From: ashwin kohli (ashwin_kohli1@yahoo.com)
Date: Fri Aug 12 2005 - 17:07:58 GMT-3
Hi gladston,
I will try to explain to you IPv6 addressing in a step by step format:
Step 1) IPV6 aggregatable global unicast addresses, also known as global addresses,
are very similar to IPv4 global unicast addresses. These addresses are used
for hosts which have interfaces connected to the public Internet, where each
address has to be unique and routable over the Internet. Addresses of this type
are designed or summarized to produce an efficient routing infrastructure.
Step 2) Aggregatable global unicast addresses starts from 2000:: /3 (001 in binary)
through to E000:: /3 (111 in binary) .
Step 3) The IANA is responsible for allocating 2000:: /16 to the ISPs.
The following link - http://www.iana.org/assignments/ipv6-unicast-address-assignments
shows how IANA has allocated the Aggregatable global unicast addresses to the
different global address registries. For americas (i.e. ARIN) the aggregatable global
address space has been broken down to - 2001:0400::/23 (as in your example).
Step 4) IPv4 addresses are broken down into classes A, B, C, D and E. IPv4 address
pool can be further broken down using VLSM. In a similar way, IPv6 have a
structure also. The specific type of an IPv6 address is indicated by the
leading bits in the address. The variable-length field comprising these leading
bits is called the Format Prefix (FP).
Step 5) ARIN will then "VLSM" 2001:0400::/23 into different subnets which will be given to different ISPs in the US. Lets say, your ISP has three different supernets you listed:
1) 2001:0420:0001:1:/48
2) 2001:0420:0001:2:/48
3) 2001:0420:0001:3:/48
Step 6) To arrive at a fully aggregatable global address, you have to take these 48 bits
and attach another 16-bit subnet ID called - Site-Level Aggregators (SLA). A 16-bit
subnet field called the Subnet ID could be used by individual organizations to
create their own local addressing hierarchy and to identify subnets. A subnet ID
is similar to a subnet in IPv4, except that an organization with an IPv6 subnet
ID can support up to 65,535 (2 to the power 16) individual subnets.
Step 7) A physical interface on router which is being allocated with a aggregatable global address, will be given a /64 (48 + 16) subnet altogehter. This is how you arrive at the EUI-64 address format.
Step 8) The addressing you are giving to your routers, you are not adding the extra 16-bit subnets i.e. SLAs to your addressing. Hence the router is giving you an error in your Test1.
Therefore to answer your queston try this (i.e. change your subnets to 64 bits):
Rack2R3(config)#int lo 1001
Rack2R3(config-if)#ipv6 address 2001:0420:0001:1::3/64
Rack2R3(config-if)#int lo 1002
Rack2R3(config-if)#ipv6 address 2001:0420:0001:2::3/64
Let me know if you have any further questions and the above made sense :)
Thanks
Ashwin Kohli
CCIE#8877 (R&S, Security)
http://www.ccbootcamp.com/nliipv6labguide.asp
http://www.ccbootcamp.com/nlicciecore.asp
gladston@br.ibm.com wrote:
Hi,
This example seems to be wrong. What do you think?
====================================================
For example, an NLA provider with the address space of 2001:0420:0001::/48 can allocate the following address space:
2001:0420:0001:1:/48
2001:0420:0001:2:/48
2001:0420:0001:3:/48
quoted from:
Configuring IPv6 with Cisco IOS
by Sam Brown
===================================================
The /48 for the customers...it is not right.
It should be other than /48 (for example /64 or /52) to the customer, otherwise it overlaps.
Agree?
The book is giving an example of TLA, sub-TLA, and NLA prefix allocation.
TLA-1 provides this prefix to sub-TLA1:
2001:0420::/35 - sub-TLA1
sub-TLA-1 provides this prefix to NLA-x:
2001:0428:0001::/48
NLA-x provides these prefixes to its customers:
customer 1 2001:0420:0001:1:/48
customer 2 2001:0420:0001:2:/48
Test 1:
Rack2R3(config)#int lo 1001
Rack2R3(config-if)#ipv6 address 2001:0420:0001:1::3/48
Rack2R3(config-if)#int lo 1002
Rack2R3(config-if)#ipv6 address 2001:0420:0001:2::3/48
% Prefix 2001:420:1:2::3/48 overlaps with another prefix
Test 2:
On this test, R5 is the NLA and R2/R3 are customers.
Rack2R3#
interface Loopback1001
no ip address
ipv6 address 2001:420:1:1::3/48
ipv6 address 2001:420:1:2::3/48
ipv6 rip RIPng enable
Rack2R3#
interface Loopback1001
no ip address
ipv6 address 2001:420:1:3::2/48
ipv6 rip RIPng enable
Rack2R5#sh ipv6 route rip | b 2001:420:1:
R 2001:420:1::/48 [120/2]
via FE80::2, Serial0/0
via FE80::3, Serial0/0
Rack2R5#pi 2001:0420:0001:3::2
debug:
Rack2R3#
Aug 12 10:41:20: IPV6: source 2001:148:5:235::5 (Serial1/0.235)
Aug 12 10:41:20: dest 2001:420:1:3::2 (Loopback1001)
Rack2R2#
Aug 12 10:41:22: IPV6: source 2001:148:5:235::5 (Serial0/0.235)
Aug 12 10:41:22: dest 2001:420:1:3::2
On the other hand, using prefix /52 to each customer it works fine:
R 2001:420:1:2000::/52 [120/2]
via FE80::2, Serial0/0
R 2001:420:1:3000::/52 [120/2]
via FE80::3, Serial0/0
This archive was generated by hypermail 2.1.4 : Sun Sep 04 2005 - 17:01:19 GMT-3