interfaces are in standby mode while one interface is active), then Redirect, Predirect and unsolicited NA messages include only a single TLLAO with Link ID set to a constant value (e.g., 0).
If the Client has multiple active underlying interfaces, then from the perspective of IPv6 ND it would appear to have a single link-local address with multiple link-layer addresses. In that case, Redirect, Predirect and unsolicited NA messages MAY include multiple TLLAOs -- each with a different Link ID that corresponds to a specific underlying interface of the Client. Further details on coordination of multiple active underlying interfaces are outside the scope of this specification.
Each AERO interface maintains a conceptual neighbor cache that includes an entry for each neighbor it communicates with on the AERO link, the same as for any IPv6 interface [RFC4861]. Neighbor cache entries are created and maintained as follows:
AERO Relays maintain a permanent neighbor cache entry for each Server on the link, and AERO Servers maintain a permanent neighbor cache entry for each Relay on the link. AERO Clients maintain a neighbor cache entry for each of their associated Servers, and AERO Servers maintain a neighbor cache for each of their associated Clients with a lifetime based on the DHCPv6 lease lifetime. AERO Clients maintain neighbor cache entries for each of their active correspondent Clients with lifetimes based on IPv6 ND messaging constants.
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 AERO address corresponding to the Client's ACP as the network-layer address and with the Client's encapsulation IP address and UDP port number as the link-layer address. The Server also records the ACP's lease lifetime and prefix length in the neighbor cache entry.
When an AERO Client receives a DHCPv6 Reply message from an AERO Server, it creates or updates a neighbor cache entry for the Server based on the Reply message link-local source address as the network-layer address, the lease lifetime as the neighbor cache entry lifetime, and the encapsulation IP source address and UDP source port number as the link-layer address.
When an AERO Client receives a valid Predirect message it creates or updates a neighbor cache entry for the Predirect target network-layer and link-layer addresses plus prefix length. The node then sets an "AcceptTime" variable for the neighbor and uses this value to determine whether packets received from the predirected neighbor can be accepted.
When an AERO Client receives a valid Redirect message it creates or updates a neighbor cache entry for the Redirect target network-layer and link-layer addresses plus prefix length. The node then sets a "ForwardTime" variable for the neighbor and uses this value to determine whether packets can be sent directly to the redirected neighbor. The node also maintains a "Retry" variable to limit the number of keepalives sent when a neighbor may have gone unreachable.
When an AERO Client receives a valid NS message corresponding to a neighbor cache entry for another Client, it (re)sets AcceptTime for the neighbor to ACCEPT_TIME.
When an AERO Client receives a valid solicited NA message corresponding to a neighbor cache entry for another Client, it (re)sets ForwardTime for the neighbor to FORWARD_TIME and sets Retry to MAX_RETRY. (When an AERO Client receives a valid unsolicited NA message, it updates the neighbor's link-layer address but DOES NOT reset ForwardTime or Retries.)
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 ND [RFC4861].
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 AcceptTime decrements below FORWARD_TIME.
It is RECOMMENDED that MAX_RETRY be set to 3 the same as described for IPv6 ND address resolution in Section 7.3.3 of [RFC4861].
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. Most importantly, ACCEPT_TIME SHOULD be set to a value that is sufficiently longer than FORWARD_TIME to allow the AERO redirection procedure to converge.
For AERO Client<->Server neighbor cache entries, AcceptTime and ForwardTime are set based on the DHCPv6 lease lifetime and may be modified based on the Router Lifetime advertised in the Server's RA messages.
When an IP packet enters a Client's AERO interface from the network layer, the Client searches its neighbor cache for an entry with an AERO address that matches the packet's destination address. If there is a match, the Client uses the link-layer address in the neighbor cache entry as the link-layer address for encapsulation then admits the packet into the tunnel. If there is no match, the Client instead uses the link-layer address of a neighboring Server as the link-layer address for encapsulation. (Note that the Client caches the ASPs for the AERO link and can thus search the neighbor cache only for destination addresses that are covered by an ASP.)
When an IP packet enters a Server's AERO interface from the link layer, the Server searches for a neighbor cache match the same as for a Client. If there is a match, the Server uses the link-layer address in the neighbor cache entry as the link-layer address for re-encapsulation. If there is no match, the Server instead uses the link-layer address of a neighboring Relay as the link-layer address for encapsulation. Servers also relay Predirect, Redirect and unsolicited Neighbor Advertisement messages received from a Client and with an AERO destination address. If the AERO destination address is the address of a neighbor, the Server changes the link-layer source address to its own address, changes the link-layer destination address to the address of the neighbor and forwards the message to the neighbor. If the AERO destination address is not a neighbor, the Server instead forwards the message to a Relay. When an AERO Relay forwards either a data packet or an IPv6 ND message to an AERO Server, the Server MUST NOT forward the packet back to the same or a different Relay.
