When assigning IPv6 addresses to a point to point link, you usually don’t go to the trouble of creating a /127 prefix. In IPv4 this was done because the preservation of addresses is important. For IPv6 this is not the case. So for any point to point connection, using a /64 prefix is more than acceptable.
The standard used by most ISPs is to provide you with a /48 prefix, and then you can further subdivide that into /64 prefixes. That gives you 65536 subnets for your whole network. Typically, you should use a whole /64 prefix for every one of your subnets, even if you have many point to point networks. It’s just much simpler.
Now having said all of that, there there are good reasons to use /127 address spaces for point-to-point links. These reasons are so good, that an RFC has been written up just for this. RFC6164 is titled “Using 127-Bit IPv6 Prefixes on Inter-Router Links”.
Now how would you configure a /127 network? Well, here’s an example where two routers are configured to communicate in a point to point connection over their FastEthernet 0/0 interfaces:
Hi all,
In IPv6 /127 was really confusing me when i realized that it is the same as using a /31 in IPv4. Is there any reason in IPv4 to prefer a /30 over a /31 for p2p links in a real world scenario? The only reason i could think of is that /30 has a broadcast address but why do we need a broadcast in a p2p? /30 seems useless now that i think about it. It is bugging me because i always see /30 for p2p in my work.
Can you please explain?
Sorry about the delay, it seems this post slipped through the cracks… my apologies once again!
Strictly speaking, according to the original IPv4 definition in the early 1980s, a subnet must have a network address and a broadcast address in addition to the host addresses available. Even with the advent of VLSM in 1995, you could not have a /31 subnet mask because of this restriction. Indeed there was no need for such a subnet mask since there were plenty of IPv4 addresses to go around in those days.
However, in 2000, with RFC 3021 the use of /31 as a subnet mask was defined and accepted. So virtually all modern networking equipment should support /31 prefixes.
Now is there an advantage of one over the other? No. They both work in exactly the same way, as long as the equipment supports it. (There may still be some vendors that won’t allow you to use /31, but as far as I know, all modern Cisco equipment supports this). Such prefixes are also correctly dealt with in the device’s routing table as well. That is, you will see /31 routes appear in the routing table, assuming the particular prefixes are shared using dynamic routing protocols.
There is no advantage to using /30 over /31. However, there is the advantage of using /31 over /30, in that you conserve the use of IPv4 addresses.
I hope this has been helpful, and apologies once again for the delayed response!
Note that in the first case, the fourth hextet is 0002, while in the second, it is 0020, and these are indeed different.
The rule here is that all leading zeros are removed. This has been done in both cases. In the first, there are three leading zeros, while in the second, there are two. More info on how to shorten IPv6 addresses can be found in this lesson:
I was wondering why I didn’t see any IPV6 prefixes that started with 2000 or even 3000. Usually you see 2001 prefixes for everything. if you visit the IANA IPv6 Global Unicast Address Assignments page you will see that IANA hasn’t handed out any prefixes that start with 2000. 2001:xxxx/23 prefixes are handed out quite a lot which explains why they show up everywhere. https://www.iana.org/assignments/ipv6-unicast-address-assignments/ipv6-unicast-address-assignments.xhtml
Many global unicast IPv6 addresses do indeed begin with 2001, and this is true because of the fact that the IANA has begun assigning IPv6 addresses to Regional Internet Registries starting with this address range.
The choice of the “2001” prefix was based on a few factors, including the fact that it is easy to remember and easy to distinguish from other prefixes. It was also chosen because it is a unicast address range that is not used for any other purpose in IPv6, and it is compatible with the requirements of the Multicast Listener Discovery (MLD) protocol.
As more and more IPv6 addresses will be assigned, more and more IPv6 prefixes will be used, so we’ll begin seeing additional numbers starting off the IPv6 addresses.
I got a /40 from ARIN as our company is in North America. This would be more than enough for our company that is world wide. I am wondering if I could use the ARIN prefix in countries in Europe and Asia or do I need to register with RIPE or other local RIRs. Thanks!
Once you get your registered IPv6 blocks from any regional internet registry (RIR), you are free to use them anywhere in the world, as long as you advertise them appropriately using BGP.
However, there is a catch. If you do so, you will likely experience suboptimal routing, increased latency, and possibly higher costs (depending on how your ISP traffic costing is done). This is because BGP typically routes the specific block assigned to each region to that area of the world. So users will initially be routed toward North America, but will then be rerouted to wherever else in the world you are advertising those addresses from.
To successfully achieve something like this, it is crucial to coordinate with your local ISP and the RIR in which those addresses will be residing to ensure proper routing and address management. It can get more complicated than expected, so you must weigh the pros and cons of such an approach carefully.
It is generally recommended to use IPv6 addresses assigned by the RIR responsible for the specific geographic region where the addresses will be used to avoid such situations. You will probably find that it is more appropriate to purchase a block of IPv6 addresses from each RIR in which you have offices, it is likely that you will have fewer potential problems this way.