Nov 20 2008
IPv6 An Introductory Overview
Internet Protocol version 6 is the next version of the Internet Protocol that will handle all of the logical address assignments at good old Layer 3 of the OSI model. The version that is still in use today is version 4. In fact Version 4 in its current use has been active since 1981. IP version 5 was a protocol called Internet Stream Protocol, which didn’t make the grade.
You may be asking yourself… why is there a newer version of IP? If something has worked well for so long then why bother changing it?
Well the biggest reason why is that we are running out of logical address to use on the Internet. There is even an article here that states that China may run out by 2011.
IPv4 has the capability to supply 2^32 number of Addresses. IPv6 has the capability to supply 2^128 number of address. That’s 2 to the 128th power. Just to break it down, that means that IPv4 can supply approximately 4,294,967,296 address to private organizations. The IPv6 can supply approx 340 undecillion addresses.
My favorite analogy is this… If we were to assign one billion addresses every second since the earth was formed, at the time of this writing, we would not have used even one trillionth of the address space available.
So what is important for us to know as Administrators? Here’s a kicker… IPv6 is Enabled, and Preferred on Windows Vista and Server 2008.
It is my prediction that due to the cost of equipment that fully supports this amazing protocol and the complexity of engineering that IPv4 will be around for a while yet as well. Note that in the previous statement I mentioned that Engineering may be complex, however administration may be even easier then IPv4.
IPv6 has a number of performance enhancements and features available. An example of this is that IPv6 no longer includes the concept of a Broadcast. Similar to its IPv4 predecessor, it still uses unicasts and multicasts, but adds a whole new method of communication called an Anycast.
- A unicast can be described as a one-to-one communication.
- A multicast is a one-to-whoever-chooses-to-tune-in communication.
- An anycast is a one-to-a-member-of-a-group style communication. Usually this is the easiest to reach member of a group. An example of a group might be collection of Routers. In the above example the initiator would request information to a group of routers, but only the most easy to reach router out of a group of pre-defined routers would respond.
Now that we understand the need to transition to this modern form of communication, let’s take a quick look at some of the protocols that IT professionals will be dealing with.
Some of the protocols are enhancements to good old fashion network science like these three.
- DNS. Names are resolved using an aaaa record and yes folks we refer to this as a ‘quad a’ record. This is going to be really important since an IPv6 address is going to be long.
- DHCPv6. There will still be plenty of use for the science of DHCP. As cool as Link-Local addresses are, there are some things that are just missing through other dynamic addressing schemes. The network address of the DNS server is one good example of what DHCPv6 is capable of that other protocols don’t provide. In IPv4 we remembered the initial address as a DORA process that’s broadcast on ports 67 and 68. The v6 process is a little more involved in that it uses Multicasts on port 546 and 547, and I guess we can call the process SARCRRR. That would stand for Solicit, Advertise, Request, Confirm, Renew, Rebind and Reply. Heck that’s only 3 more letters that what we learned before.
- IPv6 over PPP. This allows for dial-up and broadband access to the Internet through an ISP on good old OSI Layer 2.
Then we have some protocols that are new to IPv6 such as…
- LLMNR (Link-Local Multicast Name Resolution). This protocol is the “new and Improved” Netbios. It enables computer on the same subnet to communicate with each other using a FQDN. All of this magic is accomplished without a DNS server. It does this with Multicasts instead of Netbios’s Broadcasts, which means that LLMNR is more efficient.
One of the things that IETF thought of when developing IPv6 is having intolerability with the existing IPv4 infrastructure. Here’s a quick overview on some of the protocols that will accomplish this goal.
- ISATAP (Intra-site Automatic Tunnel Addressing Protocol). This protocol was developed in order that IPv6 and IPv4 can exist within the same internal network structure. It is also used the magic of this protocol is that you can deploy newer IPv6 protocols with no additional change to the IPv4 infrastructure. This protocol is responsible for IPv6 over IPv4 networks. For network engineers, understanding the specifics of this will allow for a gradual migration from IPv4 to IPv6.
- 6to4. This protocol allows v6 to communicate to either a v4 or v6 over the existing ipv4 Internet. The difference between 6to4 and ISATAP is that 6to4 protocol is that ISATAP is used locally whereas 6to4 allows for connections across the v4 portion of the Internet.
- Teredo. This protocol is very similar to 6to4. The major difference from Teredo and 6to4 is that Teredo provides for end-to-end connections across the Internet that are behind a NAT interface whereas 6to4 resolves endpoints are public addresses.
This is naturally a brief overview of some of the things we can expect to see with regards to IPv6. The next time we look at IPv6, we’ll cover Ipv6 Addressing including the IPv6 reserved addresses.
�
Reference List:
- Understanding IPv6. By Joseph Davies.
- CCNA Study Guide sixth edition by Todd Lammle.
- TCP/IP Guide… by Charles M. Kozierok.
All of which are based on the many RFC as defined by the IETF (Internet Engineering Task Force) which can be researched on the following website.
