Documentation improvement DHCPv6 prefix delegation

  • I refer to this the official documentation DHCPv6.
    In Prefix Delegation there's an example what would be a valid range for Prefix Delegation. That example helped me a lot. Thank you for that ✌
    I suggest you change the to part from FC07:1010:1010:FF00:: to FC07:1010:1010:FFF0::
    This way the whole /52 will be used so you get 256 times /60. Your example wastes 15 times /60 i.e. FF10 to FFF0.

  • Banned

    This post is deleted!

  • I don't know how you handle documentation improvement suggestions here at netgate. Shall I post in bugtracker instead? If you just prefer a better explanation I would be happy to give you one.

  • LAYER 8 Global Moderator

    @pmisch said in Documentation improvement DHCPv6 prefix delegation:

    Your example wastes 15 times /60 i.e. FF10 to FFF0.

    Huh.. They are showing you could use a "range" of your space in this case 1 of the /52 of your /48 and use to to delegate /60s to clients... So you could have what 256 clients that all get /60 delegated to them.

    Could it be worded a bit better.. Maybe they could call out that your taking the
    Which breaks down to 16 /56

    And using that for delegations of /60s

    I think they just showing that you could take a range of the the last 16 of the /56's your /48 break up into and using them to delegate /60

    When the docs were only on the wiki, pretty much anyone that asked and was given permission could edit.. But we can call in a few guys to see about documentation edits/enhancements..

    Any thoughts on this @chrismacmahon, what is the proper way to have documentation reviewed for clarity? And possible revision? @jimp or @Derelict ?

  • LAYER 8 Netgate

    I have opened a redmine with this suggestion:

    For example, if FC07:1010:1010::/48 is routed to a firewall, the Prefix Delegation Range of FC07:1010:1010:F000:: to FC07:1010:1010:FF00:: could be used with a Prefix Delegation Size of 60.

    It looks like the range here should be changed to FC07:1010:1010:FF00:: to FC07:1010:1010:FFF0:: (16 /60s) or FC07:1010:1010:F000:: to FC07:1010:1010:FFF0:: (256 /60s) or, if the range is left unchanged, the example prefix should be changed to /56 (16 /56s)

  • @johnpoz
    Building 16 times /56 at first sounds reasonable to me as this assumption is still valid with what the documentation suggests.
    So we assume that FC07:1010:1010:F000::/52 is being reserved for 256 times /60 prefixes which is
    1st: fc07:1010:1010:f000/60
    256th: fc07:1010:1010:fff0/60

    So the logical conclusion is to enter fc07:1010:1010:fff0 as this is the last network than can be handed out by Prefix Delegation.

    @Derelict said in Documentation improvement DHCPv6 prefix delegation:

    or FC07:1010:1010:F000:: to FC07:1010:1010:FFF0:: (256 /60s)

    I prefer this one because it seems like a common thing to me to go from 1 x /48 to 16 times /52 to 256 times /60 when you slice your /48.

  • LAYER 8 Netgate

    If all you are doing is PDs, maybe. The example is not going to match every use case.

  • @Derelict
    As this example already used 241out of 256 of that /52 for PD why would you want use the lacking 15 for something else?

    I know this suggestion might seem a bit too pettifogging but I'm just trying to make sane suggestions.

    I've been trying to exploit IPv6 since 2012. I might have had my issues with how to slice bigger prefixes like /48 or /32 into several chunks 🎃 . I work at a provider where I replaced my boss's ridiculous chunks with sane ones ☺ .

  • LAYER 8 Netgate

    Maybe you only need to delegate 16 /60s. Whatever example is chosen will not match everyone's needs.

Log in to reply