US 12,260,251 B2
Sharing of compute resources between the virtualized radio access network (vRAN) and other workloads
Xenofon Foukas, Cambridge (GB); and Bozidar Radunovic, Cambridge (GB)
Assigned to Microsoft Technology Licensing, LLC, Redmond, WA (US)
Filed by Microsoft Technology Licensing, LLC, Redmond, WA (US)
Filed on Jul. 28, 2020, as Appl. No. 16/941,033.
Prior Publication US 2022/0035665 A1, Feb. 3, 2022
Int. Cl. G06F 9/46 (2006.01); G06F 9/48 (2006.01); G06F 9/50 (2006.01); G06N 5/01 (2023.01); G06N 20/00 (2019.01)
CPC G06F 9/4887 (2013.01) [G06F 9/505 (2013.01); G06N 5/01 (2023.01); G06N 20/00 (2019.01); G06F 2209/5019 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A method, comprising:
identifying a plurality of virtualized Radio Access Networks (vRAN) workloads and a plurality of other workloads;
monitoring and recording vRAN traffic characteristics of a current base station state of a base station used to perform the plurality of vRAN workloads in real time during each transmission slot in a memory;
predicting a worst case execution time for each signal processing task of the plurality of vRAN workloads by using a plurality of quantile decision trees based on the vRAN traffic characteristics recorded, wherein the worst case execution time indicate a prediction for how long a signal processing task will take to complete;
calculating a number of compute resources required for the plurality of vRAN workloads to use to complete by a deadline by determining a difference between the predicted worst case execution time for each signal processing task of the plurality of vRAN workloads and an amount of time elapsed in a transmission slot for each signal processing task divided by a remaining amount of time to meet the deadline;
scheduling the plurality of vRAN workloads across the number of compute resources required;
executing the plurality of vRAN workloads based on the scheduling of the plurality of vRAN workloads;
determining a remaining number of available compute resources after scheduling the plurality of vRAN workload across the number of compute resources required;
scheduling the plurality of other workloads across the remaining number of available compute resources; and
executing the plurality of other workloads based on the scheduling of the plurality of other workloads.