Slashdot Mirror
Red Hat Opens Netscape Directory
suezz writes " Eweek is running a story that Redhat is releasing Netscape Directory (LDAP) under the GPL - this is huge at least from my point of view. I know of at least two huge companies that have standardized on Netscape Directory for their web applications."
Red hat paid $20.5 million for this LDAP. Will they get that in return? Is it possible with this type of software?
I think this is a good thing, I'm just honestly curious, having messed around with OpenLDAP, and never really doing much with ND.
What's the major differences, feature-wise not philosophy-wise (no Free vs free vs Open vs open rants).
I don't need no instructions to know how to rock!!!!
How does this improve my user experience?
How can using ND make my life, as a user/administrator/purveyor of exotic animals, easier?
I think that is a useful question to ask any time a "new" feature is presented.
I know this story is going to prompt people wanting to know how the Netscape directory server compares with OpenLDAP. I've never used the Netscape one but what I would really love to know is how does it stack up against Novell eDirectory? eDirectory isn't open source but the licenses are damn cheap, the first 250,000 licenses are free. Any LDAP experts care to share their opinions?
I've used OpenLDAP and Netscape Directory Server. NDS is a *very very very* cool product. It's easy to use, scales like there's no tomorrow (it was the backend for a lot of the older Netscape Netcenter sign on functions) and it's nice & documented. (I still have books for it)
Red Hat releasing it under the GPL is a good thing, any way that you look at it. Cool product, "big name company" supporting it, and oodles of applications that can already use many of its functions.
Now, if someone would slurp up Netscape Calendaring Server and release *that* under the GPL..
If the Netscape SuiteSpot Server suite still existed and was under the GPL, there's your Exchange-killer right there.
Can RH possibly integrate the http://hula-project.org/ into this roll out? I would really like to have THE non-M$ directory/email/calendaring system running for my school district: single sign-on and email accounts for teachers, staff, students, parents... with Mac OS X Server directory delegation, Kerberos, etc.
A killer kombination for Open Source.
My first ever experience with LDAP was with openldap, and it took 10 minutes to configure, and then maybe an hour to work out how I wanted my schema, and write an ldif of it to import. Unless it used to be significantly different than it is now, I can't see any way anyone could think its hard to configure.
Isn't Sun's Directory Server based off this as well? I thought they'd acquired all the old Netscape stuff back in the Netscape/iPlanet days.
The problem with LDAP is that adding the 'L' (lightweight) to the 'DAP' (directory access protocol) removed many features including, most noticably, proper distribution of data over multiple servers and proper chaining of requests.
Proper distribution and request chaining protools would allow Linux systems and MS systems to share a perceived common user data store. At the moment, hybrid enterprises are forced to support multiple islands of trust in the organization. It also sets the operational limits of the system to an enterprise/employee rather than a global/customer scale solution.
Still, it's a good thing that Red Hat is implementing a directory based identity management solution. It's a step in the right direction.
In the development and staging environments it was great. As other posters mentioned you could get from zero to something usable in less than 30 minutes. Everything was as you would expect.
However... in the -production- environment, with 10's of millions of ldap objects connected to SprintPCS's provisioning systems which were making 1,000+ ldap writes --a minute-- the SunOne system absolutely blew chunks.
LDAP architects will ask what the hell we were doing with the entire database in one ldap instance rather than partition the dataset, and they'd be right, but we were acting under Sun's direction since at the time we had one of (if not) the largest LDAPs in the world.
LDAP architects would also wonder why on earth you would ask an ldap server to live under such a write intensive churn, and they'd be right again.
That being said...
-- Multimaster replication would never ever work. Most of the time the entire SprintPCS userbase was hanging off one master and less than 4 replication slaves. For several months the entire messaging system was wedged into a single point of failure nightmare. (to be fair, this wasn't all slapd's fault and had 1/2 of the root cause in Sprint Datacenter practices which produced predictable results)
-- Other posters asked for SunOne Calendar server to be opensourced. My first response is to suggest you have your head examined since that thing would die for absolutely no reason on a regular basis. We actually automated the process of detecting its death and restoring from last night's backup. If you were a SprintPCS customer and your calendar ever seemed screwy now you know why. Of course further reflection suggested opensourcing it is probably the only thing that could help at this point because...
-- We used to get hotfix builds from Sun which were missing entire sections of the binaries. Whoever was managing the code would forget to use the same compilation flags for hotfixes as original code so we would receive webmail frontend builds which couldn't talk to imap backends, or calendar backends which wouldn't accept connections from calendar front ends.
-- SOL if you wanted to run more than 4G of memory in slapd.
Dont consider this post a rant, just let any CIO's/etc. reading this know that this opensource release will probably work great for you if you dont load it heavily (unlike exchange 5x, which would grenade just sitting there)
On the other hand, if you want to push the performance envelope, pretty much expect it to take alot of time and cause a bunch of headaches -in production-. Get help from people who have pushed the performance of the tools you are considering running.
Weird mood tonight.
Not since 1999-2000. The overall shape is still similar but the internal details have all been reimplemented by the OpenLDAP Project. Today OpenLDAP is miles ahead of Netscape in terms of performance, scalability, and stability.
a p/history/index.html
See for yourself:
http://www.stanford.edu/services/directory/openld
OpenLDAP 2.0 is slow, snail's pace, frozen molasses slow. That's the release that RedHat has bundled for years, up to RH9 and even beyond. It's only in the past few months that anything from them (Fedora Core) has shipped anything newer.
OpenLDAP 2.1 is over Two Hundred Times faster than OpenLDAP 2.0 and already significantly faster than Netscape 5. OpenLDAP 2.2 is 30-50% faster than OpenLDAP 2.1 and leaves Netscape in the dust. OpenLDAP 2.3 is faster yet.
-- *My* journal is more interesting than *yours*...
At the bottom of the page is the download link. It does appear to go to a "free" evaluation/beta copy.
I didn't download it though, so I don't know what the exact terms of use are.
The fact that there is a "Buy Now" would suggest that the eval copy is for testing but not production. Just a guess though.
I ran the Corporate Directory for a major US automaker for a number of years.
We used Netscape's Directory Server. There were hundreds of apps pointing at it, and the main Internet proxy server used it as the authentication service.
Over a million objects, hundreds of thousands of searches per day. It might crash once or twice per year, and never corrupted anything.
The management GUI sucked, but it was an outstanding product in all other respects.
DG
Want to learn about race cars? Read my Book