Thanks to Steve for trying to tow this barge up stream in all this shallow thought.
I think it is truely amazing that there seems to be so many people still having a hard time thinking in the abstract. SOA is about architecture, with an A, and that could also mean Service Oriented Abstractions! The simple point I was trying to make with a simple statement, is that services' interactions need to be modeled as interactions, not as imaginary open ended coversations. I actually believe that the simplest architecture model involves a system where everything is done by writing stuff on pieces of paper and passing them around, filing them (database simulation) and otherwise doing what used to be done for services 100 years ago before there were computer systems. If you have all of these pieces of paper, that are passed around, then you have what feels like a messaging system. My problem with that, is that you can then neglect to really think about responses, transactional and leasing behaviors and other invisible items that help the system be stable and make progress in doing the needed work. Now that we have computers that "never fail", we can have "reminders" (aka notifications, lease/transaction timeouts etc) that something is not done yet (i.e. when I lost that note of what I was supposed to do for you, businesses could fail). We can have "audits run" (aka validation of content, state and timing) and anything that we imagine to validate the state and progress of the "machine." The communication mechanism of all the information between the pieces is frankly, immaterial. But everything a computer system does has a "beginning" (the call out to do the work) and an end (the return to the point of the call). That is the model that helps make it very visible when a responsibility is passed off to another party. Are there attributes of some types of communcations/messaging that are helpful to a system? Absolutely YES! But, that's something that's technology based. I'm trying hard to say think abstractly about the interactions, and just say "here we need to XXXX" and XXXX is the business step not the "send a message" or "put the record in table y". If you sat down and had a CRC card exercise to develop an OO system architecture, you'd do all the "is a", "has a", "needs a" etc. steps. These steps are the most fundamental in developing a data processing abstraction. It helps you to understand better where responsibilities really should be. If you say Service instead of Object as you go through the process, I think there is power in the CRC process to organize thoughts around what the service needs to do. Then, when you have all the "is a" pieces, you also have the "has a" attributes that help dictate what service interactions need to happen. The "needs a" pieces let you see how "has a" must to be used so that you can start thinking toward the technology interactions that are really needed (I at the transaction, leased, lifecycle level, not the CORBA, TCP, FTP, HTTP level). Conveying data items between systems is a trivial task if you are in charge of all the endpoints. When you are not, then you have to make choices that can be dynamic and overlapping. By just making your architecture look like an RPC interaction, you make sure that you deal with the partial failure issues, because there is no imaginary handoff. It's in your face as you look at it. Gregg Wonderly Michael Poulin wrote: > I agree with "the conceptual architecture using a single consistent > style" but I disagree that this style is RPC. Relationships and > corresponding interactions between architectural entities are the part > of the architecture, it is not a technology; I can send a message via > MOM or via FedEx, this is immaterial at the architecture level. > > - Michael > > ------------------------------------------------------------------------ > *From:* Steve Jones <[email protected]> > *To:* [email protected] > *Sent:* Fri, January 15, 2010 10:08:07 AM > *Subject:* Re: [service-orientated-architecture] Steve on RPC > > > > > > 2010/1/14 Eric Newcomer <e_newco...@yahoo. com > <mailto:[email protected]>> > > > > I don't think this is correct. We are modeling and designing > services with various message exchange patterns, both synchronous > and asynchronous and we consider this to be consistent with SOA > based designs. > > > At the IMPLEMENTATION and TECHNOLOGY levels I completely 100% agree that > this is what we do. What I'm saying (and Gregg pointed out) is that you > are often better off doing the conceptual architecture using a single > consistent style (RPC) and then making the implementation (design) > decision afterwards. The alternative is having a high-level > architecture which exposes technical implementation styles. > > Steve > > > > Eric > > > On Thu Jan 14th, 2010 5:50 PM EST Steve Jones wrote: > > >Fight ;-o > > > >2010/1/15 Michael Poulin <m3pou...