Re: Running Routing Protocol over 6to4 Tunnels

Hi,

Like Bob said previously, you cannot run an IGP over a 6to4 tunnel due
to the fact that IGP protocols cannot interpret it as a single link.
Distance-vector protocols e.g. RIPng/EIGRPv6 use link-local address
for the next-hop information and OSPFv3 needs to map the prefixes to
the link-local addresses using link LSAs. It all has to deal with IPv6
semantics of local link reachability and ND.

However, it is possible to run MP-BGP across 6to4 tunnels for dynamic
routing, as MP-BGP extension for IPv6 allows for using a global IPv6
address for the nex-hop in MP_REACH_NLRI attribute. Simply use the
6to4 tunnel addresses to establish BGP peering and dont forget eBGP
multihop if you're running external sessions. The BGP next-hop
recursive lookup will properly resolve and install learned prefixes in
the RIB with the next-hop of the advertising peer.

In fact, this is yet another solution for running IPv6 on top of IPv4
only core SP network, though using 6PE or 6VPE is probably a better
way of doing that in MPLS-enabled networks.

Regards,

Petr

2010/6/11 Ladee Geek <ladeegeek_at_gmail.com>:
> I'm wrestling with running a routing protocol over 6to4 tunnels. I can ping
> from tunnel end point to end point, but can't get to any other 2002::/16
> addresses I've created. That was based on the static route 2002::/16 out
> the tunnel interface. So I am trying a dynamic routing protocol, OSPF in
> p-mp nbma mode. I have configured the neighbor statement so there's a next
> hop address, but I don't get a neighbor relationship.
>
> When pining the other end of the tunnel I actually see icmp traffic. But
> when I ping another 2002 address I don't see anything. I have the static
> 2002::/16 tunnel 345.
>
> So my static didn't work and neither did a dynamic protocol to get to 2002
> addresses other than the tunnel end points.
>
> What am I missing?
>
>
> Topology: R3, R4 and R5 are set up with 6-to-4 tunnels.
>
> *Rack1R3*(config-if)#do sh run int tun 345
> Building configuration...
> Current configuration : 267 bytes
> !
> interface Tunnel345
> no ip address
> no ip redirects
> ipv6 address 2002:9B01:2503::/64 eui-64
> ipv6 ospf network point-to-multipoint non-broadcast
> ipv6 ospf neighbor FE80::9B01:9204
> ipv6 ospf 1 area 345
> tunnel source FastEthernet0/0
> tunnel mode ipv6ip 6to4
> *Rack1R3(config-if)#do pin 2002:9B01:3A05::9B01:3A05 - this is the fastether
> of R5*
> **
> Type escape sequence to abort.
> Sending 5, 100-byte ICMP Echos to 2002:9B01:3A05::9B01:3A05, timeout is 2
> seconds:
> !!!!!
> Success rate is 100 percent (5/5), round-trip min/avg/max = 60/78/140 ms
>
> *Rack1R3(config)#do sirv | se 2002*
> S 2002::/16 [1/0]
> via Tunnel345, directly connected
>
>
> *Rack1R5(config)#do sh run int tun 345
> *Building configuration...
> Current configuration : 156 bytes
> !
> interface Tunnel345
> no ip address
> no ip redirects
> ipv6 address 2002:9B01:3A05::/64 eui-64
> tunnel source FastEthernet0/0
> tunnel mode ipv6ip 6to4
> end
> *Rack1R5(config)#do sh run int lo 105*
> Building configuration...
> Current configuration : 80 bytes
> !
> interface Loopback105
> no ip address
> ipv6 address 2002:9601:505:1::5/64
>
>
> *Rack1R3(config)#do ping 2002:9601:505::1*
> Type escape sequence to abort.
> Sending 5, 100-byte ICMP Echos to 2002:9601:505::1, timeout is 2 seconds:
> .....
> Success rate is 0 percent (0/5)
> Rack1R3(config)#
>
> *Rack1R5(config-if)#do shipv*
> Loopback0 [up/up]
> FE80::21A:6CFF:FE9F:9FA8
> * 2002:9601:505::1
> *
> --
> r/
> LG
>
>
> Blogs and organic groups at http://www.ccie.net
>
> _______________________________________________________________________
> Subscription information may be found at:
> http://www.groupstudy.com/list/CCIELab.html

Blogs and organic groups at http://www.ccie.net
Received on Fri Jun 11 2010 - 20:26:13 ART