When an IP packet enters a Relay's AERO interface from the network layer, the Relay searches its IP forwarding table for an entry that is covered by an ASP and also matches the destination. If there is a match, the Relay uses the link-layer address in the neighbor cache entry for the next-hop Server as the link-layer address for encapsulation. When an IP packet enters a Relay's AERO interface from the link-layer, if the destination is not covered by an ASP the Relay forwards the packet to another IP link as indicated by the IP forwarding table. If the destination is covered by an ASP, and there is a more-specific forwarding table entry that matches the destination, the Relay uses the link-layer address in the neighbor cache entry for the next-hop Server as the link-layer address for encapsulation. If there is no more-specific entry, the Relay instead drops the packet. Relays also relay Predirect, Redirect and unsolicited Neighbor Advertisement messages by searching for an IP forwarding table entry that matches the message's AERO destination address. If there is a match, the Relay proxies the packet in the same manner as described for Servers above; otherwise, the Relay drops the packet. When an AERO Server forwards either a data packet or an IPv6 ND message to an AERO Relay, the Relay MUST NOT forward the packet back to the same Server.
Note that in the above this tunnel exit determination is often based on consulting the neighbor cache instead of the IP forwarding table. IP forwarding is therefore linked to IPv6 ND via the AERO address.
When an AERO node forwards a packet back out the same AERO interface the packet arrived on, the node MUST NOT decrement the network layer TTL/Hop-count.
AERO interfaces encapsulate IP packets according to whether they are entering the AERO interface from the network layer 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 [RFC2003][RFC2473][RFC4213][RFC4301][RFC5246] (etc.) except that the interface copies the "TTL/Hop Limit", "Type of Service/Traffic Class" and "Congestion Experienced" values in the packet's IP header into the corresponding fields in the encapsulation 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 encapsulation header into the corresponding fields in the new encapsulation header (i.e., the values are transferred between encapsulation headers and *not* copied from the encapsulated packet's network-layer header).
When AERO UDP encapsulation is used, the AERO interface encapsulates the packet per the specifications in [RFC2003][RFC2473][RFC4213] except that it inserts a UDP header between the encapsulation header and the packet's 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: [RFC6935][RFC6936]) and sets the UDP length field to the length of the IP 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-registered port number for AERO) when AERO-only encapsulation is used. 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 IP protocol number in the encapsulation header to the appropriate value for the first protocol layer within the encapsulation (e.g., IPv4, IPv6, UDP, IPsec, etc.). When IPv6 is used as the encapsulation protocol, the interface then sets the flow label value in the encapsulation header the same as described in [RFC6438]. When IPv4 is used as the encapsulation protocol, the AERO interface sets the DF bit as discussed in Section 3.7.
AERO interfaces decapsulate packets destined either to the node itself or to a destination reached via an interface other than the AERO interface the packet was received on. When AERO UDP encapsulation is used (i.e., when a UDP header with destination port 8060 is present) the interface examines the first octet of the encapsulated packet. If the most significant four bits of the first octet encode the value '0110' (i.e., the version number value for IPv6) or the value '0100' (i.e., the version number value for IPv4), the packet is accepted and the encapsulating UDP header is discarded; otherwise, the packet is discarded.
Further decapsulation then proceeds according to the appropriate tunnel type [RFC2003][RFC2473][RFC4213][RFC4301][RFC5246] (etc.).
AERO nodes employ simple data origin authentication procedures for encapsulated packets they receive from other nodes. In particular, AERO Clients accept encapsulated packets with a link-layer source address belonging to one of their current AERO Servers, and AERO Clients and Servers 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 AERO interface neighbor cache entry with an AERO address that matches the packet's network-layer source address prefix, with a link-layer address that matches the packet's link-layer source address, and AcceptTime is non-zero.
An AERO Server also accepts packets with a link-layer source address that matches one of its associated Relays, and an AERO Relay accepts packets with a source address that matches one of its associated Servers.
Note that this simple data origin authentication only applies to environments in which link-layer addresses cannot be spoofed. Additional security mitigations may be necessary in other environments.
The AERO link Maximum Transmission Unit (MTU) is 64KB minus the encapsulation overhead for IPv4 as the link-layer [RFC0791] and 4GB minus the encapsulation overhead for IPv6 as the link layer [RFC2675]. 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. (If there are multiple underlying interfaces, the node sets the AERO interface MTU according to the largest underlying interface MTU, or 64KB /4G minus the encapsulation overhead if the largest MTU cannot be determined.) 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 IP fragmentation to fragment the encapsulated packet into two pieces (where the first fragment contains 1024 bytes of the original IP packet) then admits the fragments into the tunnel. If the link-layer protocol is IPv4, the node admits each fragment into the tunnel with DF set to 0 and subject to rate limiting to avoid reassembly errors [RFC4963][RFC6864]. 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 NS message with a Nonce option plus trailing padding octets added to a length of 1500 bytes without including the length of the padding in the IPv6 Payload Length field. The node then encapsulates the NS in the encapsulation headers (while including the length of the padding in the encapsulation header length fields), sets DF to 1 (for IPv4) and sends the padded NS message to the neighbor. If the neighbor returns an NA message with a correct Nonce value, the node may then send whole packets within this size range and (for IPv4) relax the rate limiting requirement. (Note that the trailing padding SHOULD NOT be included within the Nonce option itself but rather as padding beyond the last option in the NS message; otherwise, the (large) Nonce option would be echoed back in the solicited NA message and may be lost at a link with a small MTU along the reverse path.)
