Allocating distributed storage and task execution resources

• US 9,648,087 B2
• Filed: 08/05/2013
• Issued: 05/09/2017
• Est. Priority Date: 10/08/2012
• Status: Expired due to Fees

• Current Assignee: Pure Storage, Inc.
• Original Assignee: International Business Machines Corporation
• Inventors: Alnafoosi, Ahmad
• Primary Examiner(s): Shin, Kyung H
• Application Number:
  US13/959,006
• Publication Number:
  US 20140101221A1
• Time in Patent Office: 1,373 Days
• Field of Search: 718104, 718 1, 718102, 370252, 370235, 370236, 726 27, 709201
• CPC Class Codes: G06F 11/1076 Parity data used in redundant arrays of independent storages, e.g. in RAID systems
• CPC Class Codes: G06F 11/108 Parity data distribution in semiconductor storages, e.g. in SSD
• CPC Class Codes: G06F 16/1744 using compression, e.g. sparse files
• CPC Class Codes: G06F 21/64 Protecting data integrity, e.g. using checksums, certificates or signatures
• CPC Class Codes: G06F 3/0619 in relation to data integrity, e.g. data losses, bit errors
• CPC Class Codes: G06F 3/064 Management of blocks
• CPC Class Codes: G06F 3/067 Distributed or networked storage systems, e.g. storage area networks [SAN], network attached storage [NAS]
• CPC Class Codes: G06F 9/5027 the resource being a machine, e.g. CPUs, Servers, Terminals
• CPC Class Codes: H04L 43/0852 Delays
• CPC Class Codes: H04L 47/70 Admission control; Resource allocation
• CPC Class Codes: H04L 67/10 in which an application is distributed across nodes in the network software deployment G06F8/60; multiprogramming arrangements G06F9/46
• CPC Class Codes: H04L 67/1097 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 Class Codes: H04L 67/60 Scheduling or organising the servicing of application requests, e.g. requests for application data transmissions using the analysis and optimisation of the required network resources admission control or resource allocation H04L47/70

Abstract

A method begins by a dispersed storage (DS) processing module obtaining a plurality of data access requests and a request to execute a distributed computing function that includes a plurality of partial tasks. The method continues with the DS processing module allocating execution of the plurality of partial tasks to a set of distributed storage and task execution (DSTE) units in accordance with a desired executing efficiency and allocating processing of the plurality of data access requests in accordance with a desired data access efficiency. The method continues with the DS processing module establishing the desired executing efficiency and the desired data access efficiency to obtain a desired utilization of the set of DSTE units.

Claims

1. A method for execution by one or more processing modules of one or more computing devices, the method comprises:

obtaining, by the one or more processing modules, a plurality of data access requests for data and a request to execute a distributed computing function that includes a plurality of partial tasks regarding the data, wherein the data is dispersed storage error encoded to produce one or more sets of encoded data slices;

allocating, by the one or more processing modules, execution of the plurality of partial tasks to a set of distributed storage and task execution (DSTE) units in accordance with a desired executing efficiency, wherein the desired executing efficiency distributes the execution of the plurality of partial tasks among the set of DSTE units based on respective executing resource capabilities of respective DSTE units within the set of DSTE units such that, from time-to-time, differing DSTE units of the set of DSTE units are not assigned execution of a partial task of the plurality of partial tasks;

allocating, by the one or more processing modules, processing of the plurality of data access requests to the set of DSTE units in accordance with a desired data access efficiency, wherein the desired data access efficiency distributes the processing of the plurality of data access requests among the set of DSTE units based on respective data access capabilities of respective DSTE units within the set of DSTE units such that, from time-to-time, the differing DSTE units of the set of DSTE units are not processing a data access request of the plurality of data access requests; and

establishing, by the one or more processing modules, the desired executing efficiency and the desired data access efficiency to obtain a desired utilization of the set of DSTE units based on both the respective executing resource capabilities and the respective data access capabilities of the respective DSTE units within the set of DSTE units.

2. The method of claim 1 further comprises:

obtaining, overlapping in time, the plurality of data access requests and the plurality of partial tasks; and

during the overlapping time, establishing the desired utilization of the set of DSTE units to substantially balance loading of the set of DSTE units based on per unit processing resource utilization to processing resource capabilities.

3. The method of claim 1 further comprises:

substantially equally allocating the plurality of partial tasks among the set of DSTE units; and

substantially equally allocating the plurality of data access requests to the set of DSTE units.

4. The method of claim 3 further comprises:

utilizing a pattern to provide the substantially equally allocation, wherein the pattern indicates on a DSTE unit by DSTE unit basis of when and when not a DSTE unit of the set of DSTE units is allocated the partial task of the plurality of partial tasks and when and when not the DSTE unit is allocated the data access request of the plurality of data access requests.

5. The method of claim 1 further comprises at least one of:

allocating, in an imbalanced manner, the plurality of partial tasks among the set of DSTE units; and

allocating, in the imbalanced manner, the plurality of data access requests to the set of DSTE units.

6. The method of claim 1 further comprises at least one of:

adjusting allocation of remaining ones of the plurality of partial tasks among the set of DSTE units; and

adjusting allocation of remaining ones of the plurality of data access requests to the set of DSTE units.

