I digged some docs, I think I'll do LDAP+web for user self-servicing+SAML with Shibboleth+maybe OpenID. That will complicate the automated customer setup, but it'll cover many use cases.
Not entirely sure of your question. I'd think exposing your ldap service to the open internet is asking for very very bad things to happen.
I'd expect a service like this would use Amazon VPC, a secure VPN to access it, or something like that. What kind of data do you envision being stored in this directory, user credentials, or other things?
At $last_job, I was on a mission to put everything in LDAP. There is a custom OpenLDAP schema that the gnome.org sysadmin team (which I am an alumni member of) which allowed users to put in their ssh pub key via a webui. Hosts then run a cronjob every XX time period that put those ssh keys down (in a root owned directory/file so users can't change them), and was pretty slick. I also put DNS zone info and sudoers information into LDAP, as I already had a badass distributed datastore, ldap :)
That being said, can you come up with a real use case where your service makes sense? Active Directory is hard to compete against, it is super cheap and a pretty solid kerberized ldap for SMBs.
Classic use case: org want facilitate+centralize users management on owncloud+apache webdav+other oss app on external server (internet). I only saw AD used in intranets, if orgs would expose it on external servers (with/without VPN), then I'd better find an another idea.
Most orgs put an OpenLDAP proxy in front of their AD server. AD has multiple known crash vulnerabilities in its protocol parser (fuzzing attacks can easily break it) and is too slow to handle the load generated from open internet access.
Howard, obviously no one is more of an OpenLDAP expert than you... You have users who expose OpenLDAP to the internet directly? I've got no qualms against OpenLDAP, it is amazing software, but that still seems insane.
If you need reliable, high-throughput REST-style APIs in Java, I'm your guy. After 7+ years of experience in server-side Java (JEE 5&6, Spring, Wicket...), I'm specializing in lightweight stacks. I love Dropwizard, JEE6, PostgreSQL, Lucene... Please email me at username at gmail.com for more infos.
I now work from home, in the south of France, enjoying a fullfilling life with my SO and my 3 weeks old daughter : I won't relocate .
My favorite coding atmosphere is like 70% N2, 20% O2 and 1% Ar.
Big desk, two large monitors, Aeron chair (I miss it :/), Logitech illuminated keyboard, G500 mouse, 21°C temp, 11 PM, nobody, no noise, just the keyboard clicks rythming the night.
It also depends of who's the target audience. I did two Show HN for a medical literature search engine and got a total of two comments. On a medical-related reddit, I got much much more comments, thanks, suggestions, etc etc.
Nice: SQL mastering, good ORM trade-offs and inner-working understanding, good general database understanding and usage (let PostgreSQL be your friend)
But you'd also better master few pieces of software like Hadoop, Solr, Talend, Cassandra... to differentiate/specialize (try and pick what you love doing, contribute to open-source with these technologies)
Thank you for separating the technologies. I had a feeling Spring and Hibernate would be important to learn but I also wanted to make sure I didn't mess up by learning useless technologies.
The fact that you pointed out SQL was interesting to me as it seems you were one of the only ones that did so. I firmly believe that my work, both in school and at my internship, allowed me to excel in SQL. So, personally, I believe I do have that and the database understanding under control. But I do thank you for pointing those out. I was worried I wouldn't be using those skills again.
I have no plans to expand to all scientific literature, unless I find a good source of structured data to work on. If there is a Pubmed equivalent for other scientific fields, please drop me a link :)
