Please, find my comments in the text. - Michael
________________________________ From: Rob Eamon <[email protected]> To: [email protected] Sent: Monday, March 23, 2009 4:53:07 AM Subject: [service-orientated-architecture] Re: Joe on SOA without service-enabled apps My thoughts in-line below. --- In service-orientated- architecture@ yahoogroups. com, "Udi Dahan" <thesoftwaresimplis t...@...> wrote: > > > Is it the same service interface if it is offered over a "real > time" pipe or as a nightly batch? IMO, these would be different interfaces. MP: Agree > > How about - is it the same service? My thinking is that they should be. There was a discussion a few months back that explored this where the thought is that batch should be its own set of services. MP: if batch processing is not differ from a processing of single item in repeatable (or parallel) manner, I do not see why separate service is needed. However, I would be accurate here - sometimes people mean some additional processing when they talk about batch processing, i.e. some values derived from 'many' items and that are unavailable for deriving from 'single' item... > > Does this make SLAs part of the service, or its interface? Yes. :-) I think it depends on the specific service level being considered. Some, I think, would apply to all interfaces of the service (e.g. security aspects?) and thus would be part of the service definition. Others would be specific to an interface. For example, the process time of a batch-oriented interface/operation would differ from a near-real time interface/operation . MP: I still hesitate tying SLA with any INTERFACE because SLA is provided via the interface, not by the interface. I think that SLA always belong to the service. Another thing when you specify an SLA and (mandatory) mention which interfaces support it and which not. I think, this is more accurate and clear approach. I do understand a sort of slang you use when talking about 'SLA is a part of interface', but I am afraid that many would take it literally and end-up with confusion - dropping value of service itself, i.e. the entity that provides the SLA in reality. > Might the SLA be part of the difference between a contract and an > interface? What are the pros/cons of making a distinction? MP: see my comments above. > > Does any of the above change when the consumer is a subscriber to > events the service publishes? IMO, no. MP: as I explained in my first answer, IMO, yes. However, the only difference caused my 'yes' is that there is a possibility that in the pub/sub the publisher, i.e. service, may be substituted by an intermediary and subscriber, i.e. service consumer, has to arrange Service Agreement (including SAL) with this intermediary instead of the service/publisher. -Rob
