Hello, thanks a lot for your help. First of all I had a misunderstanding of SDO. I thought there is somewhere stored a central graph and I can get a part out of it. However I realized it is not this way. I'll try to explain what my scenario is and what I want to do.
My scenario is that I build an application with the spring framework. The different components should have access to the same master data. Therefore it would have liked to use SDO to have a local copy of the master data in each component. It could be that all the components access the same master data very often. So in this case if there is a read on the same data very often I want it to be cached and not read from the database every time. I define my data as ecore model. I use spring in my application and it is easy to define DAOs - so I thought I could program DAOs and maybe put them easily into a DAS. Is there a guide howto write your own DAS? - cause this would be very cool cause I want also to use SDO for my services of the components as abstraction layer later. There also is the question when some data refers to another data in another graph (I have semantic links between the data objects)- is there any support from Tuscany here? And how can I put all the different DAS together and put them into a big graph - that I can visualize or browse for example - is there a way? Cause I have data that is often needed - so I can serve it as in memory Java objects that are delivered via SDO and not so often used data is delivered via a slower data source. And when I want to access a DAS I don't want to use the XML transformation or RMI when the DAS and the access is on the same server. Is this possible via OSGI or so that I directly get plain Java objects from the source without being transformed and so on? Regards Tim -----Ursprüngliche Nachricht----- Von: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] Im Auftrag von Luciano Resende Gesendet: Donnerstag, 13. Dezember 2007 13:00 An: tuscany-user@ws.apache.org Betreff: Re: SDO and Spring Tim, as Kelvin mentioned, DAS would be the component handling heterogeneous data sources, it would be the interface between the Data (SDO Graphs) and the Data repository. From your e-mail, it looks like you want to persist the SDO utilizing Hibernate or Spring DAO. I could see two ways here : Having a new DAS implementation that would utilize this technology to persist the data (but note that Apache have license issues with Hibernate) or refactory DAS to allow plugability of the persistent layer (today RDB DAS uses JDBC under the covers to persist the data to RDB). Could you also elaborate little more on your cache question ? On Dec 13, 2007 1:39 AM, kelvin goodson <[EMAIL PROTECTED]> wrote: > Hi Tim, > welcome to Tuscany. What I'm not completely clear on from your question > is what you mean by "publish" and "central graph". There's a related > discussion on this list going on (see [1]), and its clear that Jason too had > initial difficulty seeing the picture here until he found the DAS material. > I think perhaps we need to do some work on the SDO part of the Tuscany > website to make this better understood. > > Bert Robben has taken part in that discussion, and his scenario involves > SDO as the primary data API, with conversion to other data formats via an > abstract black box snapshot format being done when necessary, such as > Hibernate being used for some of the back end persistence activity. > > So the notion behind the SDO API design is that in order to access > heterogeneous data sources, Data Access Services will be created to build > SDO data graphs. The RDB DAS is one such example, and you can view the > XMLHelper as providing the services of another. If you are using an RDB as > your central store and want SDO access to the data then the natural solution > is to use the RDB DAS. > > Tuscany has no concept of providing a façade over other data handling > implementations, and I think the disparity in the features of the data > representations you mention would mean that providing an SDO façade would > not be the natural way of solving interoperability. I'd be keen to get a > better idea of your scenario in order to be able to understand in more > detail how SDO can help you. > > Regards, Kelvin. > > [1] http://www.mail-archive.com/tuscany-user@ws.apache.org/msg02162.html > > > On 13/12/2007, Tim Frey <[EMAIL PROTECTED]> wrote: > > > > Hello, > > > > > > > > I searched the whole day how i could build up a data graph and publish it > > via SDO. > > > > I want to access the data graph via Tuscany SDO. > > > > I found out about the relational solution DAS, but I would like to have > > the > > data graph as Java objects with hibernate persistency. Or Maybe even > > better > > have also spring DAOs as nodes in the graph. > > > > How can I build up a central graph and is this even possible to do what I > > want? > > > > Or is the relational solution also capable of caching? > > > > > > > > Sorry maybe this questions are answered somewhere but I don't find out > > where > > to start. > > > > I really hope that someone can help that I can build up a central data > > store > > accessed in a unified way in my application. > > > > Or maybe someone can propose me a better solution than SDO for that. > > > > > > > > Kind regards > > > > Tim Frey > > > > > -- Luciano Resende Apache Tuscany Committer http://people.apache.org/~lresende http://lresende.blogspot.com/ --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
