# openEHR

**URL:** <https://discourse.openehr.org/t/openehr/14461>\
**Category:** Technical (archive)\
**Created:** [31 May 2007 22:54 UTC](https://discourse.openehr.org/t/openehr/14461 "2007-05-31T22:54:33Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![Thilo\_Schuler1](https://discourse.openehr.org/letter_avatar_proxy/v4/letter/t/a4c791/32.png) [@Thilo\_Schuler1](https://discourse.openehr.org/u/Thilo_Schuler1)\
**Post date:** [31 May 2007 22:54 UTC](https://discourse.openehr.org/t/openehr/14461/1 "2007-05-31T22:54:33Z")

</div>

Hi Bernard

I have just made post the should cover part of your question. As  
mentioned in this post the kernel component is central to the  
architecture of an openEHR system as it brings the two models  
together. For persistence many solutions would be possible. Like XML  
(that what Ocean Informatics uses in their EhrBank product) or an  
object-relational-mapping solution. But it should be noted that just  
instances of the reference model need to be persisted including  
references to the archetypes used to create the instances. The  
archetypes can be retrieved again separately when the persisted data  
is loaded through the kernel.

For Java kernel code have a look at  
[http://svn.openehr.org/ref\_impl\_java/TRUNK/project\_page.htm](http://svn.openehr.org/ref_impl_java/TRUNK/project_page.htm)

Thilo

---

<div class="post-metadata">

**Author:** ![Thilo\_Schuler1](https://discourse.openehr.org/letter_avatar_proxy/v4/letter/t/a4c791/32.png) [@Thilo\_Schuler1](https://discourse.openehr.org/u/Thilo_Schuler1)\
**Post date:** [31 May 2007 22:57 UTC](https://discourse.openehr.org/t/openehr/14461/2 "2007-05-31T22:57:34Z")

</div>

Here one more link to a nice overview written up by Thomas about  
persistence possibilities and techniques. Most of you will already  
know this:  
[http://openehr.org/FAQs/t\_persistence\_notes.htm](http://openehr.org/FAQs/t_persistence_notes.htm)