@yahoo. com > <mailto:m3poulin%40yahoo.com>> > > > >> > >> > >> Gervas, > >> > >> let's start with that "Get Delivery Status" is not really a > capability as > >> well as "Get Forecast", they are operations of the interface as > the most > >> > > > >How do you work that out? > > > >To obtain a delivery status for an order is a core capability that > an order > >management service must have. To get that status it might have to > query > >external systems as well as doing internal work. > > > >A Forecast is even more of a business capability, creating a forecast > >involves potentially millions of calculations and feeds from > massive number > >of other systems. If its a real-time forecast then the amount of > work can > >be stratospheric. > > > >A Capability is a thing that delivers the real world effect. > Getting the > >Delivery status or getting a forecast results in a real world > effect and > >undertakes a constrained piece of business functionality. > > > > > > > >> . That is, we have clearly distinguish between service > capabilities and > >> service interface operations. This is the first 'attack' on the > RPC as a > >> high level of 'a consistent view' - what is consistent from RPC > perspective > >> may be inconsistent from service perspective. > >> > > > >I can't really see it as an attack when you haven't explained, > beyond a > >statement how neither retrieving delivery status or a forecast > result in a > >real-world effect. > > > > > > > >> > >> "When we call "Make Forecast" on the Finance Service it needs to > ask the > >> Sales Service for its Forecast and therefore does a "Get > Forecast" call on > >> the Sales Service. We need the Forecast to be updated daily" - > from the > >> service consumer perspective (the only correct viewpoint for > analysing > >> services), the Finance Service is responsible for the Forecast > >> irrespective to whether it gets it from the Sales Service or > not. From > >> architectural viewpoint, > >> the Finance Service has to interact with the Sales Service, and > this is > >> not necessary to an RPC call (may be MOM-based request); this > depends on > >> the Finance Service's Service Description. If "we need the > Forecast to be > >> updated daily", this is another reason for NOT using RPC between > the Finance > >> Service and the Sales Service. > >> > > > >I am NOT saying that the implementation should be RPC but that the > >ARCHITECTURE is conceptually RPC. Having a consistent conceptual > model and > >allowing multiple implementation models is the under-pinning > strength of a > >business SOA approach. > > > > > >> > >> "Now when we start working through this at a systems level ..." > >> utilisation of legacy systems in Mainframes and FTP/ETL is fine > but this is > >> not dictated by the RPC style (as you confirmed: "...but instead > of making > >> an RPC call to get that information we have decided...". That > is, RPC is > >> good but not really universal to become the Development Solution > #1. This is > >> the second 'attack' on the RPC. > >> > > > >Eh? I'm not saying that RPC at the conceptual level = RPC at the > >implementation level. > > > > > > > >> > >> > >> Provided description of "RPC into Events" calls for messaging > solution > >> bypassing anything related to RPC. Given event-driven scenario > is typical > >> 'push' scenario, which operates in contra-RPC model. This is the > third > >> 'attack' on the RPC. > >> > > > >Nope, its nothing of the sort. Its proving the point that RPC as a > >conceptual framework can be used and then mapped to multiple > implementation > >approaches. The Business folks don't need to worry about the exact > >implementation approach (see Christmas SOA) but the validity of the > >conceptual model. > > > > > >> > >> REST=RPC via special transport protocol, no doubts. But REST can > play only > >> the interface role for the service. REST is technology not a > separate > >> concept, and this concept is RPC. > >> > > > >I'm not even going there... > > > > > >> > >> Finally, I totally disagree with the "Conceptual v Delivery" > and, I hope, > >> I have demonstrated on your examples limited usability of RPC > for service > >> design. This is why I repeat - service must be designed as > service and > >> its implementation must respect and preserve service constraints and > >> boundaries. The horse pools the cart; the cart, does not matter > if it is > >> a phaeton or a royal carriage, does not push the horse. > >> > > > >Nope I don't think you have. What you've demonstrated is that at the > >implementation level you can use lots of approaches (which is what > I've > >said). Having a conceptual architecture which surfaces EDA, GDA, > REST, RPC, > >MOM and all the other implementation approaches to the business is > just > >wrong and not (IMO) SOA. > > > >SOA means that you have services and they have capabilities which > can be > >exercised > >RPC means that you have interfaces and they have operations which > can be > >exercised > > > >This is why the two go well together at a *conceptual level* and > you then > >choose the right *implementation approach* as you move into design. > > > >Steve > > > > > > > > > >> > >> - Michael > >> > >> ------------ --------- --------- > >> *From:* Gervas Douglas <gervas.douglas@ gmail.com > <mailto:gervas.douglas%40gmail.com>> > >> *To:* service-orientated- architecture@ yahoogroups. com > <mailto:service-orientated-architecture%40yahoogroups.com> > > >> *Sent:* Thu, January 14, 2010 11:01:08 AM > >> *Subject:* [service-orientated -architecture] Steve on RPC > >> > >> > >> > >> Gregg Wonderly made a good comment on the Yahoo SOA list the > other day > >> > >> *I think one of the still, largely unrecognized issues is that > developers > >> really should be designing services as RPC interfaces, always. Then, > >> different service interface schemes, such as SOAP, HTTP (Rest > et.al.), Jini, > >> etc., can be more easily become a "deployment" technology > introduction > >> instead of a "foundation" technology implementation that greatly > limits how > >> and under what circumstances a service can be used. Programming > >> Language/platform IDEs make it too easy to "just use" a single > technology, > >> and then everything melds into a pile of 'technology' instead of a > >> 'service'.