DISTRIBUTED WORKLOAD REMAPPING AFTER COMMUNICATION FAILURE Russian patent published in 2019 - IPC G06F9/50 G06F9/52 H04L29/14 G06F11/20 G06F16/90 

FIELD: calculating; counting.

SUBSTANCE: invention relates to computer engineering. Disclosed is a computer-implemented method of processing a workload in a communication failure, comprising steps of: assigning, by a primary node, a workload to a first node; correlating the first generation identifier and the workload identifier with the workload; transmitting a first generation identifier, a workload identifier and a work load to a first node; receiving from first node request for access to node in data storage system for access to target data, wherein the first generation identifier and the workload identifier are provided with the request for identifying the first node and the corresponding workload; during communication failure between the main node and the first node, generating a second generation identifier, associating the second generation identifier with the workload and transmitting the second generation identifier, the workload identifier and the workload to the second node, wherein the second generation identifier indicates a priority different from the first generation identifier; receiving from the second node a request for access to said node in the data storage system for accessing the target data, wherein the second generation identifier and the workload identifier are provided with the request to identify the second node and the workload; determining, by a data storage system, whether the second generation identifier indicates a higher or lower priority than the first generation identifier; and determining whether the access request received from the second node with respect to the target data is satisfied based on the priorities indicated by the first and second generation identifiers, respectively.

EFFECT: technical result is more economical use of computational resources when processing workload in case of communication failure.

4 cl, 9 dwg