Home > Articles > Programming > Java

Java EE and .NET Interoperability: Exploring Asynchronous Integration

  • Print
  • + Share This
This chapter provides a basic introduction to asynchronous integration. Asynchronous integration occurs when a number of processes integrate but do not lock for the time of the transaction on which they are integrated.
This chapter is from the book


Asynchronous integration occurs when a number of processes integrate but do not lock for the time of the transaction on which they are integrated. In general, a caller makes a request of a server and then goes away and does its own thing. When the server finishes its part of the process, it sends the results back to the caller via a callback. This involves the caller and the server keeping information about each other in what is usually referred to as a session. It involves overhead on both sides, and as such, in large scale systems it requires careful design and usage.

In a typical scenario, asynchronous integration is appropriate where a request process takes a long period of time or if a client application can continue processing without waiting for a response. Asynchronous communication is often utilized for the Enterprise Application Integration (EAI). For example, processing an order fulfillment system often relies on asynchronous processing of incoming orders. Once the order is processed, a status confirmation is sent to the user or application that placed the order.

A classic example of an asynchronous service is when one orders a ticket for an airline. Typically the scenario starts with someone making a request, and then the system goes off and checks availability of the requested resource. When ordering an airplane seat, it assigns that seat to you and blocks it off for all other sessions.

The first asynchronous session occurs when the ticket query is made. In a typical browser session you get a ‘waiting’ screen, but what is happening here is that the HTTP session is being kept alive while the query for seats is being made. Once the query is complete, the server finishes the process, downloads the new data, and closes the HTTP session. (This isn’t a perfect example because in a true asynchronous case, the user could go and browse other Web sites with the same browser and get redirected to the seat availability screen when the server is ready, but the HTTP protocol doesn’t allow for this.)

At this point, the session is being maintained on the server, usually with an expiration time of two minutes. Should the user not complete the transaction in this timeframe, the session dies, and the resources are released. The next step of the session is that you confirm you want the seats and inform the server. In this case the session stays alive while the server finalizes the booking and blocks off the seats for you.

A third asynchronous session takes place for payment, with a call to another server to validate the given credit card information. This calls back to the booking server with a positive or a negative and based on those results the booking server calls back to the browser with the results of the successful booking or a request for an alternative payment.

In this manner two or more remote systems can interoperate in a more efficient way by hiding internal system complexity and only exposing high-level services forming an asynchronous Service Oriented Architecture (SOA).

  • + Share This
  • 🔖 Save To Your Account