Nokia Siemens Networks

6000 Connection Drive Irving TX75039 USA basavaraj.patil@nsn.com

Huawei USA

1700 Alma Dr. Suite 500 Plano TX 75075 USA xiayangsong@huawei.com

Huawei USA

1700 Alma Dr. Suite 500 Plano TX 75075 USA sarikaya@ieee.org

Samsung AIT

Networking Technology Lab P.O.Box 111 Suwon Korea 440-600 jinchoe@samsung.com

Ordyn Technologies

1st Floor, Creator Building, ITPL. Off Airport Road Bangalore India 560066 smadanapalli@gmail.com

Internet Network Working Group Neighbor Discovery Per-MS Perfix IEEE Std 802.16 is an air interface specification for fixed and mobile Broadband Wireless Access Systems. Service-specific convergence sublayers to which upper-layer protocols interface are a part of the IEEE 802.16 MAC (Medium Access Control). The Packet convergence sublayer (CS) is used for the transport of all packet-based protocols such as Internet Protocol (IP) and IEEE 802.3 LAN/MAN CSMA/CD Access Method (Ethernet). IPv6 packets can be sent and received via the IP-specific part of the Packet CS. This document specifies the addressing and operation of IPv6 over the IP-specific part of the Packet CS for hosts served by a network that utilizes the IEEE Std 802.16 air interface. It recommends the assignment of a unique prefix (or prefixes) to each host and allows the host to use multiple identifiers within that prefix, including support for randomly generated interface identifiers.

IEEE 802.16e is an air interface for fixed and mobile broadband wireless access systems. The IEEE 802.16 standard specifies the air interface, including the Medium Access Control (MAC) layer and multiple physical layer (PHY) specifications. It can be deployed in licensed as well as unlicensed spectrum. While the PHY and MAC are specified in IEEE 802.16, the details of IPv4 and IPv6 operation over the air interface are not included. This document specifies the operation of IPv6 over the IEEE 802.16 air interface. IPv6 packets can be carried over the IEEE Std 802.16 specified air interface via: the IP-specific part of the Packet CS or the 802.3-specific part of the Packet CS The scope of this specification is limited to the operation of IPv6 over IP CS only. The IEEE 802.16 specification includes the PHY and MAC details. The convergence sublayers are a part of the MAC. The packet convergence sublayer includes the IP-specific part that is used by the IPv6 layer. The mobile station (MS)/host is attached to an access router via a base station (BS). The host and the BS are connected via the IEEE Std 802.16 air interface at the link and physical layers. The IPv6 link from the MS terminates at an access router that may be a part of the BS or an entity beyond the BS. The base station is a layer 2 entity (from the perspective of the IPv6 link between the MS and access router (AR)) and relays the IPv6 packets between the AR and the host via a point-to-point connection over the air interface.

The terminology in this document is based on the definitions in "IP over 802.16 Problem Statement and Goals". IP CS - The IP-specific part of the Packet convergence sublayer is referred to as IP CS. IPv6 CS and IP CS are used interchangeably. Subscriber station (SS), Mobile Station (MS), Mobile Node (MN) - The terms subscriber station, mobile station, and mobile node are used interchangeably in this document and mean the same, i.e., an IP host.

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 RFC 2119.

The IEEE 802.16 MAC specifies two main service-specific convergence sublayers: ATM convergence sublayer Packet convergence sublayer The Packet CS is used for the transport of packet-based protocols, which include: IEEE Std 802.3(Ethernet) Internet Protocol (IPv4 and IPv6) The service-specific CS resides on top of the MAC Common Part Sublayer (CPS) as shown in Figure 1. The service-specific CS is responsible for: accepting packets (Protocol Data Units, PDUs) from the upper layer, performing classification of the packet/PDU based on a set of defined classifiers that are service specific, delivering the CS PDU to the appropriate service flow and transport connection, and receiving PDUs from the peer entity. Payload header suppression (PHS) is also a function of the CS but is optional. The figure below shows the concept of the service-specific CS in relation to the MAC:

Classifiers for each of the specific upper-layer protocols, i.e., Ethernet and IP, are defined in the IEEE 802.16 specification, which enable the packets from the upper layer to be processed by the appropriate service-specific part of the Packet CS. IPv6 can be transported directly over the IP-specific part of the Packet CS (IP CS). IPv4 packets also are transported over the IP-specific part of the Packet CS. The classifiers used by IP CS enable the differentiation of IPv4 and IPv6 packets and their mapping to specific transport connections over the air interface. The figure below shows the options for IPv6 transport over the packet CS of IEEE 802.16:

