US 12,393,855 B1
Capacity planning using machine learning
Ali Jalali, Austin, TX (US); and Pranesh Vyas, Austin, TX (US)
Assigned to Amazon Technologies, Inc., Seattle, WA (US)
Filed by Amazon Technologies, Inc., Seattle, WA (US)
Filed on Apr. 22, 2021, as Appl. No. 17/237,712.
Int. Cl. G06F 15/16 (2006.01); G06N 7/01 (2023.01); G06N 20/00 (2019.01); H04L 41/0896 (2022.01); H04L 41/147 (2022.01); H04L 41/16 (2022.01); H04L 41/5009 (2022.01)
CPC G06N 7/01 (2023.01) [G06N 20/00 (2019.01); H04L 41/0896 (2013.01); H04L 41/147 (2013.01); H04L 41/16 (2013.01); H04L 41/5009 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A computer-implemented method, comprising:
determining historical demand for a computing resource of a computing resource service provider over a first time period;
determining an acceptable quality of service for the computing resource of the computing resource service provider;
determining a second time period;
performing a quantile regression on the historical demand for the computing resource to target the acceptable quality of service over the second time period;
determining, based on the quantile regression, a set of parameters for a model that approximates the historic demand for the computing resource over the first time period;
performing one or more Monte-Carlo simulations over the second time period to determine one or more innovations;
determining, based on the set of parameters and the one or more innovations, a distribution of predicted future demand values for the computing resource over the second time period;
determining a first quartile of the distribution;
determining a third quartile of the distribution;
determining, based on the first quartile and the third quartile, an amount of the computing resource that is predicted to satisfy the acceptable quality of service over the second time period; and
allocating one or more computing resources for the computing resource service provider to meet the amount of the computing resource that is predicted to satisfy the acceptable quality of service over the second time period.