In order to provide flexibility for a developer, a network API is provided which allows for a network event (for example, the receipt of data over the network) to be handled by one thread from among a predetermined number of dedicated threads. The number of dedicated threads is specified by the application at runtime and can be modified at runtime as well. Where the number of dedicated threads specified is greater than zero, dedicated threads are created which handle network events and call the specific callback functions of the application in response to the network events. Thus the number of such callback events from threads that can occur at one time is limited by the number of dedicated threads specified by the application. When the number of dedicated threads is specified as zero, no callback events can occur except when the application assigns a certain time period for them. Thus, when the number of dedicated threads created is zero, the application has control over the amount of processing time allotted to the network layer. Network events are handled when the application issues calls a "Do Work" method. While executing this method, network events are handled.