Prescriptive analytics based storage class placement stack for cloud computing

Patent metadata

Field	Value
Publication number	US-10742567-B2
Application number	US-201816219435-A
Country	US
Kind code	B2
Filing date	Dec 13, 2018
Priority date	Dec 13, 2018
Publication date	Aug 11, 2020
Grant date	Aug 11, 2020

Abstract

A multi-layer storage class placement stack may generate a token containing storage class placement prescriptions for controlling the placement of stored items within a selection of classes for storage. An input layer of the storage class placement stack may generate time-collated activity data based on historical access data, volume metric data, and/or tagging data. The time-collated activity data may include data groupings using timestamps or other timing indicators. A transformation layer may further process the time-collated activity data to generate defined-period summation data that provides summary detail for defined durations across a period of analysis. The defined-period summation data may be used by a prescriptive engine layer to generate prescriptions for placement of individual stored items by associating the prescriptions with storage identifiers for the individual items. Using the prescriptions, the prescriptive engine layer may generate the token to send to a host interface for management of the stored items.

First claim

What is claimed is:

1. A system including:
network interface circuitry configured to:
receive historical access data for a selected stored item;
receive volume metric data for the selected stored item;
send a storage class placement (SCP) token to a host interface, the host interface configured to control storage class placement for at least the selected stored item;
placement circuitry in data communication with the network interface circuitry, the placement circuitry configured to execute a SCP stack,
the SCP stack including:
a data-staging layer;
an input layer;
a transformation layer; and
a prescriptive engine layer;

the SCP stack executable to:
obtain, via the input layer, the historical access data, and the volume metric data;
process, at the input layer, the historical access data and volume metric data to generate time-collated activity data over a historical analysis time-window;
store, at the data-staging layer, the time-collated activity data;
process, at the transformation layer, the time-collated activity data to generate defined-period summation data that includes a series of summations spanning the historical analysis time-window, the summations for durations equal to a defined-period;
store, at the data-staging layer, the defined-period summation data;
access, at the prescriptive engine layer, the defined-period summation data via a memory resource provided by the data-staging layer;
based on the defined-period summation data, determine, at the prescriptive engine layer, an efficiency coefficient for a placement selection associated with a selected storage class for the selected stored item;
based on the efficiency coefficient, perform a class determination whether to execute the placement selection by comparing the efficiency coefficient to other efficiency coefficients for the selected stored item; and
based on the class determination, generate the SCP token.

2. The system of claim 1 , where:
the SCP stack is further executable to, at a presentation layer, generate a SCP control interface including a SCP-window presentation; and
the SCP-window presentation includes a selectable option to execute a placement selection.

3. The system of claim 2 , where the placement selection is grouped within the SCP-window presentation with other placement selections each corresponding to efficiency coefficients within a pre-defined range.

4. The system of claim 2 , where the SCP-window presentation further includes a selectable option to reject execution of the placement selection.

5. The system of claim 2 , where the SCP-window presentation is configured to display multiple placement selections each with a corresponding efficiency coefficient for the selected stored item.

6. The system of claim 1 , where the placement circuitry is configured to perform the class determination by comparing the efficiency coefficient to a threshold.

7. The system of claim 1 , where the historical access data includes at least 90 days of historical access data for the selected stored item.

8. The system of claim 1 , where the selected stored item is associated with a storage identifier.

9. The system of claim 1 , where the placement circuitry is configured to determine the efficiency coefficient by iteratively analyzing, in accord with corresponding storage identifiers, defined-period summation data for multiple stored items including the selected stored item.

10. The system of claim 1 , where the placement circuitry is configured to determine the efficiency coefficient based on a feedback history generated using previous command inputs from a SCP control interface generated at a presentation layer of the SCP stack.

11. A method including:
at network interface circuitry:
receiving historical access data for a selected stored item; and
receiving volume metric data for the selected stored item;

at placement circuitry in data communication with the network interface circuitry, the placement circuitry executing a storage class placement (SCP) stack:
obtaining, via an input layer of the SCP stack, the historical access data, and the volume metric data;
processing, at the input layer, the historical access data and volume metric data to generate time-collated activity data over a historical analysis time-window;
storing, at a data-staging layer of the SCP stack, the time-collated activity data;
processing, at a transformation layer of the SCP stack, the time-collated activity data to generate defined-period summation data that includes a series of summations spanning the historical analysis time-window, the summations for durations equal to a defined-period;
storing, at the data-staging layer, the defined-period summation data;
accessing, at a prescriptive engine layer of the SCP stack, the defined-period summation data via a memory resource provided by the data-staging layer;
based on the defined-period summation data, determining, at the prescriptive engine layer, an efficiency coefficient for a storage class for the selected stored item;
based on the efficiency coefficient, performing a class determination whether to place the selected stored item within the storage class by comparing the efficiency coefficient to other efficiency coefficients for the selected stored item; and
based on the class determination, generating, at the prescriptive engine layer, an SCP token; and

sending, via network interface circuitry, the SCP token to a host interface configured to control storage class placement for at least the selected stored item.

12. The method of claim 11 , where performing the class determination includes comparing the efficiency coefficient to a threshold.

13. The method of claim 11 , where the selected stored item is associated with a storage identifier.

14. The method of claim 11 , determining the efficiency coefficient includes iteratively analyzing, in accord with corresponding storage identifiers, defined-period summation data for multiple stored items including the selected stored item.

15. The method of claim 11 , where determining the efficiency coefficient includes determining the efficiency coefficient based on a feedback history generated using previous command inputs from a SCP control interface generated at a presentation layer of the SCP stack.

16. A product including:
machine-readable media other than a transitory signal; and
instructions stored on the machine-readable media, the instructions configured to, when executed, cause a machine to:
at network interface circuitry:
receive historical access data for a selected stored item; and
receive volume metric data for the selected stored item;

at placement circuitry in data communication with the network interface circuitry, the placement circuitry executing a storage class placement (SCP) stack:
obtain, via an input layer of the SCP stack, the historical access data, and the volume metric data;
process, at the input layer, the historical access data and volume metric data to generate time-collated activity data over a historical analysis time-window;
store, at a data-staging layer of the SCP stack, the time-collated activity data;
process, at a transformation layer of the SCP stack, the time-collated activity data to generate defined-period summation data that includes a series of summations spanning the historical analysis time-window, the summations for durations equal to a defined-period;
store, at the data-staging layer, the defined-period summation data;
access, at a prescriptive engine layer of the SCP stack, the defined-period summation data via a memory resource provided b

Assignees

• Accenture Global Solutions Ltd

Inventors

• Srinivasan Madhan Kumar
• Purushothaman Arun
• Pv Guruprasad
• Kolachalam Manish Sharma

Classifications

• H04L47/83

  based on usage prediction · CPC title

• G06F3/0653 Primary

  Monitoring storage devices or systems · CPC title

• G06F3/0604

  Improving or facilitating administration, e.g. storage management · CPC title

• H04L67/1097

  for distributed storage of data in networks, e.g. transport arrangements for network file system [NFS], storage area networks [SAN] or network attached storage [NAS] · CPC title

• G06F3/0631

  by allocating resources to storage systems · CPC title