The figure above shows that while there are multiple methods by which IPv6 can be transmitted over an 802.16 air interface, the scope of this document is limited to IPv6 operation over IP CS only. Transmission of IP over Ethernet is specified in . Transmission of IPv4 over IP CS is specified in . It should be noted that immediately after ranging (802.16 air interface procedure) and exchange of SBC-REQ/RSP messages (802.16 specific), the MS and BS exchange their capabilities via REG-REQ (Registration Request) and REG-RSP (Registration Response) 802.16 MAC messages. These management frames negotiate parameters such as the Convergence Sublayer supported by the MS and BS. By default, Packet, IPv4, and 802.3/Ethernet are supported. IPv6 via the IP CS is supported by the MS and the BS only when the IPv6 support bit in the capability negotiation messages (REG-REQ and REG-RSP) implying such support is indicated in the parameter "Classification/PHS options and SDU (Service Data Unit) encapsulation support" (refer to ). Additionally, during the establishment of the transport connection for transporting IPv6 packets, the DSA-REQ (Dynamic Service Addition) and DSA-RSP messages between the BS and MS indicate via the CS-Specification TLV the CS that the connection being set up shall use. When the IPv6 packet is preceded by the IEEE 802.16 6-byte MAC header, there is no specific indication in the MAC header itself about the payload type. The processing of the packet is based entirely on the classifiers. Based on the classification rules, the MAC layer selects an appropriate transport connection for the transmission of the packet. An IPv6 packet is transported over a transport connection that is specifically established for carrying such packets. Transmission of IPv6 as explained above is possible via multiple methods, i.e., via IP CS or via Ethernet interfaces. Every Internet host connected via an 802.16 link: MUST be able to send and receive IPv6 packets via IP CS when the MS and BS indicate IPv6 protocol support over IP CS MUST be able to send and receive IPv6 packets over the Ethernet (802.3)-specific part of the Packet CS when the MS and BS indicate IPv6 protocol support over Ethernet CS. However, when the MS and BS indicate IPv6 protocol support over both IP CS and Ethernet CS, the MS and BS MUST use IP CS for sending and receiving IPv6 packets. When the MS and BS support IPv6 over IP CS, it MUST be used as the default mode for transporting IPv6 packets over IEEE 802.16 and the recommendations in this document that are followed. Inability to negotiate a common convergence sublayer for IPv6 transport between the MS and BS will result in failure to set up the transport connection and thereby render the host unable to send and receive IPv6 packets. In the case of a host that implements more than one method of transporting IPv6 packets, the default choice of which method to use (i.e., IPv6 over the IP CS or IPv6 over 802.3) is IPv6 over IP CS when the BS also supports such capability. In any case, the MS and BS MUST negotiate at most one convergence sublayer for IPv6 transport on a given link. In addition, to ensure interoperability between devices that support different encapsulations, it is REQUIRED that BS implementations support all standards-track encapsulations defined for 802.16 by the IETF. At the time of writing this specification, this is the only encapsulation, but additional specifications are being worked on. It is, however, not required that the BS implementations use all the encapsulations they support; some modes of operation may be off by configuration.

The IPv6 payload when carried over the IP-specific part of the Packet CS is encapsulated by the 6-byte IEEE 802.16 generic MAC header. The format of the IPv6 packet encapsulated by the generic MAC header is shown in the figure below. The format of the 6-byte MAC header is described in the specification. The CRC (cyclic redundancy check) is optional. It should be noted that the actual MAC address is not included in the MAC header.

For transmission of IPv6 packets via the IP CS over IEEE 802.16, the IPv6 layer interfaces with the 802.16 MAC directly. The IPv6 layer delivers the IPv6 packet to the Packet CS of the IEEE 802.16 MAC. The Packet CS defines a set of classifiers that are used to determine how to handle the packet. The IP classifiers that are used at the MAC operate on the fields of the IP header and the transport protocol, and these include the IP Traffic class, Next header field, Masked IP source and destination addresses, and Protocol source and destination port ranges. Next header in this case refers to the last header of the IP header chain. Parsing these classifiers, the MAC maps an upper-layer packet to a specific service flow and transport connection to be used. The MAC encapsulates the IPv6 packet in the 6-byte MAC header (MAC SDU) and transmits it. The figure below shows the operation on the downlink, i.e., the transmission from the BS to the host. The reverse is applicable for the uplink transmission.

