Configuring Distributed MAP-T

Distributed Scenario

In a distributed scenario, the MAP-BR and BRAS reside on the same device. The device functions as the BRAS to deliver MAP addresses and mapping rules to MAP-CEs in DHCPv6 IA_PD mode. (In this example, only key configurations of IPv6 address assignment are provided. For details, see HUAWEI NetEngine 8000 F Series Router Configuration Guide - User Access - IPv6 Address Management Configuration.) The device also functions as a MAP-BR and resides on the edge of a MAP domain. The device allows the MAP-CEs to access the public IPv4 network through the IPv6 network that is within the MAP domain. In addition, the MAP-CEs can use each other's public IPv4 address to communicate through the MAP-BR.

Figure 1 Distributed MAP-T

Configuration Roadmap

In the distributed scenario, MAP-T is configured on the MAP-BR (BRAS).

Configuring a BMR

This section describes how to configure a basic mapping rule (BMR). Configure BMR rules on the BRAS to instruct the BRAS to assign IPv6 and IPv4 addresses to MAP-CEs.

Configuring a DMR

A default mapping rule (DMR) can be created. A MAP-CE encapsulates an IPv6 prefix defined in the DMR into packets and directs the packets to a service boards. The service boards convert the packets in a MAP-T instance to which the DMR is bound.

Configuring an IPv6 Delegation Prefix Pool

In a MAP-E or MAP-T scenario, a MAP rule needs to be bound to an IPv6 delegation prefix pool. The NetEngine 8000 F uses the BMR configured using the map-rule command to assign prefixes to MAP users.

Configuring an IPv6 Delegation Address Pool

In a MAP-T scenario, a prefix pool, a DMR prefix, and an FMR flag bit need to be configured in an IPv6 delegation address pool.

Binding a DMR to a MAP-T Instance

This section describes how to bind a DMR to a MAP-T instance. The DMR is used by a MAP-CE to select a MAP-T instance to convert IPv6 packets.

Binding a BMR to a MAP-T Instance

This section describes how to bind a BMR to a MAP-T instance. The BMR is used to encapsulate and verify packets in the MAP-T instance.

(Optional) Configuring a Device to Preferentially Assign MAP-T Prefixes to Users

In a MAP-E or MAP-T scenario, if users request both MAP-E and MAP-T options, a device prioritizes the requests and assigns a prefix to a higher-priority MAP option first, and if this attempt fails, to a lower-priority MAP option.

(Optional) Configuring MAP Translation for ICMP/ICMPv6 Error Packets

In a MAP-T scenario, the MAP conversion of ICMP error packets can be performed only after the MAP conversion function is configured for ICMP error packets.

(Optional) Setting an MSS Value for MAP-T Services

The MSS value in the TCP protocol is used to specify the length of a TCP packet that can be transmitted without being fragmented. During the TCP connection establishment, the MSS value is carried in a SYN packet to notify the peer end of the maximum size of a packet that can be received by the local end.

(Optional) Setting an MTU for IPv6 Packets

This section describes how to set a maximum transmission unit. Whether to fragment translated packets depends on the smaller value between the MTU configured in an instance and the interface MTU. To customize the MTU value of the IPv6 packets in a MAP instance, set the MTU value in the MAP instance.

(Optional) Setting the Traffic Class of IPv6 Packets

When an IPv4 packet is translated into an IPv6 packet, the Traffic-Class field value in the IPv6 packet is copied from the ToS field in the IPv4 packet. To modify the traffic class of IPv6 packets, set the traffic class value of IPv6 packets in an instance.

(Optional) Setting an IPv4 ToS Value

When an IPv6 packet is translated into an IPv4 packet, the ToS value in the IPv4 packet is copied from the Traffic-Class field in the IPv6 packet by default. To change the ToS value in the IPv4 packet, set the ToS value of the IPv4 packet in an instance.

(Optional) Clearing the DF Field in IPv4 Packets

The default implementation is as follows. Before a device translates an IPv6 packet into an IPv4 packet, if the IPv6 packet does not carry the fragment extension header and the packet length is less than or equal to 1280 bytes, the DF field in the IPv4 packet is set to 0 (can be fragmented). If the IPv6 packet does not carry the fragment extension header and the packet length is greater than 1280 bytes, the DF field in the IPv4 packet is set to 1 (cannot be fragmented). If the IPv6 packet carries the fragment extension header, the DF field in the IPv4 packet is set to 0 (can be fragmented). After the DF field clearing function is enabled, the device sets the DF field in IPv4 packets to 0.

Checking the Configurations

After configuring basic MAP-T functions, verify the configurations.