jackrabbit-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Darren Hartford <binarymon...@yahoo.com>
Subject ILM JCR model best practices - start!
Date Fri, 22 May 2009 13:46:24 GMT

ILM = information lifecycle management, or another way to describe it, regularly archiving
off old content, but ensuring that old content may still be available in some fashion (see
the old memory hierarcy at http://en.wikipedia.org/wiki/Memory_hierarchy).  On additional
item for discussion also includes archiving to optical media (CD/DVD) which has two distinct
properties to be aware of for ILM - read-only and fixed size per storage unit (500mb/4g/etc).

In the JCR world, places that could be used in/for ILM include repository, workspace, nt:folder,
and the individual date properties scattered across many node types, as well as the difference
between the metadata and the binary data (particularly if separated with configurations like
Datastore).

---------------

DRH: So, what are some best practices people have been using for JCR? My first crack at something
like ILM would actually be date-based, since you are always adding new data -- and probably
break up the JCR model so each unit of time (year/month/day) that sounds reasonable for that
solution is it's own workspace.  For a solution that is relatively busy site, that has less
than 500MB of disk storage per month, a monthly workspace may work well and burn each month
off to CD. Or, alternatively, a workspace per quarter (which would fit into a DVD based on
that solution).  HOWEVER - this contradicts some of the rules/best practices about one workspace
per solution, so what have other people come up with?






      

Mime
View raw message