Guys,

I might be wrong here, but is this not a U2 problem. Should IBM not fix the
software so that all access to UV consistently uses the same security
function calls so that the method of authentication is consistent across
interfaces, and if you like transparent to U2.


Phil Walker 
+64 21 336294 
[EMAIL PROTECTED] 
Gnosys Consulting Limited 
11 Woodward Road, Mount Albert, Auckland 1003, New Zealand 
DISCLAIMER:  This electronic message together with any attachments is
confidential.  If you are not the intended recipient, do not copy, disclose
or use the contents in any way. Please also advise us by return e-mail that
you have received the message and then please destroy. Gnosys Consulting
limited is not responsible for any changes made to this message and / or any
attachments after sending by Gnosys Consulting limited. We use virus
scanning software but exclude all liability for viruses or anything similar
in this email or any attachment

-----Original Message-----
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of Craig Bennett
Sent: Tuesday, March 22, 2005 10:23 AM
To: [email protected]
Subject: Re: [U2] [AD]UniObjects and LDAP user authentication.

Gordon,

I was thinking about your problem further over the weekend. You can use 
our routines to authenticate from BASIC but I think that being unable to 
directly authenticate with UniObjects is going to make you jump through 
some hoops.

You will need to have at least one local unix user on the UniVerse box 
so that UniObjects can login at all. This will give all your uniobjects 
users the same effective Unix Username and permissions (not sure how 
important that is at your site).

Once logged in you can then call our LDAP routines to authenticate a 
user/password for access to the system. Our routines support simple and 
SASL MD5-HASH authentication, optionally over SSL (if you can make the 
U2 SSL sockets work at your site).

I think this is a better solution than using LDAP on the webserver, 
because it is UV that decides if a user should have access but it is 
still an inferior solution to direct UOJ authentication because you have 
to write and maintain code to decide if a user is authenticated separate 
to UOJ.

Problems with this setup are:

- Will solaris allow the single UOJ user to be authenticated locally?
(I would think this must be possible).

- Can solaris configure the UOJ user so that it can only be used for UOJ 
(telnet, ssh, ftp etc disabled, setting a null shell would go part way 
but it would be nice to limit their login further).

- You will need to store the UOJ username and password on the machine 
where the UOJ client runs (or distribute the username and password to 
all your users and have them authenticate twice). Both of these feel bad 
you decide which is less so.

- If you use unix level security to segregate access to data you will 
need to play with the AUTHORIZE statement to change the effective user 
id of process. I just don't know if this will work from UOJ.


Craig
-------
u2-users mailing list
[email protected]
To unsubscribe please visit http://listserver.u2ug.org/
-------
u2-users mailing list
[email protected]
To unsubscribe please visit http://listserver.u2ug.org/

Reply via email to