HIGH PERFORMANCE AND GRID COMPUTING WITH PARTITIONING QUALITY OF SERVICE CONTROL

摘要

High performance computing (HPC) and grid computing processing for seismic and reservoir simulation are performed without impacting or losing processing time in case of failures. A Data Distribution Service (DDS) standard is implemented in High Performance Computing (HPC) and grid computing platforms, to avoid the shortcomings of current Message Passing Interface (MPI) communication between computing modules, and provide quality of service (QoS) for such applications. QoS properties of the processing can be controlled. A “partitioning” quality of service is provided and the computer can be logically segregated into several “logical partitions” so that a computer can have several publisher nodes, serving several groups of compute nodes, and running different applications independently.

主权项

1. A computer implemented method of computerized processing in a data processing system of data for exploration and production of hydrocarbons, the data processing system including a plurality of master nodes established as publishers of segments of exploration and production data to be separately processed by the master nodes, based on an established quality of service standard profile including a partition setting for the master nodes establishing partitions of the data processing system for separate processing of the segments of the exploration and production data, a plurality of processor nodes designated as subscribers in the established partitions to receive an assigned segment of the exploration and production data for processing from a designated one of the plurality of the publisher master nodes according to the partition setting of the quality of service standard profile, and a data memory, the method comprising the computer processing steps of:
(a) transmitting the established quality of service standard profile from the publisher master nodes to the subscriber processor nodes; (b) establishing with the publisher master nodes the partitions of the data processing system for separate processing of the segments of the exploration and production data by the publisher master nodes and the processor nodes designated as subscriber processor nodes according to the partition setting; (c) further establishing the designated subscriber processor nodes in the established partitions as data writers to transfer the processed exploration and production data to the designated publisher master nodes of the established partitions; (d) sending source data samples of the exploration and production data being processed in each of the established partitions from the publisher master node to the designated subscriber processor nodes of the partition; (e) processing the transmitted exploration and production data in the designated subscriber processor nodes of the established partitions; (f) monitoring the processed exploration and production data of the designated subscriber processor nodes at the publisher master nodes of the established partitions; (g) determining in the publisher master nodes whether the designated subscriber processor nodes of the established partitions comply with the transmitted established quality of service standard profile from the publisher master node; and (h) if so, receiving at the publisher master nodes the processed exploration and production data from the designated subscriber processor nodes of the established partitions which comply with the transmitted established quality of service standard profile; and (i) if not, inhibiting at the publisher master nodes transfer to the publisher master node of the processed exploration and production data from the designated subscriber processor nodes of the established partitions which do not comply with the transmitted established quality of service standard profile; and (j) assembling in the data memory of the data processing system the processed exploration and production data received at the publisher master nodes for the separately processed segments of the of the exploration and production data.