Browse Prior Art Database

Dynamically modifiable endurance profiles for individual devices based on array health and failures tied to other devices in the same array

IP.com Disclosure Number: IPCOM000236948D
Publication Date: 2014-May-22
Document File: 2 page(s) / 21K

Publishing Venue

The IP.com Prior Art Database

Abstract

Disclosed is a method to track extrapolated lifetime endurance curves of flash devices and alert users when the extrapolation has changed to some value outside of a configurable acceptable amount based on workload or array status.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 51% of the total text.

Page 01 of 2

Dynamically modifiable endurance profiles for individual devices based on array health and failures tied to other devices in the same array

Customers can currently use tools delivered by the server vendor to track the endurance levels of the solid-state devices in servers. A number of measurement points can be used to extrapolate a device's remaining time based on the current workload. If a device is in an array and another device in the array fails, then the write profile is altered by both a lesser number of devices doing the same workload as well as the fact that an array in critical mode produces an abnormally large number of additional Input/Outputs (IOs). If a customer takes periodic measurements to track endurance, then a critical failure of this type can cause an incorrect assumption of endurance on other devices in the array.

The novel contribution is a method to track extrapolated lifetime endurance curves of flash devices and alert users when the extrapolation has changed to some value outside of a configurable acceptable amount based on workload or array status.

In this context, endurance refers to the amount of writes a flash device can endure before reaching a point where it cannot be guaranteed the device can complete the write.

An intelligent device can identify activity changes based on both increased/decreased customer IO as well as workload changes as the result of a change in the status of the array. As an example in a Redundant Array of Independent Disks (RAID) 5 array, if one device in the array fails and the workload remains the same, then the remaining devices in the array must account for the workload of the failed device. In addition, the array's overall workload changes because the data on the failed drive must be calculated from parity, which increases workload to the remaining drives as well.

The individual devices, driver for the device, or management software, can also be altered to track the rate at which the endurance of the device is historically checked by the monitoring Command Line Interface (CLI) or Graphical User Interface (GUI). Tracking the last two points of measurement for endurance and when those points are captured allows the device to calculate the rate of usage. If the activity rate of the device changes by some level that, when coupled with the rate at which the device is periodically checked, sho...