This is a single archived entry from Stefan Tilkov’s blog. For more up-to-date content, check out my author page at INNOQ, which has more information about me and also contains a list of published talks, podcasts, and articles. Or you can check out the full archive.

SOA Days 2006: The Central Role of Registries

Stefan Tilkov,

For some reason, I forgot to mention this conference — it was organized by Deutsche Post AG’s SOPSOLUTIONS and took place from 21-22 September. I gave a presentation on “The Central Role of Registries”. Again, this will probably not make much sense without the accompanying audio, especially since I experimented a little with a different presentation style; still: the slides are here.

On October 31, 2006 2:26 PM, Karl Krukow said:

Stefan, I really enjoyed reading the slides but, as you point out, some things don’t make sense without the audio.

I am curious about your points about registries vs. repositories. Can you elaborate on this here?

Cheers, - Karl

On October 31, 2006 7:14 PM, Stefan Tilkov said:

Thanks Karl. What exactly is it that you’d like to see elaborated? The difference between registries and repositories?

On November 2, 2006 3:42 PM, Karl Krukow said:

Yes, first of all the differences. You said that registries store metadata and references whereas repositories store actual data besides metadata. What would be examples of “actual data”?

You also say that:

“An integrated SOA registry/ repository can be used as the single point of information about SOA metadata.”

Why is an integrated reg/rep needed at all?

I guess what I would like elaborated is more or less getting the information equivalent of having been present at the talk (at least during the part where you discuss registries/repositories) ;-)

Cheers, /Karl

On November 2, 2006 3:50 PM, Stefan Tilkov said:

I see. First of all, I don’t necessarily believe an integrated reg/rep, or even only one of them, is mandatory in any way. It is a valid strategy to choose this approach, and if you do, you have to have a solution for both maintaining information about what data you have, as well as the actual data itself.

As an example, there may be a service called CustomerManagement with a WSDL describing its interface at http://example.com/services/CustomerManagement/wsdl. A registry would store this information. A repository would also store the actual WSDL (the XML) itself. A WSDL-aware repository might interpret the WSDL when it’s registered, extracting information about the interface(s) and operations, schema elements used etc.

If you want to find out, for example, what services would be affected when the Customer schema is changed, your reg/rep needs to know about both Customer and CustomerManagement (the artifacts themselves), not only their URIs.

What you call data and what you call metadata is largely a matter of taste. I generally try to avoid the terms if possible.

On November 3, 2006 12:23 PM, Karl Krukow said:

Ok, I see what you mean.

Thanks, Stefan.