Managing distributed system performance using accelerated data retrieval operations
A distributed system is adapted to manage the performance of distributed processes. In one aspect, multiple stripes associated with a data item are stored in a distributed storage. The stored stripes include one or more stripes of redundancy information for the data item. A distributed process including at least one task is performed. During performance of the distributed process, a determination is made as to whether to perform an accelerated data retrieval operation. Responsive to a determination to perform an accelerated data retrieval operation, at least one of the one or more stripes of redundancy information for the data item is requested from the distributed storage. Other stripes associated with the data item may also be requested from the distributed storage. After a sufficient subset of stripes associated with the data item is received, the data item is reconstructed using the subset.
1. A computer-implemented method for managing performance of a distributed system, the method comprising:
storing, in a plurality of storage devices of a distributed storage, a plurality of stripes associated with a data item, the plurality of stripes generated according to a coding scheme, wherein the coding scheme generates a number of stripes associated with the data item that is more than a minimum number of stripes needed to reconstruct the data item, and wherein each stripe in the plurality of stripes includes redundancy information for the data item;
performing a distributed process including a task that requires retrieval of the data item from the distributed storage;
determining a processing speed associated with the task, the processing speed comprising an amount of data processed divided by a pre-defined time; and
responsive to determining the processing speed does not meet a threshold, performing an accelerated data retrieval operation by:
requesting more than the minimum number of stripes needed to reconstruct the data item from at least two of the plurality of storage devices of the distributed storage;
determining whether a sufficient subset of the requested stripes has been received, the sufficient subset comprising at least the minimum number of stripes required to reconstruct the data item; and
responsive to a determination that the sufficient subset has been received, reconstructing the data item using the sufficient subset of the requested stripes.
2. The computer-implemented method of claim 1 , wherein the plurality of stripes associated with the data item is generated according to an error correction coding scheme.
3. The computer-implemented method of claim 1 , wherein the stripes of the plurality of stripes are each stored on a different storage device of the distributed storage.
4. The computer-implemented method of claim 1 further comprising:
storing a plurality of copies of the plurality of stripes associated with the data item in the distributed storage, and
wherein performing the accelerated data retrieval operation further comprises requesting the plurality of copies of the plurality of stripes associated with the data item from the distributed storage.
5. The method of claim 1 , wherein:
the distributed process comprises a plurality of additional tasks, each of the plurality of additional tasks having a respective processing speed, each respective processing speed comprising an amount of data processed by the respective additional task divided by the pre-defined time, an average processing speed of the plurality of additional tasks comprising a sum of the processing speeds of the additional tasks divided by the number of additional tasks; and
the threshold comprises the average processing speed of the plurality of additional tasks.
6. The method of claim 1 , wherein:
the distributed process comprises a plurality of additional tasks, each of the plurality of additional tasks having a respective processing speed, each respective processing speed comprising an amount of data processed by the respective additional task divided by the pre-defined time, an average processing speed of the plurality of additional tasks comprising a sum of the processing speeds of the additional tasks divided by the number of additional tasks; and
the threshold comprises a fraction of the average processing speed of the plurality of additional tasks.
7. The method of claim 1 , wherein:
the distributed process comprises a plurality of additional tasks, each of the plurality of additional tasks having a respective processing speed, each respective processing speed comprising an amount of data processed by the respective additional task divided by the pre-defined time; and
the threshold comprises a selected additional task's processing speed, the selected additional task selected for having a slowest processing speed of the plurality of additional tasks.
8. A non-transitory computer readable storage medium executing computer program instructions for managing performance of a distributed system, the computer program instructions comprising instructions for:
storing, in a plurality of storage devices of distributed storage, a plurality of stripes associated with a data item, the plurality of stripes generated according to a coding scheme, wherein the coding scheme generates a number of stripes associated with the data item that is more than a minimum number of stripes needed to reconstruct the data item, and wherein each stripe in the plurality of stripes includes redundancy information for the data item;
performing a distributed process including a task that requires retrieval of the data item from the distributed storage;
determining a processing speed associated with the task, the processing speed comprising an amount of data processed divided by a pre-defined time; and
responsive to determining the processing speed does not meet a threshold, performing an accelerated data retrieval operation by:
requesting more than the minimum number of stripes needed to reconstruct the data item from at least two of the plurality of storage devices of the distributed storage;
determining whether a sufficient subset of the requested stripes has been received, the sufficient subset comprising at least the minimum number of stripes required to reconstruct the data item; and
responsive to a determination that the sufficient subset has been received, reconstructing the data item using the sufficient subset of the requested stripes.
