documentation/presentations/EclipseSummit_2008/proposal2.html

Do you like to model and generate your applications with the <em>Eclipse Modeling Framework</em> (EMF)?
That's fine because EMF is really great at generating model based runtimes, user interfaces and XML serialization!
<p>

But what if:
<p>

<ul>
<li>you require more enterprise-like model persistence than XML serialization?
<li>your application has to be available for multiple users at the same time?
<li>your model resources are too huge to fit into XML files?
<li>you need object grained locking and transactions that span multiple resources?
<li>your users appreciate to always look at the latest state of the central model?
<li>they could simply switch back their view of the model to historical states?
</ul>

The CDO model repository and distributed shared model for EMF 
delivers solutions to the afore mentioned challenges and more. 
CDO is a component of the Modeling project for many years now 
and the latest generation has a completely new design and a 
whole bunch of cool new features. 
<p>

This long talk starts with guided examples for the main usage areas:
<p>

<ul>
<li>Setting up a CDO model repository server
<li>Generating a model for native usage with CDO
<li>Using a distributed shared model at client-side
</ul>

The examples are followed by an overview of some of the optimizations 
possible to tweak performance and footprint:
<p>

<ul>
<li>Caching at client and at server side
<li>Partial collection loading
<li>Intelligent prefetching
</ul>


A preview of the upcoming CDO 2.0 features rounds up this talk.
<p>

Although not strictly necessary for the understanding of this talk we recommend
to attend the preceding talk with a high-level introduction, too:
<p>

<em><a href="https://www.eclipsecon.org/submissions/ese2008/view_talk.php?id=131">CDO Model Repository - Part 1: A Functional Introduction</a></em>