Data service-aware input/output scheduling

1. A method of request scheduling in a computing environment, comprising:

obtaining a segment size for which one or more data services in the computing environment are configured to process data;

obtaining, from a host device in the computing environment, one or more requests to at least one of read data from and write data to one or more storage devices in the computing environment, wherein the one or more requests originate from one or more application threads of the host device;

aligning the one or more requests into one or more segments having the obtained segment size to generate one or more aligned segments, wherein the one or more requests are respectively provided to one or more local request queues corresponding to the one or more application threads, the one or more local request queues performing request merging, based on the obtained segment size, to generate the one or more aligned segments; and

dispatching the one or more aligned segments to the one or more data services prior to sending the one or more requests to the one or more storage devices;

wherein the computing environment is implemented via one or more processing devices operatively coupled via a communication network.

2. The method of claim 1 , wherein the segment size is specified by the one or more data services.

3. The method of claim 1 , wherein the one or more application threads respectively execute on one or more processing cores of the host device.

4. The method of claim 1 , wherein the aligning step comprises a local stage and a global stage.

5. The method of claim 4 , wherein the local stage further comprises the one or more local request queues corresponding to the one or more application threads.

6. The method of claim 5 , wherein the global stage further comprises a global request queue corresponding to the host device, wherein the global request queue is operatively coupled to the one or more local request queues.

7. The method of claim 6 , wherein the global request queue performs request merging on the requests received from the one or more local request queues based on the obtained segment size for which the one or more data services in the computing environment are configured to process data.

8. The method of claim 6 , wherein each of the one or more local request queues determine if a request is aligned and of sufficient size and, if aligned and of sufficient size, the request is dispatched to the one or more data services without being processed by the global request queue.

9. The method of claim 8 , wherein if the request is aligned but not of sufficient size, the request is provided to the global request queue.

10. The method of claim 8 , wherein if the request is not aligned, the local request queue splits the request into an aligned part and an unaligned part, sends the aligned part to the global request queue, and holds the unaligned part.

11. The method of claim 8 , wherein if the request is not aligned, the local request queue holds the entire request.

12. The method of claim 7 , wherein the global request queue splits a request into an aligned part and an unaligned part, sends the aligned part to the one or more data services, and holds the unaligned part.

13. The method of claim 7 , wherein the global request queue and the one or more local request queues are responsive to a configurable parameter relating to a preferred block segment length.

14. The method of claim 7 , wherein the global request queue and the one or more local request queues are responsive to a configurable parameter relating to a strategy about segment merge and submit operations.

15. The method of claim 7 , wherein the global request queue and the one or more local request queues are responsive to a configurable parameter relating to a merge timeout.

16. The method of claim 7 , wherein the global request queue and the one or more local request queues are responsive to configurable parameter relating to a segment length threshold.

17. The method of claim 1 , wherein the one or more data services comprise one or more of inline data deduplication and inline data compression.

18. A system for request scheduling in a computing environment, the system comprising:

at least one processor, coupled to a memory, and configured to:

obtain a segment size for which one or more data services in the computing environment are configured to process data;

obtain, from a host device in the computing environment, one or more requests to at least one of read data from and write data to one or more storage devices in the computing environment, wherein the one or more requests originate from one or more application threads of the host device;

align the one or more requests into one or more segments having the obtained segment size to generate one or more aligned segments, wherein the one or more requests are respectively provided to one or more local request queues corresponding to the one or more application threads, the one or more local request queues performing request merging, based on the obtained segment size, to generate the one or more aligned segments; and

dispatch the one or more aligned segments to the one or more data services prior to sending the one or more requests to the one or more storage devices.

19. An article of manufacture for request scheduling in a computing environment, the article of manufacture comprising a non-transitory processor-readable storage medium having encoded therein executable code of one or more software programs, wherein the one or more software programs when executed by at least one processing device implement the steps of:

obtaining a segment size for which one or more data services in the computing environment are configured to process data;

obtaining, from a host device in the computing environment, one or more requests to at least one of read data from and write data to one or more storage devices in the computing environment, wherein the one or more requests originate from one or more application threads of the host device;

aligning the one or more requests into one or more segments having the obtained segment size to generate one or more aligned segments, wherein the one or more requests are respectively provided to one or more local request queues corresponding to the one or more application threads, the one or more local request queues performing request merging, based on the obtained segment size, to generate the one or more aligned segments; and

dispatching the one or more aligned segments to the one or more data services prior to sending the one or more requests to the one or more storage devices.

20. The system of claim 18 , wherein:

the aligning step comprises a local stage and a global stage,

the local stage further comprises the one or more local request queues corresponding to the one or more application threads,

the global stage further comprises a global request queue corresponding to the host device, wherein the global request queue is operatively coupled to the one or more local request queues,

the global request queue is configured to perform request merging on the requests received from the one or more local request queues based on the obtained segment size,

the one or more local request queues are configured to determine if a request is aligned and of sufficient size and, if aligned and of sufficient size, to dispatch the request to the one or more data services without being processed by the global request queue,

wherein if the request is aligned but not of sufficient size, the one or more local request queues are configured to provide the request to the global request queue, and

if the request is not aligned, the one or more local request queues are configured to split the request into an aligned part and an unaligned part, to send the aligned part to the global request queue, and to hold the unaligned part.