Damjan,good to hear there is more interest in the DHCP arena. It could really use some more contributors.
Damjan Jovanovic schrieb:
you are right, the current code base is very rudimentary. I have been working on driving the DHCP implementation towards being usable. Unfortunately, The inclusion of this contribution is currently pending due to open questions regarding the need to incubate the contribution as it has been developed outside the project for some time. For its current status see this message:Hi I'm glad to see that Apache DS does more than just LDAP, but I see that the DHCP protocol implementation is very rudimentary - the code in SVN trunk indicates is just a skeleton and I've been told nobody's working on it any more.
http://mail-archives.apache.org/mod_mbox/directory-dev/200712.mbox/[EMAIL PROTECTED] towards the bottom (PXE proxy).
Yes, absolutely - I stumbled upon exactly the same problem. There is very little documentation about how this is meant to work. And no documentation towards the container relationships.Anyway I'm interested in improving the DHCP protocol and I've even started writing some code, only to discover that there is no "standard" DHCP schema and the one from the IETF drafts expired many years ago. That schema also seems to make little sense, everything is linked using attributes instead of structured in container relationships, there is no documentation on how the objectclasses and their attributes are supposed to be used, and it seems at best very inefficient - allocating an IP address seems to require a linear search through the list of dhcpLeases to find an unused one (for a /8 network, that's up to 16777216 comparisons).
As stated in the linked message, there is, even in the mentioned contribution, no full DHCP implementation. However, protocol- and message format-wise there is already quite a lot that is working and usable. I see the majority of required effort in the configuration and state handling area - i.e. the stuff that ends up in the directory and is/should be covered by the DHCP schema. Since the DHCP schema is unusable in its current state, I'd recommend implementing it the way you/we understand it and fixing the specification as we go along.Any suggestions on what I should do, try to implement the draft schema as I understand it, or invent a new one?
Joerg Henne
smime.p7s
Description: S/MIME Cryptographic Signature
