Distributed object storage system

1. A distributed object storage system, comprising:

a plurality of storage elements configured to store and retrieve a data object, the data object stored as a first number of redundant sub blocks; and

a controller node coupled with the plurality of storage elements, the controller node including:

a processor;

a memory;

a maintenance agent stored in the memory and executable by the processor to execute a repair task;

a clustering module stored in the memory and executable by the processor to retrieve remaining redundant sub blocks of the data object and determine a number of missing redundant sub blocks of the data object;

a decoding module stored in the memory and executable by the processor to assemble a repair data object from the remaining redundant sub blocks of the data object;

an encoding module stored in the memory and executable by the processor to disassemble the repair data object into a second number of redundant sub blocks, the second number of redundant sub blocks being greater than or equal to the number of missing redundant sub blocks of the data object;

a spreading module stored in the memory and executable by the processor to store the second number of redundant sub blocks to the plurality of storage elements such that a sum of the second number of redundant sub blocks and a number of the remaining redundant sub blocks is greater than or equal to a desired spreading width for the data object; and

a monitoring agent stored in the memory and executable by the processor to:

determine an actual concurrent failure tolerance of the data object, wherein the actual concurrent failure tolerance corresponds to an actual number of storage elements that store the remaining redundant sub blocks of the data object and are allowed to fail concurrently;

determine a difference between the actual concurrent failure tolerance of the data object and a desired concurrent failure tolerance of the data object, wherein the desired concurrent failure tolerance corresponds to a maximum number of storage elements that store the first number of redundant sub blocks that are allowed to fail concurrently; and

generate the repair task in response to determining that the difference between the actual concurrent failure tolerance of the data object and the desired concurrent failure tolerance of the data object satisfies a first threshold.

2. The distributed object storage system of claim 1 , wherein:

the first number of redundant sub blocks corresponds to a predetermined multiple of the desired spreading width; and

the desired spreading width equals a sum of: a minimal spreading requirement, corresponding to a minimal number of storage elements that store the first number of redundant sub blocks that are not allowed to fail, and the desired concurrent failure tolerance.

3. The distributed object storage system of claim 2 , wherein a redundant sub block of the first number of redundant sub blocks comprises:

a data object identifier;

encoded data of a size of the data object divided by a factor equal to a predetermined multiple of the minimal spreading requirement; and

decoding data, such that the data object can be decoded from a subset of the first number of redundant sub blocks, wherein the subset includes any combination of a number of the first number of redundant sub blocks equal to the minimal spreading requirement.

4. The distributed object storage system of claim 1 , wherein the monitoring agent is further executable by the processor to generate the repair task in response to determining that the actual concurrent failure tolerance of the data object satisfies a second threshold.

5. The distributed object storage system of claim 4 , wherein the monitoring agent is further executable by the processor to:

aggregate a repair task list comprising a plurality of repair tasks associated with a corresponding plurality of data objects that are not correctly stored in the distributed object storage system; and

determine the second threshold as a function of the lowest actual concurrent failure tolerance of the plurality of data objects associated with the plurality of repair tasks.

6. A method, comprising:

identifying that a data object stored as a first number of redundant sub blocks on a plurality of storage elements in a distributed object storage system is not correctly stored in the distributed object storage system;

determining an actual concurrent failure tolerance of the data object, wherein the actual concurrent failure tolerance corresponds to an actual number of storage elements that store remaining redundant sub blocks of the data object and are allowed to fail concurrently;

determining a difference between the actual concurrent failure tolerance of the data object and a desired concurrent failure tolerance of the data object, wherein the desired concurrent failure tolerance corresponds to a maximum number of storage elements that store the first number of redundant sub blocks that are allowed to fail concurrently;

generating, in response to determining that the difference between the actual concurrent failure tolerance of the data object and the desired concurrent failure tolerance of the data object satisfies a first threshold, a repair task for the data object; and

executing the repair task, including:

retrieving remaining redundant sub blocks of the data object;

determining a number of missing redundant sub blocks of the data object;

assembling a repair data object from the remaining redundant sub blocks of the data object;

disassembling the repair data object into a second number of redundant sub blocks, the second number of redundant sub blocks being greater than or equal to the number of missing redundant sub blocks of the data object; and

storing the second number of redundant sub blocks to the plurality of storage elements such that a sum of the second number of redundant sub blocks and a number of the remaining redundant sub blocks is greater than or equal to a desired spreading width for the data object.

