Affinity-based cache operation for a persistent storage device
Aspects described herein relate to a method comprising: receiving a request to write data to a persistent storage device, the request comprising data; determining an affinity of the data; writing the request to a cache line of a cache; associating the cache line with the affinity of the data; and reporting the data as having been written to the persistent storage device.
1. A method comprising:
receiving a request to write data to a persistent storage device, the request comprising data and a logical block addressing (LBA) value;
determining an affinity of the data, wherein the affinity of the data is a least a function of the LBA value;
writing the request to a cache line of a cache;
associating the cache line with the affinity of the data; and
reporting the data as having been written to the persistent storage device.
2. The method of claim 1 , wherein the request further comprises a request time, and the affinity of the data is at least a function the request time.
3. The method of claim 1 , wherein the request further comprises a process identifier, and the affinity of the data is at least a function the process identifier.
4. The method of claim 1 , wherein the request further comprises a central processing unit (CPU) identifier, and the affinity of the data is at least a function the CPU identifier.
5. The method of claim 1 , wherein the request further comprises a central processing unit (CPU) core identifier, and the affinity of the data is at least a function of the CPU core identifier.
6. The method of claim 1 , wherein the cache comprises a plurality of buckets, each bucket comprising a predetermined number of cache lines.
7. The method of claim 6 , wherein each bucket is associated with a range of LBA values.
8. The method of claim 1 , further comprising:
determining that data is to be evicted from the cache;
generating a stream based on one or more buckets; and
writing the stream to the persistent storage device.
9. The method of claim 8 , wherein generating a stream based on one or more buckets comprises:
determining one or more buckets based on least recent use.
10. The method of claim 8 , wherein determining that data is to be evicted from the cache comprises:
determining that the cache is below a threshold clean capacity.
11. The method of claim 8 , wherein determining that data is to be evicted from the cache comprises determining that a predetermined proportion of the buckets are full above a threshold.
12. The method of claim 1 , further comprising:
after writing the data to a cache line of a cache, marking the cache line as dirty.
13. A system comprising:
one or more processors; and
a memory;
wherein the memory stores instructions that, when executed by the one or more processors, causes the system to:
receive a request to write data to a persistent storage device, the request comprising data and a logical block addressing (LBA) value;
determine an affinity of the data, wherein the affinity of the data is a least a function of the LBA value;
write the request to a cache line of a cache;
associate the cache line with the affinity of the data; and
report the data as having been written to the persistent storage device.
14. A system comprising:
a memory; and
a plurality of non-uniform memory access (NUMA) nodes, each NUMA node comprising at least one CPU core in communication with at least a portion of the memory,
wherein the memory stores instructions that, when executed by the plurality of NUMA nodes, cause each of the NUMA nodes to:
receive a request to write data to a persistent storage device, the request comprising data and a logical block addressing (LBA) value;
determine an affinity of the data, wherein the affinity of the data is a least a function of the LBA value;
write the request to a cache line of a cache;
associate the cache line with the affinity of the data; and
report the data as having been written to the persistent storage device.
15. A non-transitory computer readable medium comprising instructions which, when executed by one or more processors, cause the one or more processors to perform steps comprising:
receiving a request to write data to a persistent storage device, the request comprising data and a logical block addressing (LBA) value;
determining an affinity of the data, wherein the affinity of the data is a least a function of the LBA value;
writing the request to a cache line of a cache;
associating the cache line with the affinity of the data; and
reporting the data as having been written to the persistent storage device.