[CID1] | | --- |-------- | || Downlink 0\/-->[CID2] | | |Reconstruct| | || classifiers0/\-->[....] | | | (undo PHS)| | || 0---->[CIDn] | | --- ------- | ||--------------| | | /|\ | | | | | | | {SDU, CID,..} | | {SDU, CID,..} | | | | | /|\ | | v | | | | ------[SAP]----------------- |-------[SAP]--------- | 802.16 MAC CPS |------>| 802.16 MAC CPS | ---------------------------- ---------------------- BS MS ]]>

In a network that utilizes the 802.16 air interface, the host/MS is attached to an IPv6 access router (AR) in the network. The BS is a layer 2 entity only. The AR can be an integral part of the BS or the AR could be an entity beyond the BS within the access network. An AR may be attached to multiple BSs in a network. IPv6 packets between the MS and BS are carried over a point-to-point transport connection which is identified by a unique Connection Identifier (CID). The transport connection is a MAC layer link between the MS and the BS. The figures below describe the possible network architectures and are generic in nature. More esoteric architectures are possible but not considered in the scope of this document. Option A:

Option B:

The above network models serve as examples and are shown to illustrate the point-to-point link between the MS and the AR.

"Neighbor Discovery for IP Version 6 (IPv6)" defines link as a communication facility or medium over which nodes can communicate at the link layer, i.e., the layer immediately below IP. A link is bounded by routers that decrement the Hop limit field in the IPv6 header. When an MS moves within a link, it can keep using its IP addresses. This is a layer 3 definition, and note that the definition is not identical with the definition of the term '(L2) link' in IEEE 802 standards.

In 802.16, the transport connection between an MS and a BS is used to transport user data, i.e., IPv6 packets in this case. A transport connection is represented by a CID, and multiple transport connections can exist between an MS and a BS. When an AR and a BS are colocated, the collection of transport connections to an MS is defined as a single link. When an AR and a BS are separated, it is recommended that a tunnel be established between the AR and a BS whose granularity is no greater than 'per MS' or 'per service flow' (An MS can have multiple service flows which are identified by a service flow ID). Then the tunnel(s) for an MS, in combination with the MS's transport connections, forms a single point-to-point link. The collection of service flows (tunnels) to an MS is defined as a single link. Each link that uses the same higher-layer protocol has only an MS and an AR. Each MS belongs to a different link. A different prefix should be assigned to each unique link. This link is fully consistent with a standard IP link, without exception, and conforms with the definition of a point-to-point link in neighbor discovery for IPv6 . Hence, the point-to-point link model for IPv6 operation over the IP-specific part of the Packet CS in 802.16 SHOULD be used. A unique IPv6 prefix(es) per link (MS/host) MUST be assigned.

In order to enable the sending and receiving of IPv6 packets between the MS and the AR, the link between the MS and the AR via the BS needs to be established. This section illustrates the link establishment procedure. The MS goes through the network entry procedure as specified by 802.16. A high-level description of the network entry procedure is as follows: The MS performs initial ranging with the BS. Ranging is a process by which an MS becomes time aligned with the BS. The MS is synchronized with the BS at the successful completion of ranging and is ready to set up a connection. The MS and BS exchange basic capabilities that are necessary for effective communication during the initialization using SBC-REQ/RSP (802.16 specific) messages. The MS progresses to an authentication phase. Authentication is based on Privacy Key Management version 2 (PKMv2) as defined in the IEEE Std 802.16 specification. On successful completion of authentication, the MS performs 802.16 registration with the network. The MS and BS perform capability exchange as per 802.16 procedures. Protocol support is indicated in this exchange. The CS capability parameter indicates which classification/PHS options and SDU encapsulation the MS supports. By default, Packet, IPv4, and 802.3/Ethernet shall be supported; thus, absence of this parameter in REG-REQ (802.16 message) means that named options are supported by the MS/SS. Support for IPv6 over the IP-specific part of the Packet CS is indicated by Bit #2 of the CS capability parameter (refer to ). The MS MUST request the establishment of a service flow for IPv6 packets over IP CS if the MS and BS have confirmed capability for supporting IPv6 over IP CS. The service flow MAY also be triggered by the network as a result of pre-provisioning. The service flow establishes a link between the MS and the AR over which IPv6 packets can be sent and received. The AR and MS SHOULD send router advertisements and solicitations as specified in neighbor discovery . The above flow does not show the actual 802.16 messages that are used for ranging, capability exchange, or service flow establishment. Details of these are in .

