| CPC G06F 11/3476 (2013.01) [G06F 11/3051 (2013.01); G06F 11/3447 (2013.01)] | 17 Claims |

|
1. A computer-implemented method comprising:
generating pre-initialization environments on at least one hardware processor for running incoming applications;
monitoring resource usage of resources configured for the pre-initialization environments in running the incoming applications;
deactivating automatic tuning of a pre-initialization environment among the pre-initialization environments responsive to determining, based on the resource usage of resources configured for the pre-initialization environment, that the pre-initialization environment is maintaining a threshold level of efficiency;
activating automatic tuning of the pre-initialization environment among the pre-initialization environments responsive to determining, based on the resource usage of resources configured for the pre-initialization environment, that the pre-initialization environment is overloaded;
activating automatic tuning of the pre-initialization environment among the pre-initialization environments responsive to determining, based on the resource usage of resources configured for the pre-initialization environment, that the pre-initialization environment is idle;
running machine learning models using features associated with the incoming applications to predict performance index associated with each incoming application of the incoming applications; and
based on mapping of the predicted performance index of an incoming application with resource configuration of the pre-initialization environment, routing the incoming application to the pre-initialization environment for the incoming application to run in the pre-initialization environment.
|