burningnode.com
Random networking related stuff.

IPv6 transition – MPLS 6VPE lab

March 11, 2013

Today we are going to work on an IPv6 transition mechanism called 6VPE.

6VPE/6PE uses the underlying IPv4 MPLS/MPLS VPN network to carry IPv6 packets.

To realize this lab, PE routers and route-reflectors need to be IPv6/VPNv6 ready. P routers forward packets based on MPLS information ignoring the layer 3.

Based on LACNIC drawings, here’s a brief explanation
mpls6vpe-drawing

The topology:

MPLS6vpelab

Pros:
– Only RR, PE and CPE need IPv6
– P core routers are left unchanged
– No need to run a dual stack network (IP/IPv6, OSPF/OSPFv3)
– Possible to have dual stack customers
– Non disruptive

Prerequisites:

– A MPLS VPN network
– RR, PE, and CPE IPv6 ready (hardware processing must be supported)

Steps:

– IPv6/VPNv6 compatible MP-BGP between RR1 and all the PEs,

– VRFs v6 for each customers on the PEs,

– IPv6 compatible (and supported…) routing protocol for PE-CE communication. It is also possible to set static routes (redistribute static and connected).

Route reflector configuration (activate IPv6 routing, VPNv6 peering)

ipv6 unicast-routing
!
router bgp 65000
address-family vpnv6
neighbor RRCLIENT send-community both
neighbor RRCLIENT route-reflector-client
neighbor 5.5.5.5 activate
neighbor 6.6.6.6 activate
exit-address-family

PE configuration
ip vrf becomes vrf definition, ip vrf forwarding becomes vrf forwarding
The statement address-family ipv4 in the VRF definition is needed to use the command ping vrf X…

ipv6 unicast-routing
ipv6 cef
!
vrf definition A
 description Customer A
 rd 65000:3
 !
 address-family ipv4
 exit-address-family
 !
 address-family ipv6
 route-target export 65000:3
 route-target import 65000:3
 exit-address-family
!
interface FastEthernet1/0
vrf forwarding A
no ip address
duplex full
speed 100
ipv6 address 2001::1/64
ipv6 address autoconfig
!
router bgp 65000
address-family vpnv6
neighbor 1.1.1.1 activate
neighbor 1.1.1.1 send-community both
exit-address-family
!
address-family ipv6 vrf A
redistribute static
redistribute connected
no synchronization
exit-address-family
!
ipv6 route vrf A 2008::/64 2001::2

CPE configuration

ipv6 unicast-routing
!
interface FastEthernet0/0
 no ip address
 speed 100
 full-duplex
 ipv6 address 2001::2/64
 ipv6 address autoconfig
!
ipv6 route ::/0 2001::1

Let’s have a closer look to the path

First the packet is sent out ACPE2 thanks to default route.

ACPE2#sh ipv route static
IPv6 Routing Table - 7 entries
S   ::/0 [1/0]
     via 2001::1

It enters PE2 in VRF A. The VRF A routing table shows us a recursive lookup inside the default IPv4 routing table.

PE2#sh ipv route vrf A
...
B   2002::/64 [200/0]
     via 6.6.6.6%default, indirectly connected
B   2009::/64 [200/0]
     via 6.6.6.6%default, indirectly connected
...

PE2#sh ip route 6.0.0.0
Routing entry for 6.0.0.0/32, 1 known subnets
O        6.6.6.6 [110/4] via 52.52.52.2, 00:12:43, FastEthernet1/1

To reach the 6.6.6.6 the packet will use MPLS switching (label 28)

PE2#sh mpls forwarding-table 6.6.6.6
Local      Outgoing   Prefix           Bytes Label   Outgoing   Next Hop
Label      Label      or Tunnel Id     Switched      interface
22         28         6.6.6.6/32       0             Fa1/1      52.52.52.2

The MPLS VPN tag is 35

PE2#sh ip bgp vpnv6 unicast all labels
   Network          Next Hop      In label/Out label
Route Distinguisher: 65000:3 (A)
   2001::/64        ::              34/nolabel
   2002::/64        ::FFFF:6.6.6.6  nolabel/34
   2008::/64        2001::2         36/nolabel
   2009::/64        ::FFFF:6.6.6.6  nolabel/35

To sum it up, the CEF entry

PE2#sh ipv cef vrf A
2009::/64
  nexthop 52.52.52.2 FastEthernet1/1 label 28 35

mpls-6vpe-cap-1

On P1 the packet is switched

P1#sh mpls forwarding-table 6.6.6.6
Local  Outgoing    Prefix            Bytes tag  Outgoing   Next Hop
tag    tag or VC   or Tunnel Id      switched   interface
28     26          6.6.6.6/32        164380     Fa2/0      23.23.23.3

mpls-6vpe-cap-2

On P2 the tag is removed due to PHP. Only one tag remains, the VPN tag.

P2#sh mpls forwarding-table 6.6.6.6
Local  Outgoing    Prefix            Bytes tag  Outgoing   Next Hop
tag    tag or VC   or Tunnel Id      switched   interface
26     Pop tag     6.6.6.6/32        201008     Fa1/0      36.36.36.6

mpls-6vpe-cap-3

On PE3 the packet is transfered according the VPN tag:

PE3#sh mpls forwarding-table labels 35
Local      Outgoing   Prefix           Bytes Label   Outgoing   Next Hop
Label      Label      or Tunnel Id     Switched      interface
35         No Label   2009::/64[V]     182312        Fa1/1      2002::2

Troubleshooting commands

ping 2009::9 repeat 30 
traceroute 2009::9

sh ipv6 int brief
sh ipv6 route vrf A

sh ipv6 cef vrf A

sh mpls forwarding-table
sh mpls forwarding-table vrf A
sh mpls forwarding-table labels 35

sh ip bgp vpnv6 unicast all
sh ip bgp vpnv6 unicast vrf A
sh ip bgp vpnv6 unicast all labels

Special thanks to Romain that made me discover this technology.

Links

http://lacnic.net/documentos/seminarios/6PE_6VPE_LACNIC.pdf
http://www.cisco.com/en/US/prod/collateral/iosswrel/ps6537/ps6553/prod_presentation0900aecd80311df4.pdf
http://www.ietf.org/rfc/rfc4659.txt
http://tools.ietf.org/html/rfc4798
http://www.ipflow.utc.fr/index.php/6VPE_-_IPv6_VPN_over_MPLS