This content has been marked as final. Show 4 replies
Design patterns really is a broad subject and only the people involved in coming up with architecture and designs for a particular project should decide which patterns should be used.
Whether to have separate DTOs for inner calls depends on:
1.) Are the fields always going to the same?
2.) Are the 'inner' calls also standalone SOA designed so that they can also be called directly without the need to include properties relevant to other calls?
3.) (related to 2 above) How dependent are the XSDs for the DTOs if they are made separate
Thanks for your valuable time
I have 3 Webservices A(Provider),B,C. Service 'A' is a consumer of service 'B' and 'C'.
A - is exposed to Consumers, A act as a proxy layer for service B and abstract 'C' from users ,Service 'A' only do the service orchestration .Service 'A' using data from Service 'B' and 'C'
B – is of same specification as that of Service A (Service B contains All operation defined on A +Internal operation for service A).Internal Webservice which act as active datasource.Request and response structure is same as that of Service 'A'.
C – is different external service ,requests for Invoking Service 'C' need to be generated from Consumer Request and operations exposed in Service 'B'. External Webservice act as active Data source. Enrich the service 'B' output using service 'C' ,This is done by Service 'A'
I have done this using following components.
WebserviceEndpoint-> WebserviceFacade(POJO) -> Helper Classes ->Aggregator/Workflow Impl which uses WSClients(JAX-WS Stub)
It will be great If some of you can suggest a good logical separation for Service 'A' ? Is above logical separation is correct?
Basically same responses as already given. Just to add that I would not use the same pojos for the data that I would send out to customers. In particular I would not make JPA entities part of object graphs sent out to customers. Store your data in the most efficient manner possible and build relevant views for the customers.