Delaying result of I/O operation based on target completion time

A method implemented at a device facilitating interaction between a compute instance and a block storage volume acting as virtual storage for the compute instance, wherein the block storage volume is hosted by one or more storage media in communication with the compute instance over a network, and wherein the device enables the compute instance to submit input/output (I/O) operations to the block storage volume via a local mass storage protocol, the method comprising:

• obtaining a performance profile for the block storage volume, the performance profile indicating an expected performance of the block storage volume as a distribution of I/O operation latencies;
• obtaining an I/O operation from the compute instance;
• submitting the I/O operation to the one or more storage media, wherein the one or more storage media are configured to process the I/O operation and return to the device a result of the I/O operation;
• determining, from the distribution of I/O operation latencies, a target completion time of the I/O operation;
• delaying returning the result of the I/O operation to the compute instance until the target completion time of the I/O operation, wherein the target completion time of the I/O operation is subsequent to a time of obtaining the result of the I/O operation at the device;
• and
• subsequent to the delay, providing the result of the I/O operation to the compute instance.

One or more non-transitory computer-readable media comprising instructions executable by a device facilitating interaction between a compute instance and a block storage volume acting as virtual storage for the compute instance, wherein the block storage volume is hosted by one or more storage media in communication with the compute instance over a network, wherein the device enables the compute instance to submit input/output (I/O) operations to the block storage volume via a local mass storage protocol, and wherein the instructions, when executed, cause the device to:

• obtain a performance profile for the block storage volume, the performance profile indicating an expected performance of the block storage volume as a distribution of I/O operation latencies;
• obtain an I/O operation from the compute instance;
• submit the I/O operation to the one or more storage media, wherein the one or more storage media are configured to process the I/O operation and return to the device a result of the I/O operation;
• determine, from the distribution of I/O operation latencies, a target completion time of the I/O operation;
• if a current time is prior to the target completion time, delay completion of the I/O operation until the target completion time of the I/O operation;
• and
• on or after the target completion time, provide the result of the I/O operation to the compute instance.