This enhancement introduces the capability to execute read operations in parallel within a single DataShard, thereby improving both latency and throughput. The aim is to optimize the performance of data retrieval processes by allowing simultaneous read execution.
Considerations
As we implement this improvement, several important considerations must be addressed:
Deal with splits by load: establish criteria and thresholds for determining when a DataShard should be split based on its load.
Decide when to use parallel reads: develop a condition when to initiate new parallel reads instead of the sequential ones.
Measure the system performance enhancement: develop methods to evaluate and demonstrate the overall improvement in system behavior. This could involve setting up benchmarks, conducting performance tests, and collecting metrics to provide quantitative evidence of enhanced latency and throughput.
Description
This enhancement introduces the capability to execute read operations in parallel within a single DataShard, thereby improving both latency and throughput. The aim is to optimize the performance of data retrieval processes by allowing simultaneous read execution.
Considerations
As we implement this improvement, several important considerations must be addressed:
Deal with splits by load: establish criteria and thresholds for determining when a DataShard should be split based on its load.
Decide when to use parallel reads: develop a condition when to initiate new parallel reads instead of the sequential ones.
Measure the system performance enhancement: develop methods to evaluate and demonstrate the overall improvement in system behavior. This could involve setting up benchmarks, conducting performance tests, and collecting metrics to provide quantitative evidence of enhanced latency and throughput.
Steps
tbd