The MTU value for IPv6 packets on an 802.16 link is configurable. The default MTU for IPv6 packets over an 802.16 link SHOULD be 1500 octets. The 802.16 MAC PDU is composed of a 6-byte header followed by an optional payload and an optional CRC covering the header and the payload. The length of the PDU is indicated by the Len parameter in the Generic MAC header. The Len parameter has a size of 11 bits. Hence, the total MAC PDU size is 2048 bytes. The IPv6 payload size can vary. In certain deployment scenarios, the MTU value can be greater than the default. Neighbor discovery for IPv6 [RFC4861] defines an MTU option that an AR MUST advertise, via router advertisement (RA), if a value different from 1500 is used. The MN processes this option as defined in . Nodes that implement Path MTU Discovery MAY use the mechanism to determine the MTU for the IPv6 packets.

The MS and the AR are connected via a point-to-point connection at the IPv6 layer. Hence, each MS can be considered to be on a separate subnet. A CPE (Customer Premise Equipment) type of device that serves multiple IPv6 hosts may be the end point of the connection. Hence, one or more /64 prefixes SHOULD be assigned to a link. The prefixes are advertised with the on-link (L-bit) flag set as specified in . The size and number of the prefixes are a configuration issue. Also, Dynamic Host Configuration Protocol (DHCP) or Authentication, Authorization, and Accounting (AAA)-based prefix delegation MAY be used to provide one or more prefixes to MS for an AR connected over 802.16. The other properties of the prefixes are also dealt with via configuration.

On completion of the establishment of the IPv6 link, the MS may send a router solicitation message to solicit a router advertisement message from the AR to acquire necessary information as per the neighbor discovery for IPv6 specification . An MS that is network attached may also send router solicitations at any time. Movement detection at the IP layer of an MS in many cases is based on receiving periodic router advertisements. An MS may also detect changes in its attachment via link triggers or other means. The MS can act on such triggers by sending router solicitations. The router solicitation is sent over the IPv6 link that has been previously established. The MS sends router solicitations to the all-routers multicast address. It is carried over the point-to-point link to the AR via the BS. The MS does not need to be aware of the link-local address of the AR in order to send a router solicitation at any time. The use of router advertisements as a means for movement detection is not recommended for MNs connected via 802.16 links as the frequency of periodic router advertisements would have to be high.

The AR SHOULD send a number (configurable value) of router advertisements to the MS as soon as the IPv6 link is established. The AR sends unsolicited router advertisements periodically as per . The interval between periodic router advertisements is however greater than the specification in neighbor discovery for IPv6, and is discussed in the following section.

The router lifetime SHOULD be set to a large value, preferably in hours. This document overrides the specification for the value of the router lifetime in "Neighbor Discovery for IP Version 6 (IPv6)" . The AdvDefaultLifetime in the router advertisement MUST be either zero or between MaxRtrAdvInterval and 43200 seconds. The default value is 2 * MaxRtrAdvInterval. 802.16 hosts have the capability to transition to an idle mode, in which case, the radio link between the BS and MS is torn down. Paging is required in case the network needs to deliver packets to the MS. In order to avoid waking a mobile that is in idle mode and consuming resources on the air interface, the interval between periodic router advertisements SHOULD be set quite high. The MaxRtrAdvInterval value specified in this document overrides the recommendation in "Neighbor Discovery for IP Version 6 (IPv6)". The MaxRtrAdvInterval MUST be no less than 4 seconds and no greater than 21600 seconds. The default value for MaxRtrAdvInterval is 10800 seconds.

The addressing scheme for IPv6 hosts in 802.16 networks follows the IETF's recommendation for hosts specified in "IPv6 Node Requirements" . The IPv6 node requirements specify a set of RFCs that are applicable for addressing, and the same is applicable for hosts that use 802.16 as the link layer for transporting IPv6 packets.

The MS has a 48-bit globally unique MAC address as specified in 802.16. This MAC address MUST be used to generate the modified EUI-64 format-based interface identifier as specified in "IP Version 6 Addressing Architecture" . The modified EUI-64 interface identifier is used in stateless address autoconfiguration. As in other links that support IPv6, EUI-64-based interface identifiers are not mandatory and other mechanisms, such as random interface identifiers, "Privacy Extensions for Stateless Address Autoconfiguration in IPv6" , MAY also be used.

