Query Management System Interval Slicing for Latency Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Distributed query management systems often become backlogged due to the sequential execution of queries, leading to increased latency for shorter queries and inefficient resource utilization, as longer queries occupy processing resources for extended periods.
Innovation Solution
The system divides queries into intervals for partial processing and suspension, allowing high-priority queries to be processed sooner and reducing backlog by selecting queries based on priority and using query slicing, where queries are executed in alternating intervals and suspended to enable immediate processing of new queries.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If queries are executed sequentially, then processing resources are used efficiently for each query, but the system becomes backlogged and latency increases for shorter queries
Solution Approach 1:
The patent segments query execution into discrete intervals, allowing the system to process different queries in different intervals rather than executing each query sequentially to completion. This segmentation enables interleaved execution where short queries can be processed in intervals between longer queries, reducing overall latency while maintaining resource utilization.
Solution Approach 2:
The system implements periodic action by executing queries in time-based intervals rather than continuous sequential execution. The query processing is divided into periodic time slots, allowing the system to switch between queries periodically, which prevents any single long query from monopolizing resources and reduces the backlog effect.
2Reliability
If long queries occupy processing resources for extended periods, then complete processing is achieved, but resource utilization efficiency decreases and backlog increases
Solution Approach 1:
The patent applies dynamics by making query execution flexible and adaptive rather than static and fixed. Query execution duration and resource allocation are dynamically adjusted based on query priority, current system state, and interval timing. This allows the system to interrupt long queries when necessary and reallocate resources to higher-priority or shorter queries, improving overall productivity while ensuring all queries eventually complete.
Data Source
AI summary
Managing queries performed on one or more data sources includes: storing at least a first query in a storage medium; selecting the first query for processing; instructing a query engine to process the first query on a first portion of data in the one or more data sources for a first query interval; receiving result data from the query engine based on processing the first query on the first portion of data; saving a state of the first query in the storage medium after the first query interval; instructing the query engine to process a second query during a second query interval after the first query interval; and instructing the query engine to process the first query on a second portion of data in the one or more data sources during a third query interval after the second query interval.