Maintaining spatial locality of write operations
View Patent ↗A method, system, computer system, and computer-readable medium that identify the number of times areas of storage are updated and use that information to synchronize primary and secondary storage areas. The invention enables quick resumption of operations upon a discovery that primary and secondary data have become unsynchronized. Regions of storage that are unsynchronized can be identified quickly without the need to read the change log. In addition, data are copied efficiently, with minimal or no duplication of effort or data. Faster and more reliable storage can be allocated for storing regions of data that are “hot spots,” thereby making better use of existing storage resources and budgets.
1. A method comprising:
in response to a write operation to a region of a plurality of regions of a storage area, incrementing a respective counter of a plurality of counters;
in response to an acknowledgement that the write operation to the region is received at a replication storage area, decrementing the respective counter for the region;
identifying a set of unreplicated write operations from the plurality of counters, wherein a non-zero counter value of the plurality of counters indicates a number of unreplicated write operations for the region; and
examining a spatial distribution of the set of unreplicated write operations among the respective regions to select a synchronization method.
2. The method of claim 1 , further comprising:
when the spatial distribution of the set of unreplicated write operations is below a concentration limit,
reading a log of write operations to the storage area, and
sending each unreplicated write operation in the log to the replication storage area.
3. The method of claim 1 , further comprising:
when the spatial distribution of the set of unreplicated write operations is greater than or equal to a concentration limit,
for each non-zero counter in the plurality of counters, copying the respective region of the storage area for the non-zero counter to the replication storage area.
4. The method of claim 1 , wherein the selected synchronization method comprises:
synchronizing the storage area and the replication storage area by performing the following:
for each non-zero counter in the plurality of counters, copying the respective region of the storage area for the non-zero counter to the replication storage area.
5. The method of claim 1 , wherein the selected synchronization method comprises:
synchronizing the storage area and the replication storage area by performing the following:
reading a log of write operations to the storage area, and
sending each unreplicated write operation in the log to the replication storage area.
6. The method of claim 1 , further comprising:
using the respective value of at least one counter of the plurality of counters to identify a type of storage to store data for the respective region.
7. The method of claim 6 , further comprising:
migrating the data for the respective region to a storage area that comprises the type of storage.
8. The method of claim 1 wherein each counter of the plurality of counters is capable of tracking a value of at least two.
9. A computer-readable storage medium comprising:
incrementing instructions for incrementing a respective counter of a plurality of counters in response to a write operation to a region of a plurality of regions of a storage area;
decrementing instructions for decrementing the respective counter for the region in response to an acknowledgement that the write operation to the region is received at a replication storage area;
identifying instructions for identifying a set of unreplicated write operations from a plurality of counters; and
examining instructions to examine a spatial distribution of the set of unreplicated write operations among the respective regions to select a synchronization method.
10. The computer-readable storage medium of claim 9 , further comprising:
first selecting instructions to select copying a given region as the synchronization method if a non-zero counter value for the given region exceeds a threshold; and
second selecting instructions to select synchronization instructions comprising the following if the non-zero counter value for the given region does not exceed the threshold:
reading instructions to read a log of write operations to the storage area, and
sending instructions to send each unreplicated write operation in the log to the replication storage area.
11. The computer-readable storage medium of claim 9 , further comprising:
using instructions to use the respective value of at least one counter of the plurality of counters to identify a type of storage to store data for the respective region.
12. The computer-readable storage medium of claim 11 , further comprising:
migrating instructions to migrate the data for the respective region to a storage area that comprises the type of storage.
13. A computer system comprising:
a memory comprising the computer-readable storage medium of claim 9 ; and a processor coupled to the memory to execute said instructions on the computer-readable storage medium.