From: ccie2be (ccie2be@nyc.rr.com)
Date: Wed Jun 08 2005 - 14:49:40 GMT-3
Thanks, Brian.
I do understand ipv4 proxy arp and how L3 to L2 resolution is done in an
ipv4 network.
But, before getting into the details, does this scenario illustrate that
using a static default route in a large ipv6 network is essentially useless
(assuming no one is going to enter in hundreds or thousands of static L3 to
L2 maps)?
Besides running an ipv6 IGP or entering lots of static L3 to L2 maps, is
there an alternative way of getting this to work?
But, I'm not sure I follow your statement, "The problem with ICMPv6 proxy ND
is that the router would have to join the solicited node multicast
address for the address it is proxying for." Which router, R4 or R6?
I assume you mean R4 would have to join since R4 would proxy on behalf of
R6, right?
Let me see if I can correctly interpret that statement and apply it to the
scenario below step by step.
In this scenario, R6 has a default static ipv6 route pointing to R4. And,
R4 can reach all other ipv6 addresses in the network.
So, when R6 has a packet it wants to send to an ipv6 address on the other
side of R4, R6 just sends the packet to R4. (R6 knows R4's link-local
address through neighbor discovery.) And, this show command proves it.
R6#sh ipv nei
IPv6 Address Age Link-layer Addr State Interface
FE80::203:6BFF:FE21:9700 12 0003.6b21.9700 STALE Ethernet0
2001:144:3:46:203:6BFF:FE21:9700 12 0003.6b21.9700 STALE Ethernet0
But, what is the debug from R6 below saying?
19:24:45: ICMPv6: Sending echo request to 2001:192:10:1:2D0:BAFF:FE7B:F580
19:24:45: IPV6: source 2001:144:3:46:260:83FF:FE7C:969 (local)
19:24:45: dest 2001:192:10:1:2D0:BAFF:FE7B:F580 (Ethernet0)
19:24:45: traffic class 0, flow 0x0, len 100+0, prot 58, hops 64,
originating
19:24:45: IPv6: Resolving next hop 2001:192:10:1:2D0:BAFF:FE7B:F580 on
interface Ethernet0
From the last line above, it looks like R6 resolves the L3 address. (Don't
see any encap failures above.)
Assuming the packet from R6 reaches R4, what happens?
Debugs on R4 don't show it getting any packets from R6, so I'm not sure if
R4 is actually seeing those pings from R6.
TIA, Tim
-----Original Message-----
From: nobody@groupstudy.com [mailto:nobody@groupstudy.com] On Behalf Of
Brian McGahan
Sent: Wednesday, June 08, 2005 12:37 PM
To: Group Study
Subject: RE: ipv6 - proxy ICMP ND not supported
If you understand how layer 3 to layer 2 resolution works on
multipoint interfaces and how IPv4 proxy ARP works then it the same
logic that applies to ICMPv6 proxy ND. The problem with ICMPv6 proxy ND
is that the router would have to join the solicited node multicast
address for the address it is proxying for. This works fine for mobile
IPv6 where you know which address you need to proxy for, but it is not
feasible to proxy for all addresses. For IPv4 proxy ARP this is fine
because the ARP request is broadcast. For ICMPv6 ND the address used is
multicast.
HTH,
Brian McGahan, CCIE #8593
bmcgahan@internetworkexpert.com
Internetwork Expert, Inc.
http://www.InternetworkExpert.com
Toll Free: 877-224-8987 x 705
Outside US: 775-826-4344 x 705
24/7 Support: http://forum.internetworkexpert.com
Live Chat: http://www.internetworkexpert.com/chat/
> -----Original Message-----
> From: nobody@groupstudy.com [mailto:nobody@groupstudy.com] On Behalf
Of
> ccie2be
> Sent: Wednesday, June 08, 2005 10:56 AM
> To: Group Study
> Subject: ipv6 - proxy ICMP ND not supported
>
> Hi guys,
>
> IPv6 is configured between R4 and R6. R4 is running RIPng on it's
other
> interfaces (not shown) but is not running any ipv6 routing protocol on
> it's
> e0/0 int.
>
> R4 e0/0 ------------- e0 R6
>
> R6 is configured with a static default route pointing out it's e0
> interface
> (towards R4).
>
> ipv6 route ::/0 e0
>
> R6#sh ipv6 route
> IPv6 Routing Table - 4 entries
> Codes: C - Connected, L - Local, S - Static, R - RIP, B - BGP
> Timers: Uptime/Expires
>
> L 2001:144:3:46:260:83FF:FE7C:969/128 [0/0]
> via ::, Ethernet0, 03:28:02/never
> C 2001:144:3:46:260:83FF:FE7C:969/64 [0/0]
> via ::, Ethernet0, 03:28:02/never
> L FE80::/64 [0/0]
> via ::, Null0, 03:28:02/never
> S ::/0 [1/0]
> via ::, Ethernet0, 03:28:00/never
>
> Yet, in order for R6 to reach other ipv6 addresses in the network, it
> needs
> to have lots of static L3 to L2 maps configured like this.
>
> ipv6 neighbor 2001:144:3:24:2D0:BAFF:FEE3:2CA0 e0 0002:165a:52a0
>
> BTW, there's no special about e0's ipv6 config:
>
> interface Ethernet0
> ip address 144.3.46.6 255.255.255.0
> ipv6 address 2001:144:3:46::/64 eui-64
>
> The explanation given is that those entries are required because ipv6
> doesn't support proxy ICMP ND.
>
> What does that mean in plain English?
>
> And, doesn't that negate the whole idea of using a default static
route?
>
> And, does this also mean that whenever a default static route is
> configured
> in ipv6, a static L3 to L2 map will be required for every ipv6
destination
> that needs to be reached or is there something in particular about
this
> scenario that makes this required?
>
> Besides running an ipv6 IGP between R4 and R6, is there any other
> alternative to having a static L3 to L2 map for each destination?
>
> Sorry for all the questions but I truly don't follow the logic at work
> here.
>
> TIA, Tim
>
>
This archive was generated by hypermail 2.1.4 : Wed Jul 06 2005 - 14:43:41 GMT-3