System and method for efficient data access for restores

1. A storage gateway for providing stored data, comprising:

a persistent storage comprising a stored data map and a data fragment library; and

a processor programmed to:

obtain a data storage access request from a client;

identify a data fragment specified by the data storage access request;

identify a plurality of storages that each store a copy of the data fragment using the stored data map;

identify, using the data fragment library, a workload type associated with the data fragment, wherein the workload type corresponds to a type of file associated with the data fragment;

obtain a plurality of connection characteristics between the plurality of storages and the client;

identify, based on the workload type, a first connection characteristic of the plurality of connection characteristics;

select a first storage of the plurality of storages, wherein the first connection characteristic is between the client and the first storage; and

obtain the copy of the data fragment from the first storage.

2. The storage gateway of claim 1 , wherein the plurality of connection characteristics specify a bandwidth of respective connections between each of the plurality of storages and the client.

3. The storage gateway of claim 1 , wherein the plurality of connection characteristics specify a latency of respective connections between each of the plurality of storages and the client.

4. The storage gateway of claim 1 , wherein the processor is further programmed to:

before obtaining the data storage access request:

obtain a data storage request for the data fragment; and

store, in response to the data storage request, a first copy of the data fragment in the first storage of the plurality of storages.

5. The storage gateway of claim 4 , wherein the processor is further programmed to:

store a second copy of the data fragment in a second storage of the plurality of storages.

6. The storage gateway of claim 5 , wherein the plurality of connection characteristics specify that the first storage has lower latency to the client than the second storage.

7. The storage gateway of claim 5 , wherein the plurality of connection characteristics specify that the first storage has greater bandwidth to the client than the second storage.

8. The storage gateway of claim 1 , wherein the processor is further programmed to:

delete the data fragment from the first storage of the plurality of storages; and

after deleting the data fragment from the first storage:

obtain a second data storage access request from a second client;

identify a second data fragment specified by the second data storage access request;

identify a second plurality of storages that each store a copy of the second data fragment using the stored data map;

select a second storage of the second plurality of storages based on a second plurality of connection characteristics between the second plurality of storages and the second client when the second data storage access request is obtained; and

obtain the copy of the second data fragment from the second storage.

9. The storage gateway of claim 1 , wherein each storage of the plurality of storages is a separate cloud storage.

10. The storage gateway of claim 1 , wherein the plurality of storages are connected to the storage gateway via operable connections.

11. The storage gateway of claim 10 , wherein the operable connections comprise the Internet.

12. The storage gateway of claim 1 , wherein the plurality of storages are geographically separated from the storage gateway.

13. The storage gateway of claim 1 , wherein the first connection characteristic is identified based on a comparison of each connection characteristic of the plurality of connection characteristics.

14. The storage gateway of claim 13 , wherein the first connection characteristic is best optimized for the workload type among the plurality of connection characteristics.

15. A method of operating a storage gateway for providing stored data, comprising:

obtaining a data storage access request from a client;

identifying a data fragment specified by the data storage access request;

identifying a plurality of storages that each store a copy of the data fragment using a stored data map;

identifying, in a data fragment library, a workload type associated with the data fragment, wherein the workload type corresponds to a type of file associated with the data fragment;

obtaining a plurality of connection characteristics between the plurality of storages and the client;

identifying, based on the workload type, a first connection characteristic of the plurality of connection characteristics;

selecting a first storage of the plurality of storages, wherein the first connection characteristic is between the client and the first storage; and

obtaining the copy of the data fragment from the first storage.

16. The method of claim 15 , further comprising:

delete the data fragment from the first storage of the plurality of storages; and

after deleting the data fragment from the first storage:

obtaining a second data storage access request from a second client;

identifying a second data fragment specified by the second data storage access request;

identifying a second plurality of storages that each store a copy of the second data fragment using the stored data map;

selecting a second storage of the second plurality of storages based on a second plurality of connection characteristics between the second plurality of storages and the second client when the second data storage access request is obtained; and

obtaining the copy of the second data fragment from the second storage.

17. A non-transitory computer readable medium comprising computer readable program code, which when executed by a computer processor enables the computer processor to perform a method for operating a storage gateway for providing stored data, the method comprising:

obtaining a data storage access request from a client;

identifying a data fragment specified by the data storage access request;

identifying a plurality of storages that each store a copy of the data fragment using a stored data map;

identifying, in a data fragment library, a workload type associated with the data fragment, wherein the workload type corresponds to a type of file associated with the data fragment;

obtaining a plurality of connection characteristics between the plurality of storages and the client;

identifying, based on the workload type, a first connection characteristic of the plurality of connection characteristics;

selecting a first storage of the plurality of storages, wherein the first connection characteristic is between the client and the first storage; and

obtaining the copy of the data fragment from the first storage.

18. The non-transitory computer readable medium of claim 17 , wherein the method further comprises:

delete the data fragment from the first storage of the plurality of storages; and

after deleting the data fragment from the first storage:

obtaining a second data storage access request from a second client;

identifying a second data fragment specified by the second data storage access request;

identifying a second plurality of storages that each store a copy of the second data fragment using the stored data map;

selecting a second storage of the second plurality of storages based on a second plurality of connection characteristics between the second plurality of storages and the second client when the second data storage access request is obtained; and

obtaining the copy of the second data fragment from the second storage.