US 12,405,918 B2
Utilizing a machine learning model to migrate a system to a cloud computing environment
Amar Ratanlal Bafna, Palghar (IN); Susan Patricia McNamara, Lamberhurst (GB); Parag Rane, Thane West (IN); Ankit Laxmichand Dedhia, Mumbai (IN); Harsh Dhiraj Vira, Mumbai (IN); and Mayank Sudhir Singh, Mumbai (IN)
Assigned to Accenture Global Solutions Limited, Dublin (IE)
Filed by Accenture Global Solutions Limited, Dublin (IE)
Filed on Aug. 3, 2022, as Appl. No. 17/880,211.
Prior Publication US 2024/0045831 A1, Feb. 8, 2024
Int. Cl. G06F 16/11 (2019.01); G06N 20/00 (2019.01)
CPC G06F 16/119 (2019.01) [G06N 20/00 (2019.01)] 17 Claims
OG exemplary drawing
 
1. A method, comprising:
receiving, by a device, logs and files associated with a system to be migrated to a cloud computing environment;
determining, by the device and based on the logs and the files, workload data identifying workload patterns of the system;
deriving, by the device, a data lineage for source data and target data included in the logs and the files,
wherein deriving the data lineage comprises deriving particular structured data from particular semi-structured and particular unstructured data using the logs and the files, and
wherein the logs include structured data, semi-structured data, and unstructured data associated with the system and the files include execution log files associated with the system;
assessing, by the device, a utilization pattern of the system based on the logs and the files;
determining whether a distributed computing feature of the system is being utilized based on the utilization pattern of the system, wherein determining whether the distributed computing feature of the system is being utilized comprises determining whether the system optimally distributes tasks among multiple computers on a network to enable workload balancing of the system;
processing, by the device, the workload data, the data lineage, and data identifying utilization of the distributed computing feature, with a machine learning model, to label utilization features of the system and to recommend a cloud architecture;
processing, by the device, the workload data, the data lineage, and the data identifying utilization of the distributed computing feature, with a natural language processing model, to determine a cost of migrating the system to the cloud computing environment;
processing, by the device, the labelled utilization features, the cloud architecture, and the cost, with a Q-matrix model, to determine migration actions for migrating the system to the cloud computing environment,
wherein the machine learning model is trained based on:
portioning a dataset into a training dataset and a test dataset,
training the machine learning model based on the training dataset,
performing a query strategy for determining whether to accept or reject predictions obtained for the test dataset,
accepting a subset of predictions, of the predictions, associated with confidence levels of a particular threshold, and
retraining the machine learning model based on the accepted predictions; and
performing, by the device, one or more actions based on the migration actions.