US 12,293,239 B1
Processing task distribution and recovery
Tingting Guo, Hong Kong (HK); Tushar Kaistha, Hong Kong (HK); and Wing Yan Ip, Hong Kong (HK)
Assigned to Morgan Stanley Services Group Inc., New York, NY (US)
Filed by Morgan Stanley Services Group Inc., New York, NY (US)
Filed on Dec. 19, 2024, as Appl. No. 18/988,714.
Int. Cl. G06F 9/50 (2006.01)
CPC G06F 9/5083 (2013.01) [G06F 9/5038 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A method for executing tasks in a distributed computing environment, the method performed by one or more processors and comprising:
receiving training data including a plurality of workloads performed in the distributed computing environment,
the training data indicating, for each workload of the plurality of workloads, a plurality of features and an amount of time taken to perform the workload;
training a machine learning model using an objective function based on the training data,
the machine learning model learning a plurality of weights associated with the plurality of features of each workload of the plurality of workloads and related to a task weight value for the workload, and
the objective function including a first hyperparameter for controlling a number of tasks to be performed in the distributed computing environment and a parameter indicating a relationship between the task weight value assigned to each workload and an estimated amount of time required to perform the workload;
receiving a workload request that includes a new workload with a new set of feature values;
computing a new task weight value for the new workload using the machine learning model;
assigning the new workload into one or more new tasks based on the new task weight value; and
transmitting the one or more new tasks for execution in the distributed computing environment.