Distributed System Failure

In a local INtime node, failures are easy to handle: if the whole system fails, there is no life for Windows nor RT anymore. If only Windows crashes (blue screen), an event is signaled (see the Error and Exception Handling chapter).

But what if a remote node dies, a network link gets broken or a Windows program communicating with a remote RT process is terminated?

To handle these kinds of problems, we introduce two process attributes: sponsor and dependent.