9. The medium of claim 8 , wherein the plurality of stripes associated with the data item is generated according to an error correction coding scheme.
10. The medium of claim 8 , wherein the stripes of the plurality of stripes are each stored on a different storage device of the distributed storage.
11. The medium of claim 8 , the computer program instructions further comprising instructions for:
storing a plurality of copies of the plurality of stripes associated with the data item in the distributed storage, and
wherein performing the accelerated data retrieval operation further comprises requesting the plurality of copies of the plurality of stripes associated with the data item from the distributed storage.
12. The medium of claim 8 , wherein:
the distributed process comprises a plurality of additional tasks, each of the plurality of additional tasks having a respective processing speed, each respective processing speed comprising an amount of data processed by the respective additional task divided by the pre-defined time, an average processing speed of the plurality of additional tasks comprising a sum of the processing speeds of the additional tasks divided by the number of additional tasks; and
the threshold comprises the average processing speed of the plurality of additional tasks.
13. The medium of claim 8 , wherein:
the distributed process comprises a plurality of additional tasks, each of the plurality of additional tasks having a respective processing speed, each respective processing speed comprising an amount of data processed by the respective additional task divided by the pre-defined time, an average processing speed of the plurality of additional tasks comprising a sum of the processing speeds of the additional tasks divided by the number of additional tasks; and
the threshold comprises a fraction of the average processing speed of the plurality of additional tasks.
14. The medium of claim 8 , wherein:
the distributed process comprises a plurality of additional tasks, each of the plurality of additional tasks having a respective processing speed, each respective processing speed comprising an amount of data processed by the respective additional task divided by the pre-defined time; and
the threshold comprises a selected additional task's processing speed, the selected additional task selected for having a slowest processing speed of the plurality of additional tasks.
15. A system comprising:
a computer readable storage medium storing processor-executable computer program instructions for managing performance of a distributed system, the instructions comprising instructions for:
storing, in a plurality of storage devices of a distributed storage, a plurality of stripes associated with a data item, the plurality of stripes generated according to a coding scheme, wherein the coding scheme generates a number of stripes associated with the data item that is more than a minimum number of stripes needed to reconstruct the data item, and wherein each stripe in the plurality of stripes includes redundancy information for the data item;
performing a distributed process including a task that requires retrieval of the data item from the distributed storage;
determining a processing speed associated with the task, the processing speed comprising an amount of data processed divided by a pre-defined time; and
responsive to determining the processing speed does not meet a threshold, performing an accelerated data retrieval operation by:
requesting more than the minimum number of stripes needed to reconstruct the data item from at least two of the plurality of storage devices of the distributed storage;
determining whether a sufficient subset of the requested stripes has been received, the sufficient subset comprising at least the minimum number of stripes required to reconstruct the data item; and
responsive to a determination that the sufficient subset has been received, reconstructing the data item using the sufficient subset of the requested stripes;
and
a processor for executing the computer program instructions.
16. The system of claim 15 , wherein the plurality of stripes associated with the data item is generated according to an error correction coding scheme.
17. The system of claim 15 , wherein the stripes of the plurality of stripes are each stored on a different storage device of the distributed storage.
18. The system of claim 15 , the instructions further comprising instructions for:
storing a plurality of copies of the plurality of stripes associated with the data item in the distributed storage, and
wherein performing the accelerated data retrieval operation further comprises requesting the plurality of copies of the plurality of stripes associated with the data item from the distributed storage.
19. The system of claim 15 , wherein:
the distributed process comprises a plurality of additional tasks, each of the plurality of additional tasks having a respective processing speed, each respective processing speed comprising an amount of data processed by the respective additional task divided by the pre-defined time, an average processing speed of the plurality of additional tasks comprising a sum of the processing speeds of the additional tasks divided by the number of additional tasks; and
the threshold comprises a fraction of the average processing speed of the plurality of additional tasks.
20. The system of claim 15 , wherein:
the distributed process comprises a plurality of additional tasks, each of the plurality of additional tasks having a respective processing speed, each respective processing speed comprising an amount of data processed by the respective additional task divided by the pre-defined time; and
the threshold comprises a selected additional task's processing speed, the selected additional task selected for having a slowest processing speed of the plurality of additional tasks.