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 [RFC4821] when sending large packets.
AERO Clients discover the link-layer addresses of AERO Servers 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 connection-specific DNS suffix for the Client's underlying network connection. After discovering the link-layer addresses, the Client associates with one or more of the corresponding Servers.
To associate with a Server, the Client acts as a requesting router to request an ACP through DHCPv6 PD [RFC3315][RFC3633][RFC6355] using 'fe80::ffff:ffff:ffff:ffff' as the IPv6 source address, 'All_DHCP_Relay_Agents_and_Servers' as the IPv6 destination address and the link-layer address of the Server as the link-layer destination address. The Client includes a DHCPv6 Unique Identifier (DUID) in the Client Identifier option of its DHCPv6 messages (as well as a DHCPv6 authentication option if necessary) to identify itself to the DHCPv6 server. The Client also includes a DHCPv6 Client Link Layer Address Option (CLLAO) [RFC6939] with the link-layer address format shown in Figure 2 with Link ID followed by Preference followed by the values 0 for both the UDP Port Number and IP Address.The Client sets the CLLAO 'option-length' field to 22 (2 plus the length of the link-layer address) and sets the 'link-layer type' field to TBD1 (see: IANA Considerations). If the Client is pre-provisioned with an ACP associated with the AERO service, it MAY also include the ACP in its DHCPv6 PD Request to indicate its preferred ACP to the DHCPv6 server. The Client then sends the encapsulated DHCPv6 request via an underlying interface.
When the Client receives its ACP and the set of ASPs via a Reply from the DHCPv6 server, i.e., via an AERO Server acting as a DHCPv6 relay, it creates a neighbor cache entry with the Server's link-local address (i.e., fe80::ID) as the network-layer address and the Server's encapsulation address as the link-layer address.
The Client then applies the AERO address to the AERO interface and sub-delegates the ACP to nodes and links within its attached EUNs (the AERO address thereafter remains stable as the Client moves). The Client also assigns a default IP route to the AERO interface as a route-to-interface, i.e., with no explicit next-hop.
The Client subsequently renews its ACP delegation by performing DHCPv6 Renew/Reply exchanges with its AERO address as the IPv6 source address, 'All_DHCP_Relay_Agents_and_Servers' as the IPv6 destination address, the link-layer address of a Server as the link-layer destination address and the same DUID and authentication information. If the Client wishes to associate with multiple Servers, it can perform DHCPv6 Renew/Reply exchanges via each of the Servers.
The Client then sends an RS message to each of its associated Servers to receive an RA message with a Router Lifetime and any other link-specific parameters. The Client uses the Router Lifetime to set the lifetime for the neighbor cache entry for this Server. The Client further ignores any RS messages it might receive, since only Servers may process RS messages.
The Client then sends periodic RS messages to obtain new RA messages, and further initiates a new DHCPv6 Renew/Reply exchange before the Router Lifetime expires. The Client can also forward IP packets destined to networks beyond its local EUNs via a Server as a default router.
Since the Client's AERO address is configured from the unique ACP delegation it receives, 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 access to the network by the DHCPv6 server due to an unacceptable link-layer address and/or security parameters (see: Security Considerations).
AERO Clients ignore the IP address and UDP port number in any S/TLLAO options in ND messages they receive directly from another AERO Client, but examine the Link ID and Preference values to match the message with the correct link-layer address information.
When a source Client forwards a packet to a prospective destination Client (i.e., one for which the packet's destination address is covered by an ASP), the source Client initiates an AERO route optimization procedure as specified in Section 3.12.
AERO Servers configure a DHCPv6 relay function on their AERO links. AERO Servers arrange to add their encapsulation layer IP addresses (i.e., their link-layer addresses) to the DNS resource records for the FQDN "linkupnetworks.[domainname]" before entering service. Here, "linkupnetworks" is a constant text string, and "[domainname]" is the connection-specific DNS suffix for this underlying network connection.
When an AERO Server relays a prospective Client's DHCPv6 PD messages to the DHCPv6 server, it wraps each message in a "Relay-forward" message per [RFC3315] and includes a DHCPv6 Interface Identifier option that encodes a value that identifies the AERO link to the DHCPv6 server. Without creating internal state, the Server then modifies the Client's link-layer address in the CLLAO [RFC6939] by writing the client's UDP Port number and IP adddress in the corresponding fields of the option. The Server finally includes a DHCPv6 Echo Request Option (ERO) [RFC4994] that encodes the option code for the CLLAO in a 'requested-option-code-n' field, then relays the message to the DHCPv6 server. The CLLAO information will therefore subsequently be echoed back in the DHCPv6 server's "Relay-reply" message.