DAD SHOULD be performed as per "Neighbor Discovery for IP Version 6 (IPv6)", and "IPv6 Stateless Address Autoconfiguration" . The IPv6 link over 802.16 is specified in this document as a point-to-point link. Based on this criteria, it may be redundant to perform DAD on a global unicast address that is configured using the EUI-64 or generated as per RFC 4941 for the interface as part of the IPv6 Stateless Address Autoconfiguration Protocol as long as the following two conditions are met: The prefixes advertised through the router advertisement messages by the access router terminating the 802.16 IPv6 link are unique to that link. The access router terminating the 802.16 IPv6 link does not autoconfigure any IPv6 global unicast addresses from the prefix that it advertises.

When stateless address autoconfiguration is performed, it MUST be performed as specified in and .

When stateful address autoconfiguration is performed, it MUST be performed as specified in and .

"Multicast Listener Discovery Version 2 (MLDv2) for IPv6" SHOULD be supported as specified by the hosts and routers attached to each other via an 802.16 link. The access router that has hosts attached to it via a point-to-point link over an 802.16 SHOULD NOT send periodic queries if the host is in idle/dormant mode. The AR can obtain information about the state of a host from the paging controller in the network.

This document does not introduce any new vulnerabilities to IPv6 specifications or operation. The security of the 802.16 air interface is the subject of . It should be noted that 802.16 provides capability to cipher the traffic carried over the transport connections. A traffic encryption key (TEK) is generated by the MS and BS on completion of successful authentication and is used to secure the traffic over the air interface. An MS may still use IPv6 security mechanisms even in the presence of security over the 802.16 link. In addition, the security issues of the network architecture spanning beyond the 802.16 base stations are the subject of the documents defining such architectures, such as WiMAX Network Architecture in Sections 7.2 and 7.3 of Stage 2, Part 2.

The authors would like to acknowledge the contributions of the 16NG working group chairs Soohong Daniel Park and Gabriel Montenegro as well as Jari Arkko, Jonne Soininen, Max Riegel, Prakash Iyer, DJ Johnston, Dave Thaler, Bruno Sousa, Alexandru Petrescu, Margaret Wasserman, and Pekka Savola for their review and comments. Review and comments by Phil Barber have also helped in improving the document quality.

This document specifies problems in running the IETF IP protocols over IEEE 802.16 networks by identifying specific gaps in the 802.16 MAC for IPv4 and IPv6 support. This document also provides an overview of IEEE 802.16 network characteristics and convergence sublayers. Common terminology used for the base guideline while defining the solution framework is also presented. IEEE 802.16 is an air interface specification for wireless broadband access. IEEE has specified the service specific convergence sublayers (CS) in the IEEE 802.16 MAC to be used by upper layer protocols. Asynchronous Transfer Mode Convergence Sublayer (ATM CS) and Packet Convergence Sublayer (Packet CS) represent the two main service specific convergence sublayers for the IEEE 802.16. The packet CS is used for transport for all packet-based protocols such as Internet Protocol (IP), IEEE 802.3 (Ethernet) and IEEE 802.1Q (VLAN). The IP specific part of the Packet CS enables transport of IPv4 packets directly over the IEEE 802.16 MAC. This document specifies the frame format, the Maximum Transmission Unit (MTU) and address assignment procedures for transmitting IPv4 packets over IP Convergence Sublayer (IPCS) of the IEEE 802.16. This document also provides the details of why the ARP cannot be sent over the IEEE 802.16 links using IPCS and a recommendation for this. This document describes the transmission of IPv4 over Ethernet as well as IPv6 over Ethernet in an access network deploying the IEEE 802.16 cellular radio transmission technology. The Ethernet on top of IEEE 802.16 is realized by bridging between point-to-point radio links, which are provided by IEEE 802.16 between a base station and its associated subscriber stations. Due to the resource constraints of radio transmission systems and the limitations of the IEEE 802.16 MAC functionality for the realization of an Ethernet, the transmission of IP over Ethernet over IEEE 802.16 may considerably benefit by adding IP specific support functions in the Ethernet over IEEE 802.16 while maintaining full compatibility with standard IP over Ethernet behavior.