7. The method of claim 6 , wherein:

the first number of redundant sub blocks corresponds to a predetermined multiple of the desired spreading width; and

the desired spreading width equals a sum of: a minimal spreading requirement, corresponding to a minimal number of storage elements that store the first number of redundant sub blocks that are not allowed to fail, and the desired concurrent failure tolerance.

8. The method of claim 7 , wherein a redundant sub block of the first number of redundant sub blocks comprises:

a data object identifier;

encoded data of a size of the data object divided by a factor equal to a predetermined multiple of the minimal spreading requirement; and

decoding data, such that the data object can be decoded from a subset of the first number of redundant sub blocks, wherein the subset includes any combination of a number of the first number of redundant sub blocks equal to the minimal spreading requirement.

9. The method of claim 6 , further comprising generating the repair task in response to determining that the actual concurrent failure tolerance of the data object satisfies a second threshold.

10. The method of claim 9 , further comprising:

aggregating a repair task list comprising a plurality of repair tasks associated with a corresponding plurality of data objects that are not correctly stored in the distributed object storage system; and

determining the second threshold as a function of the lowest actual concurrent failure tolerance of the plurality of data objects associated with the plurality of repair tasks.

11. A distributed object storage system, comprising:

means for storing and retrieving a data object, the data object stored as a first number of redundant sub blocks;

a processor;

a memory;

means, stored in the memory and executable by the processor, for executing a repair task;

means, stored in the memory and executable by the processor, for retrieving remaining redundant sub blocks of the data object and determining a number of missing redundant sub blocks of the data object;

means, stored in the memory and executable by the processor, for assembling a repair data object from the remaining redundant sub blocks of the data object;

means, stored in the memory and executable by the processor, for disassembling the repair data object into a second number of redundant sub blocks, the second number of redundant sub blocks being greater than or equal to the number of missing redundant sub blocks of the data object;

means, stored in the memory and executable by the processor, for storing the second number of redundant sub blocks such that a sum of the second number of redundant sub blocks and a number of the remaining redundant sub blocks is greater than or equal to a desired spreading width for the data object;

means, stored in the memory and executable by the processor, for determining an actual concurrent failure tolerance of the data object, wherein the actual concurrent failure tolerance corresponds to an actual number of storage elements that store the remaining redundant sub blocks of the data object and are allowed to fail concurrently;

means, stored in the memory and executable by the processor, for determining a difference between the actual concurrent failure tolerance of the data object and a desired concurrent failure tolerance of the data object, wherein the desired concurrent failure tolerance corresponds to a maximum number of storage elements that store the first number of redundant sub blocks that are allowed to fail concurrently; and

means, stored in the memory and executable by the processor, for generating the repair task in response to determining that the difference between the actual concurrent failure tolerance of the data object and the desired concurrent failure tolerance of the data object satisfies a first threshold.

12. The distributed object storage system of claim 11 , wherein:

the first number of redundant sub blocks corresponds to a predetermined multiple of the desired spreading width; and

the desired spreading width equals a sum of: a minimal spreading requirement, corresponding to a minimal number of storage elements that store the first number of redundant sub blocks that are not allowed to fail, and the desired concurrent failure tolerance.

13. The distributed object storage system of claim 11 , further comprising means for generating the repair task in response to determining that the actual concurrent failure tolerance of the data object satisfies a second threshold.

14. The distributed object storage system of claim 13 , further comprising:

means for aggregating a repair task list comprising a plurality of repair tasks associated with a corresponding plurality of data objects that are not correctly stored in the distributed object storage system; and

means for determining the second threshold as a function of the lowest actual concurrent failure tolerance of the plurality of data objects associated with the plurality of repair tasks.

15. The distributed object storage system of claim 12 , wherein a redundant sub block of the first number of redundant sub blocks comprises:

a data object identifier;

encoded data of a size of the data object divided by a factor equal to a predetermined multiple of the minimal spreading requirement; and

decoding data, such that the data object can be decoded from a subset of the first number of redundant sub blocks, wherein the subset includes any combination of a number of the first number of redundant sub blocks equal to the minimal spreading requirement.