7. The method of claim 1 further comprises:

the plurality of data access requests including a plurality of read requests regarding reading a data file from the set of DSTE units as a plurality of sets of encoded data slices;

determining a read threshold that indicates, for a set of the plurality of sets of encoded data slices, a number of encoded data slices to read to ensure recovery of a data segment of the data file from the number of encoded data slices; and

establishing, at least in part, the desired data access efficiency based on the read threshold.

8. The method of claim 1 further comprises:

the plurality of data access requests including a plurality of write requests regarding storing a data file as a plurality of sets of encoded data slices in the set of DSTE units;

determining a write threshold that indicates, for a set of the plurality of sets of encoded data slices, a number of encoded data slices to write to the set of DSTE units to ensure proper storage of a data segment of the data file; and

establishing, at least in part, the desired data access efficiency based on the write threshold.

9. The method of claim 1 further comprises:

obtaining, by the one or more processing modules, the plurality of data access requests regarding storage of a data file; and

obtaining, by the one or more processing modules, the distributed computing function regarding the data file.

10. A distributed storage and/or task (DST) client module comprises:

an outbound DST processing module, when operable within a computing device, causes the computing device to;

obtain a plurality of data access requests for data and a request to execute a distributed computing function that includes a plurality of partial tasks regarding the data, wherein the data is dispersed error encoded to produce one or more sets of encoded data slices;

allocate execution of the plurality of partial tasks to a set of distributed storage and task execution (DSTE) units in accordance with a desired executing efficiency, wherein the desired executing efficiency distributes the execution of the plurality of partial tasks among the set of DSTE units based on respective executing resource capabilities of respective DSTE units within the set of DSTE units such that, from time-to-time, differing DSTE units of the set of DSTE units are not assigned execution of a partial task of the plurality of partial tasks;

allocate processing of the plurality of data access requests to the set of DSTE units in accordance with a desired data access efficiency, wherein the desired data access efficiency distributes the processing of the plurality of data access requests among the set of DSTE units based on respective data access capabilities of respective DSTE units within the set of DSTE units such that, from time-to-time, the differing DSTE units of the set of DSTE units are not processing a data access request of the plurality of data access requests; and

an efficiency module, when operable within the computing device, causes the computing device to;

establish the desired executing efficiency and the desired data access efficiency to obtain a desired utilization of the set of DSTE units based on both the respective executing resource capabilities and the respective data access capabilities of the respective DSTE units within the set of DSTE units.

11. The DST client module of claim 10 further comprises:

the outbound DST processing module further functions to obtain, overlapping in time, the plurality of data access requests and the plurality of partial tasks; and

during the overlapping time, the efficiency module further functions to establish the desired utilization of the set of DSTE units to substantially balance loading of the set of DSTE units based on per unit processing resource utilization to processing resource capabilities.

12. The DST client module of claim 10 further comprises:

the outbound DST processing module further functions to;

substantially equally allocate the plurality of partial tasks among the set of DSTE units; and

substantially equally allocate the plurality of data access requests to the set of DSTE units.

13. The DST client module of claim 12 further comprises:

the outbound DST processing module further functions to utilize a pattern to provide the substantially equally allocation, wherein the pattern indicates on a DSTE unit by DSTE unit basis of when and when not a DSTE unit of the set of DSTE units is allocated the partial task of the plurality of partial tasks and when and when not the DSTE unit is allocated the data access request of the plurality of data access requests.

14. The DST client module of claim 10 further comprises at least one of:

the outbound DST processing module further functions to;

allocate, in an imbalanced manner, the plurality of partial tasks among the set of DSTE units; and

allocate, in the imbalanced manner, the plurality of data access requests to the set of DSTE units.

15. The DST client module of claim 10 further comprises at least one of:

the outbound DST processing module further functions to;

adjust allocation of remaining ones of the plurality of partial tasks among the set of DSTE units; and

adjust allocation of remaining ones of the plurality of data access requests to the set of DSTE units.

16. The DST client module of claim 10 further comprises:

the outbound DST processing module further functions to;

the plurality of data access requests includes a plurality of read requests regarding reading a data file from the set of DSTE units as a plurality of sets of encoded data slices;

determine a read threshold that indicates, for a set of the plurality of sets of encoded data slices, a number of encoded data slices to read to ensure recovery of a data segment of the data file from the number of encoded data slices; and

establish, at least in part, the desired data access efficiency based on the read threshold.

17. The DST client module of claim 10 further comprises:

the outbound DST processing module further functions to;

the plurality of data access requests includes a plurality of write requests regarding storing a data file as a plurality of sets of encoded data slices in the set of DSTE units;

determine a write threshold that indicates, for a set of the plurality of sets of encoded data slices, a number of encoded data slices to write to the set of DSTE units to ensure proper storage of a data segment of the data file; and

establish, at least in part, the desired data access efficiency based on the write threshold.

18. The DST client module of claim 10 further comprises:

the outbound DST processing module further functions to;

obtain the plurality of data access requests regarding storage of a data file; and

obtain the distributed computing function regarding the data file.