Well, it's a start...

LITS Open Directory Infrastructure

Okay...not much here yet, but just a quick overview to get the ball rolling.

So anyway, what is LITS Open Directory Infrastructure (LODI)

LODI is a an idea I had for a standardized, FOSS directory server. It began when I started working on a new mail server and wanted groupware functionality. I started with the typical tools that you'd find in a mail server with virtual users -- an MTA (Postfix), a database to store virtual user accounts (OpenLDAP), an MDA (maildrop), IMAP server (Dovecot), Cyrus SASL, MailScanner (SpamAssassin, Razor, ClamAV, DNSBLs, etc.).

As I dug in a little deeper, I really began to see why LDAP was all the news a couple of years back. And so, I set out on a little journey to see what I could come up with to try and mimic the functionality of a typical Microsoft Active Directory deployment. I started digging in to find a way to make a Windows box play nicely with a Unix directory server and stumbled across lots of great documentation on pGina, the Samba LDAP work, NFS, OpenAFS, Linux attr and acl support, and all the other intricate little pieces that play very well together, but take a ton of work to setup properly. I thought it'd be cool to be able to automate the entire process, and one of my primary goals here is to make Windows interoperability a major focus so that adoption by Windows users (and administrators) is as easy, and as transparent as possible. And that is where I am today. No, I don't want to recreate Microsoft, Novell, or Apple's directories...though some of their ideas will be reused. This is a new project that I want documented from the ground up. I want easy administration and use, to do that, we have to provide easy to understand documentation behind every design decision. It'll probably start out as a book and possibly a script, much like BLFS in that it follows on from a base LFS, but linear in nature. I'm not even sure on the name, I just came up with it a few weeks ago. A quick Google search suggested that “ODI” was taken, so I needed a new name and prefixed it with LITS (LucasIT Services). Hell, I obtained a great deal of my Linux experience from LFS, so LFS probably deserves the L if this gets some momentum. :-)

I've got some scripts laying around, I've toyed with a few ideas, Kerberized daemons, put some other object types in my test directories to describe printers, workstations, exported storage, pDNS integration, etc....but there is still a ton of work to be done. I will definitely need some help...and that's if this thing ever even takes off. I've reactivated LucasIT Services, my former (well current now) business, to house the project in it's inception. I'll probably need to open a SF page or ask for some resources in LFS land if this begins to draw attention. So far, I have began a write-up on the mail server that I did, which no, I haven't submitted as a LFS hint yet. I have also requested a PEN from IANA to account for new schema objects and attributes that I'll (we'll?) need to reach the goals of my (our?) directory design. Maybe Gerrard should do this under the LFS hat too. I suppose they (IANA) are probably waiting to see something from LITS because until today (20080407), there was a mail server and nothing else here besides “Directory list of /”. So, to IANA, yes I really do exist and intend to use the PEN that I've requested. :-)

BTW, if you are reading this and wondering what the heck “LFS” is, I am very proud of it and can't help but insert this plug: Please do take a peek at http://www.linuxfromscratch.org/ . I am a book editor for the follow-up project “Beyond Linux From Scratch” where I write and maintain instructions for installing a couple hundred Open Source projects from only the upstream source code. My personal responsibilities include the Sun JDK and OpenOffice. I have also worked in just about every chapter in the book at one time or another. I had provided the initial inclusion of the modular X.org installation, though it has changed considerably by the careful hands of Dan and Alexander. Gerrard, Matt, Randy, Bruce, how the heck did you guys manage to lead a group of hackers to such a well refined distribution?!?!? :-) I am quite proud to be a part of the LFS family of professionals who have extended the 'from scratch' approach well beyond anyone's imagination, and very much look forward to leveraging each of the interested developers expertise in my (our?) project. Thanks guys!

Anyway, this short little write-up is getting very long winded so I had better wrap it up quick. I really hope that I can generate some interest with the ground up approach I am looking at, but I'm going to have to get y'all more concrete information on the project goals before I can even begin to ask for more help. I know I'm busy in dev land right now on the BLFS book, but I promise, more documentation will appear in the weeks following BLFS-6.3's release. I'll probably start looking at the {,*}LFS groups for talent (and I have already invited a few of yas), but if you should stumble across this page by accident, and are genuinely interested in shaping the future by extending FOSS acceptance into the enterprise, please feel free to drop me an email at “dj_AT_lucasit_DOT_com”.

Thanks for reading.

-- DJ Lucas