When the DHCPv6 server issues the ASPs and ACP prefix delegation in a "Relay-reply" message via the AERO Server (acting as a DHCPv6 relay), the Server obtains the Client's link-layer address from the echoed CLLAO option and also obtains the Client's delegated ACP and lease lifetime from the message. The Server then creates a neighbor cache entry for the Client's AERO address with the Client's link-layer address as the link-layer address and with lifetime set to no more than the lease lifetime. The Server finally injects the ACP as an IP route into the inter-Server/Relay routing system (see: Section 3.11) then relays the DHCPv6 message to the Client while using fe80::ID 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 NS/RS messages from Clients on their AERO interfaces by returning an NA/RA message. When the Server returns an RA message, it sets Router Lifetime to the neighbor cache entry lifetime but does not include any Prefix Information Options (PIOs) since the AERO link is link-local-only. The server decrements the neighbor cache entry lifetime according to the system clock.
Servers ignore any RA messages they may receive from a Client, but they MAY examine RA messages received from other Servers for consistency verification purposes.
The DHCPv6 server observes both the base DHCPv6 specification [RFC3315] and the DHCPv6 PD specification [RFC3633]. The DHCPv6 server further MUST honor the DHCPv6 Echo Request Option (ERO) and Client Link-Layer Address Option (CLLAO) as discussed in Section 3.10.1.
The DHCPv6 server also includes a DHCPv6 Vendor-Specific Information Option with 'enterprise-number' set to "TBD2" (see: IANA Considerations). The option is formatted as shown in[RFC3315] and with the AERO enterprise-specific format shown in Figure 3:
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | OPTION_VENDOR_OPTS | option-len | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | enterprise-number ("TBD2") | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Reserved | Prefix Length | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | | + ASP (1) + | | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Reserved | Prefix Length | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | | + ASP (2) + | | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Reserved | Prefix Length | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | | + ASP (3) + | | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ . (etc.) . . . +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
Figure 3: AERO Vendor-Specific Information Option
Figure 3, the option includes one or more ASP. The Prefix Length field must contain a value between 0 - 64 and the ASP field contains the leading 64 bits of the ASP. When the Client receives an AERO Vendor-Specific Information Option it accepts the option and caches each ASP that observes the format specified above. If the Client cannot parse the ASPs, it ignores the option.
Relays require full topology information of all Client/Server associations, while individual Servers only require partial topology information, i.e., they only need to know the ACPs associated with their current set of associated Clients. This is accomplished through the use of an internal instance of the Border Gateway Protocol (BGP) [RFC4271] coordinated between Servers and Relays. This internal BGP instance does not interact with the public Internet BGP instance; therefore, the AERO link is presented to the IP Internetwork as a small set of ASPs as opposed to the full set of individual ACPs.
In a reference BGP arrangement, each AERO Server is configured as an Autonomous System Border Router (ASBR) for a stub Autonomous System (AS) (possibly using a private AS Number (ASN) [RFC1930]), and each Server further peers with each Relay but does not peer with other Servers. Similarly, Relays need not peer with each other, since they will receive all updates from all Servers and will therefore have a consistent view of the AERO link ACP delegations.
Each Server maintains a working set of associated Clients, and dynamically announces new ACPs and withdraws departed ACPs in its BGP updates to Relays. Relays do not send BGP updates to Servers, however, such that the BGP route reporting is unidirectional from the Servers to the Relays.
The Relays therefore discover the full topology of the AERO link in terms of the working set of ACPs associated with each Server, while the Servers only discover the ACPs of their associated Clients. Since Clients are expected to remain associated with their current set of Servers for extended timeframes, the amount of BGP control messaging between Servers and Relays should be minimal. However, BGP peers SHOULD dampen any route oscillations caused by impatient Clients that repeatedly associate and disassociate with Servers.