* > >> > >> > >> > >> The point here is that *conceptually* RPC is very easy for > everyone to > >> understand and at the highest levels it provides a consistent > view. Now > >> before people shriek that *"But RPC sucks"* I'll go through how > it will > >> work. > >> > >> First off lets take a simple three service system where from an > "RPC" > >> perspective we have the following: > >> > >> The Sales *Service* which has *capabilities* for "Buy Product" > and "Get > >> Forecast" > >> > >> The Finance *Service* which has *capabilities* for "Report Sale" > and "Make > >> Forecast" > >> > >> The Logistics *Service* which has *capabilities* for "Ship > Product" and > >> "Get Delivery Status" > >> > >> There is also a customer who can "Receive Invoice" > >> > >> Now we get into the conceptual design stage and we want to start > talking > >> through how these various services work and we use an "RPC > language" to > >> start working out how things happen. > >> > >> *RPC into Push* > >> *When we call "Make Forecast" on the Finance Service it needs to > ask the > >> Sales Service for its Forecast and therefore does a "Get > Forecast" call on > >> the Sales Service. We need the Forecast to be updated daily.* > >> > >> Now when we start working through this at a systems level we see > that the > >> mainframe solution of the Finance team is really old and creaky > but it > >> handles batch processing really well. Therefore given our > requirement for a > >> daily forecast what we do is take a nightly batch out of the CRM > solution > >> and Push it into the Mainframe. Conceptually we are still doing > exactly what > >> the RPC language says in that the data that the mainframe is > processing has > >> been obtained from the Sales area, but instead of making an RPC > call to get > >> that information we have decided in implementation to do it via > Batch, FTP > >> and ETL. > >> > >> *RPC into Events* > >> The next piece that is looked at is the sales to invoice process > Here the > >> challenge is that historically there has been a real delay in > getting > >> invoices out to customers and it needs to be tightened up much more. > >> Previously a batch has been sent at the end of each day to the > logistics and > >> finance departments and they've run their own processing. This > has led to > >> problems with customers being invoiced for products that aren't > shipped and > >> a 48 hour delay in getting invoices out. > >> > >> The solution is to run an event based system where Sales sends > out an event > >> on a new Sale, this is received by both Finance and the > Logistics department > >> . The Logistics department then Ships the Product (Ship Product) > after which > >> it sends a "Product Shipped" event which results in the Finance > department > >> sending the invoice. > >> > >> So while we have the conceptual view in RPC speak we have an > implementation > >> that is in Event Speak. > >> > >> *RPC into REST* > >> The final piece is buying the products and getting the delivery > status > >> against an order. The decision was made to do this via REST on a > shiny new > >> website. Products are resources (of course), you add them to a > shopping > >> basket (by POSTing the URI of the product into the basket) this > shopping > >> basket then gets paid for and becomes an Order. The Order has a > URI and you > >> just simply GET to have the status. > >> > >> So conceptually its RPC but we've implemented it using REST. > >> > >> *Conceptual v Delivery* > >> > >> The point here is that we can extend this approach of thinking > about things > >> in RPC terms through an architecture and people can talk to each > other in > >> this RPC language without having to worry about the specific > implementation > >> approach. By thinking of simply "Services" and "Capabilities" > and mentally > >> placing them as "Remote" calls from one service to another we > can construct > >> a consistent architectural model. > >> > >> Once we've agreed on this model, that this is what we want to > deliver, we > >> are then able to *design* the services using the most appropriate * > >> technology* approach. I'd contend that there really aren't any other > >> conceptual models that work consistently. A Process Model > assumes steps, a > >> Data Model assumes some sort of entity relationship a REST model > assumes its > >> all resources and an Event model assumes its all events. > Translating between > >> these different conceptual models is much harder than jumping from a > >> conceptual RPC model that just assumes Services and Capabilities > with the > >> Services "containing" the capabilities. > >> > >> So the basic point is that architecture, and particularly business > >> architecture, should always be RPC in flavour. Its conceptually > easier to > >> understand and its the easiest method to transcribe into different > >> implementation approaches.>> > >> * > > > You can read Steven's blog at: http://service- architecture. > blogspot. > >> com/2010/ 01/think- in-rpc-develop- in-anything. > html<http://service- architecture. blogspot. com/2010/ 01/think- > in-rpc-develop- in-anything. html > > <http://service-architecture.blogspot.com/2010/01/think-in-rpc-develop-in-anything.html>> > >> > >> Gervas* > >> > >> > >> > >> > > > >
