Abstract: A method is provided for sharing and/or transporting data within a single node of a multi-node data processing. The method avoids the necessity of making more than one copy of the data to be shared or transported. The method is tunable based on the size of the data segment involved. A shared memory area between tasks running in different address spaces on the node is used to coordinate the process and for task to task communication. The increase in efficiency provided by the intranodal process described herein also provides advantages to the internodal communication process since more CPU cycles are available for that aspect of system operation.

Abstract: A method is provided for sharing and/or transporting data within a single node of a multi-node data processing. The method avoids the necessity of making more than one copy of the data to be shared or transported. The method is tunable based on the size of the data segment involved. A shared memory area between tasks running in different address spaces on the node is used to coordinate the process and for task to task communication. The increase in efficiency provided by the intranodal process described herein also provides advantages to the internodal communication process since more CPU cycles are available for that aspect of system operation.

Abstract: A system and method for message processing in a distributed, multi-node data processing system is structured to permit a sending process running on one node to send messages to a selectable subset of nodes via an interface mechanism which places a sending process in an inactive or idle state pending receipt of either all responses from the selected destination nodes or of a notification via the interface that one or more responses will not arrive.

Abstract: Interconnected data processing nodes can from time to time experience temporary failures or losses of communication links between system nodes. In systems which have a mechanism for determining whether the data processing nodes are operative or not, by providing an epoch number for use as an instance identifier and by providing a mechanism for controlling the generation and propagation of the epoch number, such systems are able to resume their respective tasks when failed nodes or communication links are restored.