Figure 4 depicts the AERO redirection reference operational scenario, using IPv6 addressing as the example (while not shown, a corresponding example for IPv4 addressing can be easily constructed). The figure shows an AERO Relay ('R'), two AERO Servers ('S1', 'S2'), two AERO Clients ('A', 'B') and two ordinary IPv6 hosts ('C', 'D'):
+--------------+ +--------------+ +--------------+ | Server S1 | | Relay R | | Server S2 | | Nbr: A | |(C->S1; D->S2)| | Nbr: B | +--------------+ +--------------+ +--------------+ fe80::2 fe80::1 fe80::3 L2(S1) L2(R) L2(S2) | | | X-----+-----+------------------+-----------------+----+----X | AERO Link | L2(A) L2(B) fe80::2001:db8:0:0 fe80::2001:db8:1:0 +--------------+ +--------------+ | AERO Client A| | AERO Client B| | (default->S1)| | (default->S2)| +--------------+ +--------------+ 2001:DB8:0::/48 2001:DB8:1::/48 | | .-. .-. ,-( _)-. 2001:db8:0::1 2001:db8:1::1 ,-( _)-. .-(_ IP )-. +---------+ +---------+ .-(_ IP )-. (__ EUN )--| Host C | | Host D |--(__ EUN ) `-(______)-' +---------+ +---------+ `-(______)-'
Figure 4: AERO Reference Operational Scenario
In Figure 4, Relay ('R') applies the address fe80::1 to its AERO interface with link-layer address L2(R), Server ('S1') applies the address fe80::2 with link-layer address L2(S1),and Server ('S2') applies the address fe80::3 with link-layer address L2(S2). Servers ('S1') and ('S2') next arrange to add their link-layer addresses to a published list of valid Servers for the AERO link.
AERO Client ('A') receives the ACP 2001:db8:0::/48 in a DHCPv6 PD exchange via AERO Server ('S1') then assigns the address fe80::2001:db8:0:0 to its AERO interface with link-layer address L2(A). Client ('A') configures a default route and neighbor cache entry via the AERO interface with next-hop address fe80::2 and link-layer address L2(S1), then sub-delegates the ACP to its attached EUNs. IPv6 host ('C') connects to the EUN, and configures the address 2001:db8:0::1.
AERO Client ('B') receives the ACP 2001:db8:1::/48 in a DHCPv6 PD exchange via AERO Server ('S2') then assigns the address fe80::2001:db8:1:0 to its AERO interface with link-layer address L2(B). Client ('B') configures a default route and neighbor cache entry via the AERO interface with next-hop address fe80::3 and link-layer address L2(S2), then sub-delegates the ACP to its attached EUNs. IPv6 host ('D') connects to the EUN, and configures the address 2001:db8:1::1.
Again, with reference to Figure 4, when source host ('C') sends a packet to destination host ('D'), the packet is first forwarded over the source host's attached EUN to Client ('A'). Client ('A') then forwards the packet via its AERO interface to Server ('S1') and also sends a Predirect message toward Client ('B') via Server ('S1'). Server ('S1') then re-encapsulates and forwards both the packet and the Predirect message out the same AERO interface toward Client ('B') via Relay ('R').
When Relay ('R') receives the packet and Predirect message, it consults its forwarding table to discover Server ('S2') as the next hop toward Client ('B'). Relay ('R') then forwards both messages to Server ('S2'), which then forwards them to Client ('B').
After Client ('B') receives the Predirect message, it process the message and returns a Redirect message toward Client ('A') via Server ('S2'). During the process, Client ('B') also creates or updates a neighbor cache entry for Client ('A').
When Server ('S2') receives the Redirect message, it re-encapsulates the message and forwards it on to Relay ('R'), which forwards the message on to Server ('S1') which forwards the message on to Client ('A'). After Client ('A') receives the Redirect message, it processes the message and creates or updates a neighbor cache entry for Client ('C').
Following the above Predirect/Redirect message exchange, forwarding of packets from Client ('A') to Client ('B') without involving any intermediate nodes 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 [RFC4861], but also include a new "Prefix Length" field taken from the low-order 8 bits of the Redirect message Reserved field. (For IPv6, valid values for the Prefix Length field are 0 through 64; for IPv4, valid values are 0 through 32.) The Redirect/Predirect messages are formatted as shown in Figure 5:
0 1 2 3 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Type (=137) | Code (=0/1) | Checksum | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Reserved | Prefix Length | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | | + + | | + Target Address + | | + + | | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | | + + | | + Destination Address + | | + + | | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Options ... +-+-+-+-+-+-+-+-+-+-+-+-
Figure 5: AERO Redirect/Predirect Message Format
When a Client forwards a packet with a source address from one of its ACPs toward a destination address covered by an ASP (i.e., toward another AERO Client connected to the same AERO link), the source Client MAY send a Predirect message forward toward the destination Client via the Server.
In the reference operational scenario, when Client ('A') forwards a packet toward Client ('B'), it MAY also send a Predirect message forward toward Client ('B'), subject to rate limiting (see Section 8.2 of [RFC4861]). Client ('A') prepares the Predirect message as follows:
Note that the act of sending Predirect messages is cited as "MAY", since Client ('A') may have advanced knowledge that the direct path to Client ('B') would be unusable. If the direct path later becomes unusable after the initial route optimization, Client ('A') simply allows packets to again flow through Server ('S1').
