WRITE THRESHOLD UTILIZATION IN A DISPERSED STORAGE SYSTEM

• US 20110126042A1
• Filed: 08/25/2010
• Published: 05/26/2011
• Est. Priority Date: 11/25/2009
• Status: Active Grant

• Current Assignee: Pure Storage, Inc.
• Original Assignee: Cleversafe Incorporated (International Business Machines Corporation)
• Inventors: DHUSE, GREG
• Granted Patent: US 8,621,268 B2
• Time in Patent Office: Days
• US Class Current: 714/5.11
• CPC Class Codes: G06F 11/1008 in individual solid state devices G06F11/1004 takes precedence
• CPC Class Codes: G06F 11/2094 Redundant storage or storage space G06F11/2056 takes precedence

Abstract

A method begins with a processing module sending a plurality of dispersed storage write commands to plurality of dispersed storage (DS) units for storing a plurality of encoded data slices. The method continues with the processing module receiving, within a time period, acknowledgements from at least some of the plurality of DS units to produce received acknowledgements. The method continues with the processing module determining whether a number of received acknowledgements compares favorably to a write threshold. The method continues with the processing module changing at least one of the write threshold and at least one of the plurality of DS units when the number of received acknowledgements does not compare favorably to the write threshold.

Claims

1. A method for execution by a dispersed storage processing module, the method comprises:

sending a plurality of dispersed storage (DS) write commands to plurality of dispersed storage (DS) units for storing a plurality of encoded data slices;

receiving, within a time period, acknowledgements from at least some of the plurality of DS units to produce received acknowledgements;

determining whether a number of received acknowledgements compares favorably to a write threshold; and

when the number of received acknowledgements does not compare favorably to the write threshold, changing at least one of the write threshold and at least one of the plurality of DS units.

2. The method of claim 1 further comprises:

determining the write threshold based on storage requirements and an estimated performance; and

when the write threshold is to be changed;

determining second storage requirements;

determining second estimated performance; and

determining a changed write threshold based on the second storage requirements and second estimated performance.

3. The method of claim 2 further comprises:

determining the storage requirements or the second storage requirements based on one or more of;

a vault lookup, a user identity, a data type, a priority indicator, an availability indicator, a security indicator, a performance indicator, and an estimated frequency of retrieval indicator, wherein each of the storage requirements and the second storage requirements includes one or more of an availability requirement, a reliability requirement, a requirements preference indicator, and a write threshold enforcement indicator.

4. The method of claim 2 further comprises:

determining the estimated performance or the second estimated performance based on one or more of;

a history of receiving the acknowledgments, a history of availability, a history of reliability, a history of access latency, a history of access bandwidth, a vault lookup, and a real-time measurement, wherein the estimated performance or the second estimated performance includes at least one of estimated probability of receiving the acknowledgments, estimated availability, estimated reliability, estimated access latency, and estimated access bandwidth.

5. The method of claim 1 further comprises:

determining the write threshold based on at least one of;

selecting a slicing pillar width as the write threshold when storage requirements indicate retrieval reliability preference over write availability;

selecting a read threshold as the write threshold when the storage requirements indicate the write availability preference over the retrieval reliability; and

selecting the write threshold to be between the slicing pillar width and the read threshold when the storage requirements indicate that the write availability preference is comparable to the retrieval reliability.

6. The method of claim 1 further comprises:

when the number of received acknowledgments compares favorably to the write threshold;

updating a history of receiving the acknowledgments;

sending a commit command to the plurality of DS units;

receiving a second number of commit acknowledgment messages; and

sending a finalize command to the plurality of DS units.

7. The method of claim 1 further comprises:

when the number of received acknowledgments compares unfavorably to the write threshold, determining whether a write threshold enforcement indicator is active;

when the threshold enforcement indicator is active;

sending an error message;

identifying DS units of the plurality of DS units that did not provide an acknowledgment to produce uncommitted DS units; and

re-sending some of the plurality of DS write commands to the uncommitted DS units until the number of received acknowledgements compares favorably to the write threshold.

8. A method comprises:

selecting a write threshold to a first range of values between a read threshold and a slicing pillar width for a first storage preference, wherein the write threshold indicates a level of dispersed storage (DS) units required to acknowledge a DS write request; and

selecting the write threshold to a second range of values between the read threshold and the slicing pillar width for a second storage preference.

9. The method of claim 8 further comprises:

selecting the write threshold to a third range of values between the read threshold and the slicing pillar width for a third storage preference.

10. The method of claim 9 further comprises:

the first storage preference corresponding to a retrieval reliability preference over write availability;

the second storage preference corresponding to the write availability preference over the retrieval reliability; and

the third storage preference corresponding to the write availability preference is comparable to the retrieval reliability.

