US 12,086,040 B2
Prediction-based resource orchestration in a service fabric
Neda M. Pistinjat, Belgrade (RS); Nikola Puzovic, Belgrade (RS); Milan Micić, Belgrade (RS); Maja Stikic, Belgrade (RS); Nikola Pavlovic, Belgrade (RS); Jelena Petrovic, Belgrade (RS); Drazen Sumic, Belgrade (RS); Aleksa Brkic, Belgrade (RS); Vesna Todorovic, Belgrade (RS); Matthew T. Snider, Seattle, WA (US); and Ivan Nedic, Kucura (RS)
Assigned to Microsoft Technology Licensing, LLC, Redmond, WA (US)
Filed by Microsoft Technology Licensing, LLC, Redmond, WA (US)
Filed on Oct. 31, 2022, as Appl. No. 18/051,375.
Prior Publication US 2024/0143461 A1, May 2, 2024
Int. Cl. G06F 11/20 (2006.01); G06F 11/34 (2006.01)
CPC G06F 11/203 (2013.01) [G06F 11/3423 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A method, implemented by a processor, for reducing an interruption rate associated with a customer application offered as a collection of services, wherein the collection of services is offered via a service fabric cluster comprising service replicas for execution by nodes associated with the service fabric cluster, the method comprising:
using a trained machine learning model, predicting one or more quiet time periods associated with each of the service replicas, wherein each of the one or more quiet time periods corresponds to a low value of a predicted load specifying a consumption of a metric by a respective service replica on a given node; and
during the one or more quiet time periods predicted by the trained machine learning model, performing an impact-less failover for one or more of the service replicas associated with a stateful service by scheduling a move for the one or more service replicas from a first node associated with the service fabric cluster to a second node associated with the service fabric cluster such that the impact-less failover is performed to eliminate or reduce any interruptions of the customer application.