Building Service-Oriented Applications in C#.NET

Generator Code 39 in C#.NET Building Service-Oriented Applications

The first and most obvious use of a service is to create a Service-Oriented Application, or SOA This overloaded, overused term means that your application (like the heating system I mention in the previous section) uses services to assist a remote client with communications with a server Unfortunately, the term SOA is no longer useful in the marketplace So many so-called experts have chimed in with large, unmanageable ideas for the creation of systemwise services that it is impossible for anyone to separate the wheat from the chaff The concept is simple: The user interface calls a service at some point in the process to communicate with the server You do this for two reasons: Scalability: It s the main reason to use a service inside an application especially a Web application Up to a point, most Web applications have built-in scaling If you need more access points, you just add more servers and then use a device to sort the traffic to another machine, as shown in Figure 1-3 Web applications are different, though, because layers have different scalability needs Sometimes the database is loaded, and sometimes the Web server is You must be able to separate the layers of the application, as shown in Figure 1-4 that s where services start to become useful

Because the services use a common format usually XML over HTTP you can relatively easily install parts of the application on their own machines to isolate them physically Because the functionality of the deleted part is called via a service, you can scale the application horizontally Reusability: Every organization has a list of its participants clients, users, voters, cooks, or mailing list subscribers, for example Regardless of the type of participant, all the people on the list have first and last names and other identifying characteristics It seems like nearly every application written today has a table of People Savvy programmers use slick tricks to keep these tables in sync or to share information, for example, but only one way exists to share the People table by using a data silo, as shown in Figure 1-5 A data silo works this way: 1 A database on some server somewhere contains all participants demographic information 2 The database is surrounded by a service layer containing all allowed operations 3 The service layer is consumed by all other applications in the system, as shown in Figure 1-5 This set of steps shows you the concept of reuse It isn t about bits of code (no matter what you might hear) it s about data Services help to provide access to the data silo

A common use of services is to give other programmers public access to your information by using an open API that can be consumed anywhere You might recall the WSDL file, described in the earlier section Contract driven it can make your cool function or valuable data available to anyone who needs it In the demographic silo example, described in the earlier section Building Service-Oriented Applications, if you have a valuable mailing list and you want to give your customers access to it, you can send them the list If you do, however, they can use it forever Instead, suppose that you could bill your customers every time they use the list If you provide Web services to implemented functions on the list, you can then track actual usage of the list You can see this concept, known as provision past the boundary, everywhere Take Twitter, for example Rather than making users to go the Web page to search through billions of messages, Twitter provides a WSDL contract that gives developers the ability to search in their own applications Web services can be controlled like any other Web application can The field-level or operation-level security that used to be handled at the RDBMS level can now be built into a semipublic API