Machine learning for determining confidence for reclamation of storage volumes

Patent metadata

Field	Value
Publication number	US-10445659-B2
Application number	US-201615173562-A
Country	US
Kind code	B2
Filing date	Jun 3, 2016
Priority date	Jun 3, 2016
Publication date	Oct 15, 2019
Grant date	Oct 15, 2019

Abstract

A method, system and computer product for performing storage maintenance is described. A training set for storage volume reclamation is received. The training set includes a set of storage parameters, each set of storage parameters corresponds to a respective candidate storage volume of a set of candidate storage volumes. The training set also includes a set of user decisions made whether a respective candidate storage volume is reclaimable. The training set is used to train a machine learning system to recognize common features of reclaimable candidate storage volumes. A set of candidate storage volumes is provided for potential reclamation, each with a set of storage parameters. A graphical user interface presents respective members of the set of candidate storage volumes for reclamation if a confidence level is calculated that the respective candidate storage volume is reclaimable exceeds a threshold.

First claim

The invention claimed is:

1. A method for performing storage maintenance comprising:
receiving a training set for storage volume reclamation, the training set comprising a set of storage parameters, each set of storage parameters for a respective candidate storage volume of a set of candidate storage volumes and a set of user decisions made whether a respective candidate storage volume is reclaimable;
using the training set to train a machine learning system to recognize common features of reclaimable candidate storage volumes;
providing a candidate storage volume for reclamation with a set of storage parameters; and
displaying a graphical user interface presenting the candidate storage volume for reclamation if a confidence level is calculated that the candidate storage volume is reclaimable exceeds a threshold.

2. The method as recited in claim 1 , wherein the sets of storage parameters comprise last I/O, host attachments, and historical access patterns for a respective storage volume.

3. The method as recited in claim 1 , further comprising:
receiving a new set of user decisions for storage volume reclamation, the new set of user decisions comprising the user decision for reclaiming a respective candidate storage volume and the set of storage parameters for a respective candidate storage volume;
using the new set of user decisions to update the training of the machine learning to recognize common features of reclaimable candidate storage volumes; and
updating the common features of reclaimable candidate storage volumes according to the new set of user decisions.

4. The method as recited in claim 3 , wherein the new set of user decisions contains a first user decision, the first user decision is a presentation of a candidate storage volume to a first user which has remained unclaimed for greater than a predetermined time threshold.

5. The method as recited in claim 1 , further comprising:
scoring candidate storage volumes within the set of candidate storage volumes according to a comparison between common features of a set of candidate storage volume to the common features of the training set of candidate storage volumes; and
presenting a plurality of candidate storage volumes having scores exceeding a predetermined threshold.

6. The method as recited in claim 5 , wherein the plurality of candidate storage volumes are presented according to user context.

7. The method as recited in claim 6 , further comprising:
determining a set of permissions of a first user;
determining the threshold according to the set of permissions of the first user; and
wherein the threshold is set lower for a first user having a greater set of permissions than for a second user having a lesser set of permissions.

8. The method as recited in claim 6 , wherein the user context comprises a set of storage volumes the user is authorized to reclaim.

9. Apparatus, comprising:
a processor;
computer memory holding computer program instructions executed by the processor for performing database maintenance, the computer program instructions comprising:
a storage manager for managing storage volumes in a storage facility, the managing including reclaiming storage volumes;
a machine learning system for receiving a training set for storage volume reclamation and using the training set to recognize common features of reclaimable candidate storage volumes, the training set comprising a set of storage parameters, each set of storage parameters for a respective candidate storage volume of a set of candidate storage volumes and a set of decisions made whether a respective candidate storage volume is reclaimable;
program code, operative to provide a candidate storage volume for reclamation with a set of storage parameters; and
program code, operative to display a graphical user interface presenting a confidence level calculated that the candidate storage volume is reclaimable.

10. The apparatus as recited in claim 9 , further comprising:
program code, operative to receive a new set of user decisions for storage volume reclamation, the new set of user decisions comprising the user decision for reclaiming a respective candidate storage volume and the set of storage parameters for a respective candidate storage volume;
program code, operative to use the new set of user decisions to update the training of the machine learning to recognize common features of reclaimable candidate storage volumes; and
program code, operative to update the common features of reclaimable candidate storage volumes according to the new set of user decisions.

11. The apparatus as recited in claim 9 , further comprising:
program code, operative to score candidate storage volumes within the set of candidate storage volumes according to a comparison between common features of a set of candidate storage volume to the common features of the training set of candidate storage volumes; and
program code, operative to present a plurality of candidate storage volumes having scores exceeding a predetermined threshold.

12. The apparatus as recited in claim 11 , further comprising:
program code, operative to set the predetermined threshold according to user context and permissions.

13. The apparatus as recited in claim 11 , further comprising:
program code, operative to logically group storage volumes according to a set of storage customers;
program code, operative to select a first training set for a first storage customer, wherein the first training set comprises a first set of candidate storage volumes, user decisions regarding reclaiming respective members of the first set of candidate storage volumes and sets of storage parameters for respective members of the first set of candidate storage volumes;
program code, operative to select a second training set for a second storage customer, wherein the second training set comprises a second set of candidate storage volumes, user decisions regarding reclaiming respective members of the second set of candidate storage volumes and sets of storage parameters for respective members of the second set of candidate storage volumes; and
program code, operative to train a first machine learning model for storage volumes belonging to the first storage customer and a second machine learning model for storage volumes belonging to the second storage customer.

14. The apparatus as recited in claim 11 , wherein the storage manager manages a plurality of storage controllers, and the apparatus further comprises:
program code, operative to logically group storage volumes according to a set of storage controllers;
program code, operative to select a first training set for a first storage controller, wherein the first training set comprises a first set of candidate storage volumes, user decisions regarding reclaiming respective members of the first set of candidate storage volumes and sets of storage parameters for respective members of the first set of candidate storage volumes;
program code, operative to select a second training set for a second storage controller, wherein the second training set comprises a second set of candidate storage volumes, user decisions regarding reclaiming respective members of the second set of candidate storage volumes and sets of storage parameters for respective members of the second set of candidate storage volumes; and
program code, operative to train a first machine learning model for storage volumes belonging to the first storage controller and a second machine learning model for storage volumes belonging to the second storage controller.

15. A computer program product in a non-transitory comp

Assignees

• IBM

Inventors

• Bowers John A
• Laforteza Andrew J
• Mcnair Ryan D
• Randall Benjamin J
• Swingler Teresa S

Classifications

• G06F3/067

  Distributed or networked storage systems, e.g. storage area networks [SAN], network attached storage [NAS] · CPC title

• G06F3/0629

  Configuration or reconfiguration of storage systems · CPC title

• G06F3/0608

  Saving storage space on storage systems · CPC title

• G06N20/00 Primary

  Machine learning · CPC title