System and method for improving performance of backup storage system with future access prediction
View Patent ↗A request is received to read a portion of a file from the storage system. In response to the request, future requests to the file are predicted based in part on the received request and metadata of the file. Access of a subsequent portion of the file is optimized based on the predicted future requests and the data contained in the portion of the file is then returned.
1. A computer-implemented method for providing a storage system, the method comprising:
receiving, by an access manager executed by a processor, a request from a client application to read a first portion of a file from a persistent storage device of the storage system, wherein the file references a plurality of deduplicated data segments stored in the storage system that are shared by a plurality of files;
predicting, by an access predictor executed by the processor, future requests to the file based in part on the received request and metadata of the file, including determining whether any deduplicated segment referenced by the first portion will likely be accessed subsequently in a second portion of the file;
optimizing access of a subsequent portion of the file based on the predicted future requests, including retaining a first deduplicated segment in a cache memory device in response to determining that the first deduplicated segment will be accessed subsequently in the second portion of the file; and
returning data contained in the first portion of the file to the client application, wherein the data is constructed based on deduplicated segments retrieved from the cache memory device and the persistent storage device.
2. The method of claim 1 , wherein the storage system is a deduplicating storage system.
3. The method of claim 1 , wherein optimizing the storage system includes managing a data cache based on the predicted future requests.
4. The method of claim 1 , wherein optimizing access of the subsequent portion of the file includes prefetching data into a data cache based on the predicted future requests.
5. The method of claim 1 , wherein optimizing access of the subsequent portion of the file includes scheduling an order in which requests are handled based on the predicted future requests.
6. The method of claim 1 , wherein predicting future requests is based in part on the type of file.
7. The method of claim 6 , wherein if the file contains a backup image, predicting future requests includes predicting that the file will be read sequentially from beginning to end.
8. The method of claim 1 , further comprising receiving from a software program an indication of a task and predicting future requests based on the received task.
9. The method of claim 8 , wherein the software program is a backup software and the task is to restore from backup, and predicting future requests includes predicting that the file will be read sequentially from beginning to end.
10. The method of claim 8 , wherein the software program is antivirus software and the task is to scan for viruses, and predicting future requests includes predicting that the file will be read sequentially from beginning to end.
11. A non-transitory computer-readable storage medium having instructions stored therein, which when executed by a processor, cause the processor to perform a method for providing a storage system, the method comprising:
receiving a request from a client application to read a first portion of a file from a persistent storage device of the storage system, wherein the file references a plurality of deduplicated data segments stored in the storage system that are shared by a plurality of files;
predicting future requests to the file based in part on the received request and metadata of the file, including determining whether any deduplicated segment referenced by the first portion will likely be accessed subsequently in a second portion of the file;
optimizing access of a subsequent portion of the file based on the predicted future requests, including retaining a first deduplicated segment in a cache memory device in response to determining that the first deduplicated segment will be accessed subsequently in the second portion of the file; and
returning data contained in the first portion of the file to the client application, wherein the data is constructed based n deduplicated segments retrieved from the cache memory device and the persistent storage device.
12. The non-transitory computer-readable storage medium of claim 11 , wherein the storage system is a deduplicating storage system.
13. The non-transitory computer-readable storage medium of claim 11 , wherein optimizing the storage system includes managing a data cache based on the predicted future requests.
14. The non-transitory computer-readable storage medium of claim 11 , wherein optimizing access of the subsequent portion of the file includes prefetching data into a data cache based on the predicted future requests.
15. The non-transitory computer-readable storage medium of claim 11 , wherein optimizing access of the subsequent portion of the file includes scheduling an order in which requests are handled based on the predicted future requests.
16. The non-transitory computer-readable storage medium of claim 11 , wherein predicting future requests is based in part on the type of file.
17. The non-transitory computer-readable storage medium of claim 16 , wherein if the file contains a backup image, predicting future requests includes predicting that the file will be read sequentially from beginning to end.
18. The non-transitory computer-readable storage medium of claim 11 , wherein the method further comprises receiving from a software program an indication of a task and predicting future requests based on the received task.
19. The non-transitory computer-readable storage medium of claim 18 , wherein the software program is a backup software and the task is to restore from backup, and predicting future requests includes predicting that the file will be read sequentially from beginning to end.
20. The non-transitory computer-readable storage medium of claim 18 , wherein the software program is antivirus software and the task is to scan for viruses, and predicting future requests includes predicting that the file will be read sequentially from beginning to end.
21. A storage system, comprising:
a processor;
an access manager coupled to the processor to receive a request from a client application to read a first portion of a file from a persistent storage device of the storage system, wherein the file references a plurality of deduplicated data segments stored in the storage system that are shared by a plurality of files; and
an access predictor coupled to the access manager to predict future requests to the file based in part on the received request and metadata of the file, including determining whether any deduplicated segment referenced by the first portion will likely be accessed subsequently in a second portion of the file, wherein the access manager is configured to optimize access of a subsequent portion of the file based on the predicted future requests and to return data contained in the first portion of the file to the client application, wherein optimizing access includes retaining a first deduplicated segment in a cache memory device in response to determining that the first deduplicated segment will be accessed subsequently in the second portion of the file, and wherein the data is constructed based on the deduplicated segments retrieved from the cache memory device and the persistent storage device.
22. The system of claim 21 , wherein the storage system is a deduplicating storage system.
23. The system of claim 21 , wherein optimizing the storage system includes managing a data cache based on the predicted future requests.
24. The system of claim 21 , wherein optimizing access of the subsequent portion of the file includes prefetching data into a data cache based on the predicted future requests.