11. The method of claim 8 further comprises:

determining the first storage preference or the second storage preference based on one or more of;

a history of receiving acknowledgments, a history of availability, a history of reliability, a history of access latency, a history of access bandwidth, a vault lookup, and a real-time measurement, wherein the estimated performance includes at least one of estimated probability of receiving the acknowledgments, estimated availability, estimated reliability, estimated access latency, and estimated access bandwidth; and

selecting a plurality of DS units based on the first storage preference or the second storage preference.

12. A computer comprises:

an interface; and

a processing module operable to;

send, via the interface, a plurality of dispersed storage (DS) write commands to plurality of dispersed storage (DS) units for storing a plurality of encoded data slices;

receive, via the interface, within a time period, acknowledgements from at least some of the plurality of DS units to produce received acknowledgements;

determine whether a number of received acknowledgements compares favorably to a write threshold; and

change at least one of the write threshold and at least one of the plurality of DS units when the number of received acknowledgements does not compare favorably to the write threshold.

13. The computer of claim 12, wherein the processing module further functions to:

determine the write threshold based on storage requirements and an estimated performance; and

when the write threshold is to be changed;

determine second storage requirements;

determine second estimated performance; and

determine a changed write threshold based on the second storage requirements and second estimated performance.

14. The computer of claim 13, wherein the processing module further functions to:

determine the storage requirements or the second storage requirements based on one or more of;

a vault lookup, a user identity, a data type, a priority indicator, an availability indicator, a security indicator, a performance indicator, and an estimated frequency of retrieval indicator, wherein each of the storage requirements and the second storage requirements includes one or more of an availability requirement, a reliability requirement, a requirements preference indicator, and a write threshold enforcement indicator.

15. The computer of claim 13, wherein the processing module further functions to:

determine the estimated performance or the second estimated performance based on one or more of;

a history of receiving the acknowledgments, a history of availability, a history of reliability, a history of access latency, a history of access bandwidth, a vault lookup, and a real-time measurement, wherein the estimated performance or the second estimated performance includes at least one of estimated probability of receiving the acknowledgments, estimated availability, estimated reliability, estimated access latency, and estimated access bandwidth.

16. The computer of claim 12, wherein the processing module further functions to:

determine the write threshold based on at least one of;

selecting a slicing pillar width as the write threshold when storage requirements indicate retrieval reliability preference over write availability;

selecting a read threshold as the write threshold when the storage requirements indicate the write availability preference over the retrieval reliability; and

selecting the write threshold to be between the slicing pillar width and the read threshold when the storage requirements indicate that the write availability preference is comparable to the retrieval reliability.

17. The computer of claim 12, wherein the processing module further functions to:

when the number of received acknowledgments compares favorably to the write threshold;

update a history of receiving the acknowledgments;

send, via the interface, a commit command to the plurality of DS units;

receive, via the interface, a second number of commit acknowledgment messages; and

send, via the interface, a finalize command to the plurality of DS units.

18. The computer of claim 12, wherein the processing module further functions to:

when the number of received acknowledgments compares unfavorably to the write threshold, determine whether a write threshold enforcement indicator is active;

when the threshold enforcement indicator is active;

send, via the interface, an error message;

identify DS units of the plurality of DS units that did not provide an acknowledgment to produce uncommitted DS units; and

re-send, via the interface, some of the plurality of DS write commands to the uncommitted DS units until the number of received acknowledgements compares favorably to the write threshold.

19. A computer comprises:

an interface; and

a processing module operable to;

select a write threshold to a first range of values between a read threshold and a slicing pillar width for a first storage preference, wherein the write threshold indicates a level of dispersed storage (DS) units required to acknowledge a DS write request; and

select the write threshold to a second range of values between the read threshold and the slicing pillar width for a second storage preference.

20. The computer of claim 19, wherein the processing module further functions to:

select the write threshold to a third range of values between the read threshold and the slicing pillar width for a third storage preference.

21. The computer of claim 20, wherein the write threshold further comprises one of:

the first storage preference corresponding to a retrieval reliability preference over write availability;

the second storage preference corresponding to the write availability preference over the retrieval reliability; and

the third storage preference corresponding to the write availability preference is comparable to the retrieval reliability.

22. The computer of claim 19, wherein the processing module further functions to:

determine the first storage preference or the second storage preference based on one or more of;

a history of receiving acknowledgments via the interface, a history of availability, a history of reliability, a history of access latency, a history of access bandwidth, a vault lookup, and a real-time measurement, wherein the estimated performance includes at least one of estimated probability of receiving the acknowledgments via the interface, estimated availability, estimated reliability, estimated access latency, and estimated access bandwidth; and

selecting a plurality of DS units based on the first storage preference or the second storage preference.