When Server ('S1') receives a Predirect message from Client ('A'), it first verifies that the requested redirection is authorized. If the redirection is not permitted, Server ('S1') discards the message. Otherwise, Server ('S1') validates the message according to the ICMPv6 Redirect message validation rules in Section 8.1 of [RFC4861], except that the Predirect has Code=1. Server ('S1') also verifies that Client ('A') is authorized to use the Prefix Length in the Predirect when applied to the AERO address in the network-layer source address by searching for the AERO address in the neighbor cache. If validation fails, Server ('S1') discards the Predirect; otherwise, it copies the correct UDP Port number and IP Address for Client ('A') into the (previously empty) TLLAO.
Server ('S1') then examines the network-layer destination address of the Predirect to determine the next hop toward Client ('B') by searching for the AERO address in the neighbor cache. Since Client ('B') is not one of its neighbors, Server ('S1') re-encapsulates the Predirect and relays it via Relay ('R') by changing the link-layer source address of the message to 'L2(S1)' and changing the link-layer destination address to 'L2(R)'. Server ('S1') finally forwards the re-encapsulated message to Relay ('R') without decrementing the network-layer TTL/Hop Limit field.
When Relay ('R') receives the Predirect message from Server ('S1') it determines that Server ('S2') is the next hop toward Client ('B') by consulting its forwarding table. Relay ('R') then re-encapsulates the Predirect while changing the link-layer source address to 'L2(R)' and changing the link-layer destination address to 'L2(S2)'. Relay ('R') then relays the Predirect via Server ('S2').
When Server ('S2') receives the Predirect message from Relay ('R') it determines that Client ('B') is a neighbor by consulting its neighbor cache. Server ('S2') then re-encapsulates the Predirect while changing the link-layer source address to 'L2(S2)' and changing the link-layer destination address to 'L2(B)'. Server ('S2') then forwards the message to Client ('B').
When Client ('B') receives the Predirect message, it accepts the Predirect only if the message has a link-layer source address of one of its Servers (e.g., L2(S2)). Client ('B') further accepts the message only if it is willing to serve as a redirection target. Next, Client ('B') validates the message according to the ICMPv6 Redirect message validation rules in Section 8.1 of [RFC4861], except that it accepts the message even though Code=1 and even though the network-layer source address is not that of it's current first-hop router.
In the reference operational scenario, when Client ('B') receives a valid Predirect message, it either creates or updates a neighbor cache entry that stores the Target Address of the message as the network-layer address of Client ('A') , stores the link-layer address found in the TLLAO as the link-layer address(es) of Client ('A') and stores the Prefix Length as the length to be applied to the network-layer address for forwarding purposes. Client ('B') then sets AcceptTime for the neighbor cache entry to ACCEPT_TIME.
After processing the message, Client ('B') prepares a Redirect message response as follows:
After Client ('B') prepares the Redirect message, it sends the message to Server ('S2').
When Server ('S2') receives a Redirect message from Client ('B'), it first verifies that the requested redirection is authorized. If the redirection is not permitted, Server ('S2') discards the message. Otherwise, Server ('S2') validates the message according to the ICMPv6 Redirect message validation rules in Section 8.1 of [RFC4861]. Server ('S2') also verifies that Client ('B') is authorized to use the Prefix Length in the Redirect when applied to the AERO address in the network-layer source address by searching for the AERO address in the neighbor cache. If validation fails, Server ('S2') discards the Predirect; otherwise, it copies the correct UDP Port number and IP Address for Client ('B') into the (previously empty) TLLAO.
Server ('S2') then examines the network-layer destination address of the Predirect to determine the next hop toward Client ('A') by searching for the AERO address in the neighbor cache. Since Client ('A') is not one of its neighbors, Server ('S2') re-encapsulates the Predirect and relays it via Relay ('R') by changing the link-layer source address of the message to 'L2(S2)' and changing the link-layer destination address to 'L2(R)'. Server ('S2') finally forwards the re-encapsulated message to Relay ('R') without decrementing the network-layer TTL/Hop Limit field.
When Relay ('R') receives the Predirect message from Server ('S2') it determines that Server ('S1') is the next hop toward Client ('A') by consulting its forwarding table. Relay ('R') then re-encapsulates the Predirect while changing the link-layer source address to 'L2(R)' and changing the link-layer destination address to 'L2(S1)'. Relay ('R') then relays the Predirect via Server ('S1').
When Server ('S1') receives the Predirect message from Relay ('R') it determines that Client ('A') is a neighbor by consulting its neighbor cache. Server ('S1') then re-encapsulates the Predirect while changing the link-layer source address to 'L2(S1)' and changing the link-layer destination address to 'L2(A)'. Server ('S1') then forwards the message to Client ('A').
When Client ('A') receives the Redirect message, it accepts the message only if it has a link-layer source address of one of its Servers (e.g., ''L2(S1)'). Next, Client ('A') validates the message according to the ICMPv6 Redirect message validation rules in Section 8.1 of [RFC4861], except that it accepts the message even though the network-layer source address is not that of it's current first-hop router. Following validation, Client ('A') then processes the message as follows.
