Technological Advantages of IPv6
IPv6 is a technical evolution of IPv4. Many things that are familiar from working with IPv4 will remain or be similar. We do not have to master a completely new technology. The driving reason to develop a new protocol, was the expected address depletion. But the working group determined in the early nineties, that they had enough time to not only extend the address space in the protocol, but build in additional features which would make it a more efficient protocol to meet the growing requirements of future networks and services. So let us examine what the new key features are.
Extended address space
An IPv4 address has 32 bits or 4 Bytes.
Example: 192.168.0.1 (decimal notation)
An IPv6 address has 128 bits or 16 Bytes.
Example: FE80:0000:0000:0000:0202:B3FF:FE1E:8329 (hexadecimal notation)
There are rules, to abbreviate the address, by omitting leading zeros and replacing a series of zeros by two colons.
So the address above can be written as: FE80::202:B3FF:FE1E:8329.
Many people argue, that we have enough IPv4 address space and that with technologies like NAT, where many users with private IP addresses can hide behind a single globally routable IP address, there is no issue about address space. There are several misconceptions behind this viewpoint. NAT imposes severe limitations when it comes to scalability and end-to-end
- security. A device behind a
NAT can initiate a connection to a global host, but someone from outside cannot initiate a
connection to someone behind a NAT. We also have to consider that more than 70% of the global IPv4
address space belongs to corporations and organizations within the United States. There are
providers in the US that have almost as much address space as the whole of Asia today. If countries
like China, Japan, India and many others want to connect their people to the Internet, imagine what
numbers of addresses they need. They are using IPv6 today, they have no choice.
When it comes to services like real-time banking applications with high security requirements or services like Voice over IP, NATs can be killers. The almost unlimited global address space of IPv6 will resolve these issues.
In addition to this, we do not only need more address space to connect more people to the Internet. In the future all sorts of always-on devices like mobile phones, sensor devices, tv-sets, digital radios, refrigerators, air conditioning devices, cars, and many more to imagine, will need a permanent IP address. This growth of address demand can only be met with IPv6.
The almost unlimited IPv6 address space will let us reestablish the end-to-end paradigm. This was a fundamental design rule of the Internet, which was broken with NAT.
Perhaps the most intriguing new feature of IPv6 is its autoconfiguration mechanism. A booting device in the IPv6 world can come up and ask for its network prefix. It can get one or more network prefixes from a router on its network. Using this prefix information, it can autoconfigure for one or more valid, global IP addresses by using either his MAC identifier or a private random number to build a unique IP address. In the IPv4 world we have to assign a unique IP address to every device either by manual configuration or by DHCP. You can still use DHCP with IPv6 if you wish, but you have other options.
Autoconfiguration will facilitate the lives of network managers and save substantial cost in maintaining IP networks. But not only this. If we imagine the number of devices we may have in our homes in the future, that need an IP address, this feature becomes indispensable. Imagine reconfiguring your DHCP server at home when you bought a new tv-set! Autoconfiguration also allows for easy connection of mobile devices, like for instance your mobile phone or handheld, when moving to foreign networks.
Simplification of the header format
The IPv6 header is much simpler than the IPv4 header and has a fixed length of 40 Bytes. This allows for faster processing. All unnecessary fields have been removed and additional options can be added in the form of extension headers, which are only inserted into the packet, if they are needed. So for instance all the fields in the IPv4 header for fragmentation are removed from the basic IPv6 header. If a packet needs to be fragmented, a fragmentation extension header is inserted. There is a basic set of six extension headers defined in the current specification, but the model makes it easy to create specifications for additional extension headers, when the need arises, without changing anything in the basic IP header.
The currently defined headers are used for routing information, for RSVP (Resource Reservation Protocol), for Mobile IPv6, for QoS services (Quality of Service, Flow Labelling) and for Security Options like authentication and privacy options.
