US 12,332,971 B2
Proactive service requests for forecast storage system failures
Owen Martin, Hopedale, MA (US); and Ramesh Doddaiah, Westborough, MA (US)
Assigned to Dell Products, L.P., Hopkinton, MA (US)
Filed by Dell Products, L.P., Hopkinton, MA (US)
Filed on Oct. 13, 2022, as Appl. No. 17/965,164.
Prior Publication US 2024/0126837 A1, Apr. 18, 2024
Int. Cl. G06F 18/214 (2023.01); G06F 11/07 (2006.01); G06N 3/04 (2023.01); G06N 5/04 (2023.01)
CPC G06F 18/214 (2023.01) [G06F 11/0772 (2013.01); G06N 3/04 (2013.01); G06N 5/04 (2013.01)] 20 Claims
OG exemplary drawing
 
11. A method of forecasting storage system failures and generating proactive service requests, comprising:
deploying a checkpoint of a trained machine learning process as an inference model, the machine learning process being trained to learn recursions between time series sets of Streaming Machine Telemetry (SMT) event counters as independent variables and error types as dependent variables using a set of labeled training examples, each labeled training example including a time series set of SMT event counters generated by software executing on a storage system during a set of SMT monitoring intervals preceding an error of the software, each training example also including a label identifying a type of error associated with the time series set of SMT event counters;
supplying a current time series set of SMT event counters from monitored software of a storage system to the inference model; and
predicting, by the inference model, a predicted occurrence of an error on the monitored software of the storage system before the error occurs.