In the reference operational scenario, when Client ('A') receives the Redirect message, it either creates or updates a neighbor cache entry that stores the Target Address of the message as the network-layer address of Client ('B'), stores the link-layer address found in the TLLAO as the link-layer address of Client ('B') and stores the Prefix Length as the length to be applied to the network-layer address for forwarding purposes. Client ('A') then sets ForwardTime for the neighbor cache entry to FORWARD_TIME.
Now, Client ('A') has a neighbor cache entry with a valid ForwardTime value, while Client ('B') has a neighbor cache entry with a valid AcceptTime value. Thereafter, Client ('A') may forward ordinary network-layer data packets directly to Client ("B") without involving any intermediate nodes, and Client ('B') can verify that the packets came from an acceptable source. (In order for Client ('B') to forward packets to Client ('A'), a corresponding Predirect/Redirect message exchange is required in the reverse direction; hence, the mechanism is asymmetric.)
In some environments, the Server nearest the destination Client may need to serve as the redirection target, e.g., if direct Client-to-Client communications are not possible. In that case, the Server prepares the Redirect message the same as if it were the destination Client (see: Section 3.9.6), except that it writes its own link-layer address in the TLLAO option. The Server must then maintain a neighbor cache entry for the redirected source Client.
AERO nodes perform NUD by sending unicast NS messages to elicit solicited NA messages from neighbors the same as described in [RFC4861]. When an AERO node sends an NS/NA message, it MUST use its AERO address as the IPv6 source address and the link-local address of the neighbor as the IPv6 destination address. When an AERO node receives an NS message or a solicited 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 by sending an initial NS message to elicit a solicited NA response. While testing the path, the source Client can optionally continue sending packets via the Server, maintain a small queue of packets until target reachability is confirmed, or (optimistically) allow packets to flow directly to the target. The source Client SHOULD thereafter continue to test the direct path to the target Client (see Section 7.3 of [RFC4861]) periodically in order to keep neighbor cache entries alive.
In particular, while the source Client is actively sending packets to the target Client it SHOULD also send NS messages separated by RETRANS_TIMER milliseconds in order to receive solicited NA messages. If the source Client is unable to elicit a solicited NA response from the target Client after MAX_RETRY attempts, it SHOULD set ForwardTime to 0 and resume sending packets via the Server which may or may not result in a new redirection event. Otherwise, the source Client considers the path usable 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 target Client receives an NS message from a source Client, it resets AcceptTime to ACCEPT_TIME if a neighbor cache entry exists; otherwise, it discards the NS message.
When a source Client receives a solicited NA message from a target Client, it resets ForwardTime to FORWARD_TIME if a neighbor cache entry exists; otherwise, it discards the NA message.
When ForwardTime for a neighbor cache entry expires, the source Client resumes sending any subsequent packets via the Server and may (eventually) attempt to re-initiate the AERO redirection process. When AcceptTime for a neighbor cache entry expires, the target Client discards any subsequent packets received directly from the source Client. When both ForwardTime and AcceptTime for a neighbor cache entry expire, the Client deletes the neighbor cache entry.
When a Client needs to change its link-layer address, e.g., due to a mobility event, it performs an immediate DHCPv6 Renew/Reply via each of its Servers using the new link-layer address as the source. The DHCPv6 Server will re-authenticate the Client and (assuming authentication succeeds) the DHCPv6 Renew/Reply exchange will update each Server's neighbor cache.
Next, the Client sends unsolicited NA messages to each of its active neighbors using the same procedures as specified in Section 7.2.6 of [RFC4861], except that it sends the messages as unicast to each neighbor via a Server instead of multicast. In this process, the Client should send no more than MAX_NEIGHBOR_ADVERTISEMENT messages separated by no less than RETRANS_TIMER seconds to each neighbor.
With reference to Figure 4, Client ('B') sends unicast unsolicited NA messages to Client ('A') via Server ('S2') as follows:
When Server ('S1') receives the NA message, it relays the message in the same way as described for relaying Redirect messages in Section 3.12.7. In particular, Server ('S1') copies the correct UDP port number and IP address into the TLLAO, changes the link-layer source address to its own address, changes the link-layer destination address to the address of Relay ('R'), then forwards the NA message via the relaying chain the same as for a Redirect.
When Client ('A') receives the NA message, it accepts the message only if it already has a neighbor cache entry for Client ('B') then updates the link-layer address for Client ('B') based on the address in the TLLAO. However, Client ('A') MUST NOT update ForwardTime since Client ('B') will not have updated AcceptTime.
Note that these unsolicited NA messages are unacknowledged; hence, Client ('B') has no way of knowing whether Client ('A') has received them. If the messages are somehow lost, however, Client ('A') will soon learn of the mobility event via the NUD procedures specified in Section 3.13.
When a Client associates with a new Server, it issues a new DHCPv6 Renew message via the new Server as the DHCPv6 relay. The new Server then relays the message to the DHCPv6 server and processes the resulting exchange. After the Client receives the resulting DHCPv6 Reply message, it sends an RS message to the new Server to receive a new RA message.
