Quote:
|
Originally Posted by Xaccers
Basically some networking gurus took some really bad acid and came up with IPv6
Then a sensible person came along and said "hang on, if we use NATing and host headers, companies won't need so many IPv4 addresses so we won't need to go to IPv6 quite so soon" and there was much rejoicing  |
Well, that's simply not true, IPv6 has been in the working for sometime now and is not seen as merely a fix on the v4 addressing structure. Furthermore, there is no hurry to move to v6, since no 'killer' application for it has yet been found - although the IETF are also still working on the underlying features that will accompany v6 architecture.
IPv6 comes with a host of features that should improve the way in which the underlying protocols of the Internet function. Issues directly addressed by IPv6 are things such as configuration, Network-Level Security, Real-time data delivery, Quality of Service - as well as address space utilisation and aggregation, which will help reduce the size of the routing tables. The introduction of the 3 [now 4] Regional Internet Registries originally helped slow down the distribution of v4 address space by ensuring that addresses were only assigned to organisations having fully justified their requirements for an actual physical need for the address space. There's still more than plenty v4 address space to keep us going for some time now, and further measures could [in theory] be put in place in order to reclaim un-used address space, if the need arose. NAT is not a solution to the exhaustion of v4 address space, since it is not always possible for networks to adopt this practice, due to various implications that may adversely affect the performance of the network.
IPv6 addresses are made up of 128 bits, rather than 32 bits as used in IPv4. The use of 128 bits allows for a total of 340,282,366,920,938,463,463,374,607,431,768,211,45 6 unique addresses, compared to the 4,294,967,295 addresses available under v4.
v6 addresses are written in colon-hexidecimal format, taking the form of:
xxxx:xxxx:xxxx:xxxx:xxxx:xxxx:xxxx:xxxx, for example:
2001:0700:7654:3210:FEDC:BA98:7654:0321
2001:0700::/32 is the prefix assigned to the network operators. They in turn have control of all address space under this prefix, which helps with aggregation and the reduction of the routing tables.
v6 can be deployed across v4 networks, provided that the routers on the network are v6 enabled. The most likely action will be that there will be a long period when v4 and v6 co-exist using dual stack routers on the network. It's quite likely that your network provider [although doubtful with NTL] will actually provide v4 and v6 addresses to end users. New operating systems are already v6 enabled, as the original mail in this posting shows - although it's important to note that the likes of XP will run v6 [if enabled on the local machine] in preference to v4 in situations where v6 transit is accessible - such as on wavelans. This could result in longer routes and lengthy connection times if the v6 access is not as effecient as v4, so it's advisable to only initialise v6 on machines that are likely to use it. There'll be a few changes in the types of resource records used for v6, but there's not a great deal of difference in terms of functionality.
In answer to the initial question, yes, DNS servers are still used in v6 and the underlying functions are reasonably similar in terms of name to number resolution - which means that NTL are again likely to go fsck it all up for us
