US 12,067,251 B2
Data automation and predictive modeling for planning and procuring solid state drive replacements
Malak Alshawabkeh, Franklin, MA (US); Sunil Gumaste, Bengaluru (IN); Ravish Sachdeva, Jammu (IN); Pankaj Soni, Sikar (IN); and Christopher Allison, Milford, MA (US)
Assigned to Dell Products, L.P., Hopkinton, MA (US)
Filed by Dell Products, L.P., Hopkinton, MA (US)
Filed on Apr. 11, 2022, as Appl. No. 17/717,208.
Prior Publication US 2023/0325092 A1, Oct. 12, 2023
Int. Cl. G06F 3/06 (2006.01)
CPC G06F 3/0616 (2013.01) [G06F 3/0653 (2013.01); G06F 3/0679 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A method of predicting future aggregate % usage levels of a plurality of each monitored Solid-State Drive (SSD) of a set of monitored SSDs to prevent the aggregate % usage levels of the monitored SSDs from exceeding a repurposing threshold, comprising:
receiving % usage time series data from each of the monitored SSDs over a plurality of lags, each lag being an equal length of time;
for each SSD:
using the % usage time series data for the respective SSD to determine an order of an Auto Regression Integrated Moving Average (ARIMA) model type to be used to model the % usage time series data of the respective SSD, the determined order of the ARIMA model type being determined specifically for the respective SSD;
fitting a respective ARIMA model of the determined ARIMA model type to the % usage time series data for the respective SSD to create an SSD specific ARIMA model for each respective SSD; and
using the respective fit ARIMA model for the respective SSD in a predictive manner to forecast a first time when an expected aggregate % usage value of the respective SSD is likely to exceed a first % usage threshold value after which first forecast time the respective SSD will not able to be repurposed for another use.