Talk:IPv6 subnetting
- resolved added relevant links to and from different articles, also included external link to RFC. - Klaver (talk) 11 August 2010 at 14:36 (UTC)
Not an article
[edit]This is not an article, contains almost no useful information, and has a completely meaningless lead sentence ("the optimized network size"?). In addition the intent of this page completely misses the points of subnetting in IPv6. It should be deleted. Kbrose (talk) 14:18, 14 August 2010 (UTC)
- I concur. CIDR has meaning for any prefix length, not just those evenly divisible by four. Why stop at /8; smaller prefixes are also used. But who wants to know what astronomical number of addresses are available with a /37 prefix? —— Dandor iD (talk) 14:25, 15 August 2010 (UTC)
- Maybe deletion is a big step and all this article needed was a bit more descriptive text like IPv4_subnetting_reference? - Klaver (talk) 12:17, 20 August 2010 (UTC)
- No, the presentation of those tables is rather meaningless and does not represent the intent and purpose of the IPv6 addressing infrastructure. In IPv6, an address is treated similarly to a UUID of an interface with additional internal structure. Presenting the number of addresses for every conceivable 'network mask' is not only a trivial math operation not worthy of display, but it completely distorts the intent of having a large address. Not the number of addresses is important, it is guaranteed that there are enough, but the efficiency gained in a more refined hierarchical routing system. In analogy, not many people are interested in how many unique identifiers can be formulated within the various UUID generation schemes, they know there are enough. In IPv4 this was indeed an important concern, you had to justify how many hosts you're going to connect in a certain time period to get an allocation.
- Maybe deletion is a big step and all this article needed was a bit more descriptive text like IPv4_subnetting_reference? - Klaver (talk) 12:17, 20 August 2010 (UTC)
- Secondly, every subnet in IPv6 is a 64-bit space, subdividing it further makes little sense, is against intent and recommendations, and only for point-to-point links, and even then, rarely. Subnetting in the traditional sense of routing does occur on larger boundaries, but it's not routinely described as subnetting, you can't form a subnet with a, say, 96-bit interface id. On a practical level, one subnet still has 64 bits. It's described as route aggregation and people use the term 'subnet id', to divide their upstream allocation into 64-bit subnets. Kbrose (talk) 14:37, 20 August 2010 (UTC)