With the coming wide scale implementation of IPv6, it has become vital that network engineers become familiar with IPv6 and how it operates. While there are many different parts of IPv6 that mimic the same behaviors of IPv4, there are some subtle changes that engineers must be aware of. This article takes a look at the IPv6 address types and how they compare with the existing IPv4 address types.
Like IPv4, IPv6 has three different address types that are commonly used; these include:
The IPv6 unicast address type operates in the same way as IPv4; a unicast address is used to send traffic from one source device to one destination device. This is the most used address type as most traffic (currently) is sourced from one place and destined for one other place; this may change as network telephone and video services become used more.
The IPv6 multicast address type also operates in the same was as with IPv4; a multicast address is used to send traffic to a group of devices. This is used when there is a single source that needs to send traffic to multiple destinations; this traffic can be treated differently by the intermediate devices compared with unicast. If this same traffic would be sent via unicast methods, these devices would need to process each traffic stream from source to destination separately causing additional traffic delay.
Another large difference between IPv4 and IPv6 is with the IPv4 broadcast address type. IPv4 utilized a broadcast address type that allowed devices the ability to send a single packet that would then be delivered to all devices on the local subnet. While this capability was very useful, it would also cause all devices on the local subnet to process the packet. This was not a problem when the device was the anticipated target. However, in most situations, the majority of the devices that received the packet were NOT the anticipated targets and thus caused a considerable amount of processing waste. With IPv6 the capabilities that used the IPv4 broadcast address type are replaced by a number of different multicast addresses, for example, FF01::1 is used to speak with all local nodes and FF01::2 is used to speak with all local routers.
The IPv6 anycast address type is where IPv4 and IPv6 differ. The anycast address type offers something that was not offered with IPv4 which is the ability to send a packet to the closest device. When using anycast, a group of devices are given the same anycast address; when a source looks to send a packet to that address, the routing protocol will calculate the closest match to the source device and send the traffic there.
While there are some differences between the address types for IPv4 and IPv6, for the most part these differences are easy to understand. One important change is the removal of a broadcast mechanism; this mechanism was commonly abused and caused a great deal of excess overhead on local devices. The idea is that without this ability, devices will be able to spend less time processing those packets that it does not care about and use this time for those that it does. Once IPv6 replaces IPv4 on the majority of devices, these design decisions will be truly proven one way or the other. The content of this article will hopefully help the reader understand the differences between the two IP versions and help in the continuing transition to IPv6.