I/O performance enhancement of solid-state data storage devices
An infrastructure, method and controller card for managing flash memory in a storage infrastructure. A system is provided that includes flash memory; and a controller that includes: an I/O request handler for handling standard read and write (R/W) operations requested from a host; a garbage collection (GC) system that performs a GC process on the flash memory in response to a threshold condition, wherein the GC process includes GC-induced R/W operations; and a scheduler that interleaves standard R/W operations with GC-induced R/W operations, wherein the scheduler calculates minimum and maximum boundaries for GC-induced R/W operations for a GC process based on an estimated GC latency.
1. A storage infrastructure, comprising:
flash memory; and
a controller that includes:
an I/O request handler for handling standard read and write (R/W) operations requested from a host;
a garbage collection (GC) system that performs a GC process on the flash memory in response to a threshold condition, wherein the GC process includes GC-induced R/W operations and has an estimated GC latency divided into a plurality of time slots; and
a scheduler that interleaves standard R/W operations with GC-induced R/W operations, wherein the scheduler calculates minimum and maximum numbers of GC-induced R/W operations for each of the plurality of time slots.
2. The storage infrastructure of claim 1 , wherein the minimum and maximum numbers of GC-induced R/W operations are further determined based on a total number of free pages in the flash memory and a total number of GC-induced R/W operations required for the GC process.
3. The storage infrastructure of claim 2 , wherein the estimated GC latency is divided into n time slots and during the GC process, minimum and maximum boundaries are calculated each time a time slot elapses until the GC process finishes.
4. The storage infrastructure of claim 3 , wherein:
the minimum boundary is defined as k·v/n; and
the maximum boundary is defined as k·P f /n,
where k is a current time slot, v is the total number of GC-induced R/W operations, and P f is the total number of free memory pages.
5. The storage infrastructure of claim 1 , wherein the estimated GC latency is a fixed value.
6. The storage infrastructure of claim 1 , wherein the estimated GC latency is based on a previous GC latency.
7. The storage infrastructure of claim 1 , wherein the estimated GC latency is based on statistically analyzing standard R/W operations during prior time windows.
8. A controller card for managing flash memory in a storage infrastructure, comprising:
an I/O request handler for handling standard read and write (R/W) operations requested from a host;
a garbage collection (GC) system that performs a GC process on the flash memory in response to a threshold condition, wherein the GC process includes GC-induced R/W operations and has an estimated GC latency divided into a plurality of time slots; and
a scheduler that interleaves standard R/W operations with GC-induced R/W operations, wherein the scheduler calculates minimum and maximum numbers of GC-induced R/W operations for each of the plurality of time slots.
9. The controller card of claim 8 , wherein the minimum and maximum numbers of GC-induced R/W operations are further determined based on a total number of free pages in the flash memory and a total number of GC-induced R/W operations required for the GC process.
10. The controller card of claim 9 , wherein the estimated GC latency is divided into n time slots and during the GC process, minimum and maximum boundaries are calculated each time a time slot elapses until the GC process finishes.
11. The controller card of claim 10 , wherein:
the minimum boundary is defined as k*v/n; and
the maximum boundary is defined as k*P f /n,
where k is a current time slot, v is the total number of GC-induced R/W operations, and P f is the total number of free memory pages.
12. The controller card of claim 8 , wherein the estimated GC latency is a fixed value.
13. The controller card of claim 8 , wherein the estimated GC latency is based on a previous GC latency.
14. The controller card of claim 8 , wherein the estimated GC latency is based on statistically analyzing standard R/W operations during prior time windows.
15. A method of managing flash memory in a storage infrastructure, comprising:
receiving standard read and write (R/W) operations requested from a host;
performing a garbage (GC) process on the flash memory in response to a threshold condition, wherein the GC process includes GC-induced R/W operations and has an estimated GC latency divided into a plurality of time slots; and
interleaving standard R/W operations with GC-induced R/W operations, wherein the interleaving includes calculating minimum and maximum numbers of GC-induced R/W operations for each of the plurality of time slots.
16. The method of claim 15 , wherein the minimum and maximum numbers of GC-induced R/W operations are further determined based on a total number of free pages in the flash memory and a total number of GC-induced R/W operations for the GC process.
17. The method of claim 16 , wherein the estimated GC latency is divided into n time slots and during the GC process, minimum and maximum boundaries are calculated each time a time slot elapses until the GC process finishes.
18. The method of claim 17 , wherein:
the minimum boundary is defined as k*v/n; and
the maximum boundary is defined as k*Pf/n,
where k is a current time slot, v is the total number of GC-induced R/W operations, and Pf is the total number of free memory pages.
19. The method of claim 15 , wherein the estimated GC latency is one of a fixed value or one based on a previous GC latency.
20. The method of claim 15 , wherein the estimated GC latency is based on statistically analyzing standard R/W operations during prior time windows.