It is expected, that IPv4 and IPv6 will co-exist in our networks for many years to come. Therefore the developers put a lot of attention to developing co-existence and transition mechanisms to make the transition as smooth as possible.
The mechanisms available today can go into one of three categories, and they are:
- Dual-stack techniques
allow IPv4 and IPv6 to coexist in the same devices and networks. This will probably be the most used and easiest technique. It allows to set up dual-stacked hosts, that can access IPv4 applications using their IPv4 stack and access IPv6 applications using their IPv6 stack. It is improbable, that all applications can be ported to the new protocol at the same time. This scenario allows to use the new protocol where it makes sense, without neglecting older applications that haven't been ported.
- Tunneling techniques
allow the transport of IPv6 traffic over existing IPv4 infrastructure. These techniques allow an organization to migrate parts of the network to IPv6, even while the backbone is still running IPv4. Or to migrate to IPv6 and connect to the outside world, while the provider (ISP) is still IPv4-only. This means, for building islands of IPv6 networks, you do not need to rip out your backbone routers. You can wait until the backbone routers fulfill their life cycle and then replace them by IPv6 capable devices. And you do not have to wait until your ISP offers commercial IPv6 services (although many of them are 'sort of' ready, only they do not announce it yet).
- Translation techniques
allow IPv6-only nodes to communicate with IPv4-only nodes. With this technique, it is even possible that through an ALG (Application Level Gateway) an IPv4 host can talk to an IPv6 host and vice versa. This is your last resort, if everything else fails. It does not really let you take advantage of the advanced features of IPv6 and the ALG creates a bottleneck and performance hit. But it may help in specific scenarios.
- Dual-stack techniques
- Common myths about IPv6
There are several misconceptions circulating in discussions about IPv6. Let us explore the most common of them:
- All devices need to be upgraded to IPv6.
With all the transition techniques available, IPv6 can be introduced independently at any point and on any number of hosts in the network. Single hosts can start using IPv6 and talk to each other directly, if on the same link, or with ISATAP in a network without IPv6 aware routers. They can talk to hosts outside their network by using any of the transition mechanisms like 6to4, a Tunnel broker or Teredo, if sitting behind a NAT.
- The core of the network is too difficult and expensive to upgrade to IPv6.
Again the transition mechanisms make it possible to migrate to IPv6 at the edge of the network with no depency on the backbone. IPv6 packets traveling over the backbone to reach another IPv6 segment within the company, can be tunneled in IPv4 packets. You do not need to upgrade the backbone first.
- Our ISP doesn't offer IPv6 services, so why should we upgrade?
Same answer, you do not have to wait for your ISP to upgrade to IPv6. As long as he does not offer IPv6 services, tunnel your IPv6 packets over his IPv4 network.
- It would be too hard to upgrade all applications to IPv6.
Basically the statement is true, it is not realistic to expect that all applications can be switched to use IPv6 at the same time. But there are enough transition scenarios and techniques, which make it possible to live in a dual-stacked world and have access to some applications over IPv4 and to other applications over IPv6. We expect this to become the most used transition scenario in the real world.
A side note: the effort needed to port applications to IPv6 is often overestimated. It depends on the quality of the application code.
- All devices need to be upgraded to IPv6.
Continue to part three: IPv6 in the World
For greater insight into IPv6 we recommend Silvia's latest publication, 'IPv6 Essentials' published by O'Reilly in late 2002.
Silvia Hagen, owner and CEO of Sunny Connection AG is
the author of a number of books. She regularly speaks at international conferences like Brainshare
US and Europe, NUI Events, IPv6 Summits and other technical conferences. Sunny Connection AG (www.sunny.ch) is a leading IT consulting and
education company based in Zurich Switzerland. Our main expertise is in directory services
integration and in network analysis and protocols like TCP/IP and IPv6. We have more than ten years
of experience in consulting middle and large sized companies, mainly in the area of industry,
banking and insurance.
This was first published in June 2004