Thank you Mikael for this review, > > Hi, I am reading the document and typing as I go along: > > I do not know what a "disjoint" delegated prefix is. When googling, > "disjoint" seems to be a computer science term. Does it warrant explanation > in the document?
When I google for disjoint, I find http://en.wikipedia.org/wiki/Disjoint_sets. But fair enough. I will add a small explanation of what disjoint prefixes are. > > "Flooding mechanism", the description of that seems very specific? Shouldn't > it be about synchronizing information between the nodes instead of talking > about "reliable broadcast »? In a purely theoretical view, I think reliable broadcast and reliable synchronization are equivalent. But I guess your point is valid. ‘reliable broadcast’ is not appropriate. I will think about something else. Proposals are welcome too. > > I have now spent 30 minutes trying to grasp how the state machine works. I > don't know exactly why, but it's really hard, and it's hard to understand why > some of the wording was chosen the way they were. "Best Prefix" is confusing > when I at the same time read about "Current Assignment". Perhaps some text > explaining the rationale for the choice of words and why they are needed, ie > not only describing what needs to be done, but also why. "case study" or > "state study" is perhaps what I'm looking for. I think I understand what's > going on, but having some concrete examples would help. Would it be clearer if I add that the Current Assignment is the prefix assigned by *the node executing the algorithm* ? > > Another thing: If I know that I will receive 3 /56:es from my ISP, and I > always want $LINK to use the $56PREFIX02::/64 for all delegated prefixes, I > believe this is possible within the current algorithm, but it's not > specifically stated. It's stated that manually chosen prefixes can be done, > but not examples of how these might be chosen. I know this example from > cisco: > http://www.cisco.com/c/en/us/support/docs/ip/ip-version-6-ipv6/113141-DHCPv6-00.html I see your point, but I wonder if it is appropriate. The algorithm is capable of *many* different behaviors. Section 5 specifies that prefixes may be chosen based on any custom selection process specified by configuration. I don’t think it is part of the algorithm to list all the additional features that an implementation may propose. > > Here you can set "ipv6 address prefix-from-provider ::1:0:0:0:1/64" to > achieve what I just described. In case of multiple delegated /56:es, some > users might prefer that the same number subnet is chosen from all delegated > prefixes. That feature is even available in the current homenet implementation using the ‘link_id’ option (http://www.homewrt.org/doku.php?id=run-conf). Cheers, - Pierre _______________________________________________ homenet mailing list [email protected] https://www.ietf.org/mailman/listinfo/homenet
