| 摘要 |
A method and apparatus for controlling data flow on a network having a server and a non-blocking client node in which the client node may not drop requests received from the server. A process on the non-blocking client node allocates resources on the non-blocking client node and notifies the server of the allocated resources. The server then determines the resources on the non-blocking client node required to fulfill a request by an application on the server and whether sufficient resources are available on the non-blocking client node to fulfill the request by the application. If sufficient resources are available, the server then sends the request to the client node and if sufficient resources do not exist, the server waits until sufficient resources do exist to issue the request to the non-blocking client node. |