When a Client disassociates with an existing Server, it sends a "terminating RS" message to the old Server. The terminating RS message is prepared exactly the same as for an ordinary RS message, except that the Code field contains the value '1'. When the old Server receives the terminating RS message, it withdraws the IP route from the routing system and deletes the neighbor cache entry for the Client. The old Server then returns an RA message with default router lifetime set to 0 which the Client can use to verify that the termination signal has been processed. The client then deletes both the default route and the neighbor cache entry for the old Server. The old Server SHOULD impose a small delay before deleting the neighbor cache entry so that any packets already in the system can still be delivered to the Client.
Clients SHOULD NOT move rapidly between Servers in order to avoid causing unpredictable oscillations in the Server/Relay routing system. Such oscillations could result in intermittent reachability for the Client itself, while causing little harm to the network due to routing protocol dampening. Examples of when a Client may change to a different Server include a Server that has gone unreachable, topological movements of significant distance, etc.
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 target and source Clients have no means for reaching each other directly (since they connect to underlying networks of different IP protocol versions) and so must ignore any redirection messages 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 link-layer address of a Server.
When the underlying network supports multicast, AERO nodes use the multicast address mapping specification found in [RFC2529] 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 encapsulated packet is "M", then the IPv6 multicast destination address of the encapsulating header is also "M".)
When the AERO link does not provide DHCPv6 services, operation can still be accommodated through administrative configuration of ACPs on AERO Clients. In that case, administrative configurations of AERO interface neighbor cache entries on both the Server and Client are also necessary. However, this may interfere with the ability for Clients to dynamically change to new Servers, and can expose the AERO link to misconfigurations unless the administrative configurations are carefully coordinated.
In some AERO link scenarios, there may be no Servers on the link and/or no need for Clients to use a Server as an intermediary trust anchor. In that case, each Client acts as a Server unto itself to establish neighbor cache 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 ACP.
When there is no Server on the link, Clients must arrange to receive ACPs and publish them via a secure alternate prefix delegation authority through some means outside the scope of this document.
An application-layer implementation is in progress.
The IANA is instructed to assign a new 2-octet Hardware Type number "TBD1" for AERO in the "arp-parameters" registry per Section 2 of [RFC5494]. The number is assigned from the 2-octet Unassigned range with Hardware Type "AERO" and with this document as the reference.
The IANA is further instructed to assign a 4-octet Enterprise Number "TBD2" for AERO in the "enterprise-numbers" registry per [RFC3315].
AERO link security considerations are the same as for standard IPv6 Neighbor Discovery [RFC4861] except that AERO improves on some aspects. In particular, AERO uses a trust basis between Clients and Servers, where the Clients only engage in the AERO mechanism when it is facilitated by a trust anchor. AERO also uses DHCPv6 authentication for Client authentication and network admission control.
AERO links must be protected against link-layer address spoofing attacks in which an attacker on the link pretends to be a trusted neighbor. Links that provide link-layer securing mechanisms (e.g., IEEE 802.1X WLANs) and links that provide physical security (e.g., enterprise network wired LANs) provide a first line of defense that is often sufficient. In other instances, additional securing mechanisms such as Secure Neighbor Discovery (SeND) [RFC3971], IPsec [RFC4301] or TLS [RFC5246] may be necessary.
AERO Clients MUST ensure that their connectivity is not used by unauthorized nodes on EUNs to gain access to a protected network, i.e., AERO Clients that act as routers MUST NOT provide routing services for unauthorized nodes. (This concern is no different than for ordinary hosts that receive an IP address delegation but then "share" the address with unauthorized nodes via a NAT function.)
On some AERO links, establishment and maintenance of a direct path between neighbors requires secured coordination such as through the Internet Key Exchange (IKEv2) protocol [RFC5996] to establish a security association.
Discussions both on IETF lists and in private exchanges helped shape some of the concepts in this work. Individuals who contributed insights include Mikael Abrahamsson, Fred Baker, Stewart Bryant, Brian Carpenter, Wojciech Dec, Brian Haberman, Joel Halpern, Sascha Hlusiak, Lee Howard, Joe Touch and Bernie Volz. Members of the IESG also provided valuable input during their review process that greatly improved the document. Special thanks go to Stewart Bryant, Joel Halpern and Brian Haberman for their shepherding guidance.
This work has further been encouraged and supported by Boeing colleagues including Keith Bartley, Dave Bernhardt, Cam Brodie, Balaguruna Chidambaram, Wen Fang, Anthony Gregory, Jeff Holland, Ed King, Gen MacLean, Kent Shuey, Brian Skeen, Mike Slane, Julie Wulff, Yueli Yang, and other members of the BR&T and BIT mobile networking teams.
Earlier works on NBMA tunneling approaches are found in [RFC2529][RFC5214][RFC5569].