The WiMAX (Worldwide Interoperability for Microwave Access) forum has defined a network architecture in which the air interface is based on the IEEE 802.16 standard. The addressing and operation of IPv6 described in this document are applicable to the WiMAX network as well. WiMAX is an example architecture of a network that uses the 802.16 specification for the air interface. WiMAX networks are also in the process of being deployed in various parts of the world, and the operation of IPv6 within a WiMAX network is explained in this appendix. The WiMAX network architecture consists of the Access Service Network (ASN) and the Connectivity Service Network (CSN). The ASN is the access network that includes the BS and the AR in addition to other functions such as AAA, mobile IP foreign agent, paging controller, location register, etc. The ASN is defined as a complete set of network functions needed to provide radio access to a WiMAX subscriber. The ASN is the access network to which the MS attaches. The IPv6 access router is an entity within the ASN. The term ASN is specific to the WiMAX network architecture. The CSN is the entity that provides connectivity to the Internet and includes functions such as mobile IP home agent and AAA. The figure below shows the WiMAX reference model:

Three different types of ASN realizations called profiles are defined by the architecture. ASNs of profile types A and C include BS' and ASN-gateway(s) (ASN-GW), which are connected to each other via an R6 interface. An ASN of profile type B is one in which the functionality of the BS and other ASN functions are merged together. No ASN-GW is specifically defined in a profile B ASN. The absence of the R6 interface is also a profile B specific characteristic. The MS at the IPv6 layer is associated with the AR in the ASN. The AR may be a function of the ASN-GW in the case of profiles A and C and is a function in the ASN in the case of profile B.  When the BS and the AR are separate entities and linked via the R6 interface, IPv6 packets between the BS and the AR are carried over a Generic Routing Encapsulation (GRE) tunnel. The granularity of the GRE tunnel should be on a per-MS basis or on a per-service-flow basis (an MS can have multiple service flows, each of which is identified uniquely by a service flow ID). The protocol stack in WiMAX for IPv6 is shown below:

| | --------------- -------/ | | IPv6| |-------| | \Relay/ | | | |- - - | | | | | \ / | | GRE | | | | | | | \ /GRE | - | | | | | | |- - - | |-----| |------| | | | | IPv6CS| |IPv6CS | IP | - | IP | | | | | ..... | |...... |-----| |------|--------| |-----| | MAC | | MAC | L2 | - | L2 | L2 |- - - | L2 | |-------| |------ |-----| |----- |--------| |-----| | PHY |- - - | PHY | L1 | - | L1 | L1 |- - - | L1 | -------- --------------- ----------------- ------- MS BS AR/ASN-GW CSN Rtr ]]>

As can be seen from the protocol stack description, the IPv6 end-points are constituted in the MS and the AR. The BS provides lower-layer connectivity for the IPv6 link.

WiMAX is an example of a network based on the IEEE Std 802.16 air interface. This section describes the IPv6 link in the context of a WiMAX network. The MS and the AR are connected via a combination of: The transport connection that is identified by a Connection Identifier (CID) over the air interface, i.e., the MS and BS, and A GRE tunnel between the BS and AR that transports the IPv6 packets From an IPv6 perspective, the MS and the AR are connected by a point-to-point link. The combination of transport connection over the air interface and the GRE tunnel between the BS and AR creates a (point-to-point) tunnel at the layer below IPv6. The collection of service flows (tunnels) to an MS is defined as a single link. Each link has only an MS and an AR. Each MS belongs to a different link. No two MSs belong to the same link. A different prefix should be assigned to each unique link. This link is fully consistent with a standard IP link, without exception, and conforms with the definition of a point-to-point link in .

The mobile station performs initial network entry as specified in 802.16. On successful completion of the network entry procedure, the ASN gateway/AR triggers the establishment of the initial service flow (ISF) for IPv6 towards the MS. The ISF is a GRE tunnel between the ASN-GW/AR and the BS. The BS in turn requests the MS to establish a transport connection over the air interface. The end result is a transport connection over the air interface for carrying IPv6 packets and a GRE tunnel between the BS and AR for relaying the IPv6 packets. On successful completion of the establishment of the ISF, IPv6 packets can be sent and received between the MS and AR. The ISF enables the MS to communicate with the AR for host configuration procedures. After the establishment of the ISF, the AR can send a router advertisement to the MS. An MS can establish multiple service flows with different quality of service (QoS) characteristics. The ISF can be considered as the primary service flow. The ASN-GW/AR treats each ISF, along with the other service flows to the same MS, as a unique link that is managed as a (virtual) interface.

The WiMAX forum has specified the Max SDU size as 1522 octets. Hence, the IPv6 path MTU can be 1500 octets. However, because of the overhead of the GRE tunnel used to transport IPv6 packets between the BS and AR and the 6-byte MAC header over the air interface, using a value of 1500 would result in fragmentation of packets. It is recommended that the MTU for IPv6 be set to 1400 octets in WiMAX networks, and this value (different from the default) be communicated to the MS. Note that the 1522-octet specification is a WiMAX forum specification and not the size of the SDU that can be transmitted over 802.16, which has a higher limit.