This tutorial presents the IPv6 protocol, which is the successor to the IPv4 protocol used by the internet today. This protocol will allow, besides other things, the resolution of the limited number of available IP addresses problem.
IPv6 protocol description
The object of this tutorial is, before all else, to centralise the knowledge and to provide as much information as possible about IPv6. It is not to recopy the RFC word for word.
We should start the tutorial by stating the problem that has been thought about since 1992 by the Engineering Task Force, in relation to the saturation of IP addresses. In 1994 we saw the first recommendations. In 1996, the first IPv6 prototypes were updated. In 2003, we have started to migrate segments of IPv4 to IPv6.
The current IPv4 standard started to show some serious limitations. First of all, the number of available IP addresses was missing the necessary robustness and finesse needed for security and real time connections. This is where IPv6 comes in (or IP ng for “Next Generation”). This is necessary for the survival of the internet and has 2 main objectives:
Fix the problem of address spaces on the Internet, but also for other things like your car, dog, wife – anything!
We should anticipate that in the future multimedia applications will communicate differently with things becoming more mobile.
But this new standard must be able to be progressively added to routers, with a transparent existence with the current version of IP (IPv4).
With the new version of IP, the addresses will pass from 4 octets to 16 octets (making them 128 bits). You should also see that the addresses starting with 010 (1/8th of the total space) will be reserved for machines that are connected, saving the remaining 125 bits. The size of the reserved size can be variable with the notion of classes with a umber of fixed bits, has been fixed.
IPv6 also works with autonomous network notions and has planned addresses that start with 1111 1110 1 which will be freely used local addresses. The light bits will be used to identify the mask and address of the station that wants to “leave” the network.
IPv4 guarantees no delay on the path packets take and therefore must use tricks to be able to run applications like Windows Media Player and other “buffering” programs. It is for this reason that IPv6 has fields that “identify streams” and “prioritises” them. The general principle is the following: when I’m writing this, it isn’t exactly like this, but anyway:
When a server wishes to send a data stream (audio, video, etc) to a destination, it will attribute a unique identifier to all of its packets.
This signals to the routers which packets are coming, and treats them all equally, the characteristics of this treatment (for example, to reserve the determined transmission capacity). So that the packets are transmitted by the network management protocol or by the packets themselves (the extension is in the headers).
The “priority” field adds another setting that can be used by the sender, and treats the packets that have been sent by him. This priority is not complete, but relative to a resolved sender: 0 to 7 for normal traffic and 8 to 15 for real time items. In this case, the receiver is responsible for letting the sender know if there are any packet losses.
The designers have left these 2 fields so that they can correctly manage the priority hierarchy between the different data streams and the network (router) in respecting the 2 primary principles of the Internet; no bandwidth limitations for a particular delivery and autonomy for each network node.
The optional information is not included in the header (different from IPv4) but in the specific extensions that follow. The size of these extensions are multiplied by 8 and are not limited (eg imposed routing, encapsulation, fragmentation information) anything that you want as long as it is correct!
For security, the authentication guarantees the integrity of the transmissions. The encapsulations use the DES algorithms (Data Encryption Standard), but all this will evolve depending on legislation and user rights.
This is all very nice, but the problem is passing from version 4 to v6 (v5 or ST Datagram mode is almost not used: http://www.olympus-zone.net/00-Home/04-Protocoles/05-IPv5/'Theme=Blue&langue=fr) or how to have two protocols happily existing together on the same network, OUCH!
Yes, so these new machines possess a 16 octet address, which signifies to the routers and machines that they should retransmit IPv4 packets and vice versa.
More phases to this:
Total implementation of v4 and v6.
@v4 contained in @v6 and encapsulation of v6 packet in v4.
The translation is then done in the headers by the node to node routers.
The only constraint that is known is the update of the DNS servers (Domain Name Server: a machine that hosts lists that allows them to convert the alphabetic identifiers into IP addresses). We have now finished! I imagine that there are a few computer professionals out there that have now learnt something new.
To finish, people that wish to have more details about the RFC, I recommend that you look at these addresses:
http://www.urec.cnrs.fr/ipv6/RFC.html (This is a page from the CNRS site, where all the RFC papers that concern IPv6 are available, and there are quite a few of them).
|Previous review||Next review|
|DNS Server on Windows 2000 Server||Remotely take control with Real VNC|