Experimental Physics and
Industrial Control System

"Jeff Hill" <[email protected]> · Mon, 17 Nov 2003 17:40:11 -0700

> Just two comments.
> 
> 1) Why we need to restrict ourselfe to LDAP? I agree that LDAP is a good
> choice at the moment,
> but we don't know 5 years later. And EPICS core life will be much longer
> than that.
> At least, we have to have a layer to separate LDAP dependent part from
> other.
> 
> 2) This functionality should be optional. User of EPICS should be able
> to start a simple thing in a simple way.
> 

How this is integrated is certainly a big issue. For instance, I'm not fond
of requiring that an EPICS site must build and install LDAP prior to
installing or using EPICS base. However, a competing thought is that it
would be nice to come up with a fairly consistent solution between sites
providing wildcard based name resolution. I do agree that placing all of our
eggs in the LDAP basket might be risky. Needs more thought.

The CA gateway has built-in caching directory service capabilities which are
already being used as a mechanism for scaling EPICS systems. However, this
does not provide wild carding capabilities. Since they are open source, and
already implemented, then the wildcard features in systems such as LDAP are
certainly attractive.

Jeff

> -----Original Message-----
> From: Noboru Yamamoto [mailto:[email protected]]
> Sent: Monday, November 17, 2003 4:42 PM
> To: Jeff Hill
> Cc: 'Pete R Jemian'; EPICS-tech-talk
> Subject: Re: Among the things for which I wish...
> 
> Jeff Hill wrote:
> 
> > <snip>
> 
> >We are considering making changes so that an IOC uploads its record
> list to
> >an LDAP server when its databases are loaded. Among other benefits this
> >would allow for wildcard queries using open source LDAP client
> libraries.
> >One of my papers at ICALEPCS 03 discusses some of the issues
> surrounding the
> >use of EPICS in WAN.
> >
> >Jeff
> >
> >
> >
> 
> Just two comments.
> 
> 1) Why we need to restrict ourselfe to LDAP? I agree that LDAP is a good
> choice at the moment,
> but we don't know 5 years later. And EPICS core life will be much longer
> than that.
> At least, we have to have a layer to separate LDAP dependent part from
> other.
> 
> 2) This functionality should be optional. User of EPICS should be able
> to start a simple thing in a simple way.
> 
> Noboru

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System