Date Added: May 2010
Enterprise application integrations involve the participation of several existing applications with which the integration solution exchanges data over LANs and the Internet. In these scenarios, operations might occasionally produce exceptional results at runtime due to impairments introduced by the electronic infrastructure such as node crashes, messages lost, delayed or incorrectly composed by applications. To address the problem, the paper suggests a domain specific language to specify the integration solution: it produces platform-independent models and has built-in primitives to produce events that notify of potential exceptional situations. The paper also shows how these events can be processed by an event condition action - based monitor to trigger recovery actions.