Distributed Data Pre-moving and Adaptive Compute Allocation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In geographically distributed computing systems, latency is increased due to the inefficiency of data exchange across network bottlenecks, where data needs to be copied to a centralized location for processing, which can be costly in terms of network communication and time.
Innovation Solution
Pre-moving input data and adaptively allocating intermediate steps of a multi-step compute operation among geographically distributed computing devices to reduce data exchange duration, prioritizing high-value data sets and setting a threshold for network communication usage to avoid unnecessary latency without gain.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If data is copied to a centralized location for processing, then compute operations can be performed, but network communication cost and time increase
Solution Approach 1:
The patent pre-moves input data to computing devices before queries are received, based on predictions of which data will be needed. This preliminary data placement eliminates the need for time-consuming data copying when queries arrive, directly reducing data exchange duration while maintaining compute operation capability
Solution Approach 2:
The patent divides compute operations into multiple intermediate steps and allocates different steps to different geographically distributed computing devices. This segmentation allows parallel execution of steps across devices, reducing overall processing time and network communication overhead compared to centralized processing
2Loss of time
If data is pre-moved among geographically distributed computing devices, then data exchange duration is reduced, but network communication usage increases
Solution Approach 1:
The patent selectively pre-moves only the portions of data that will be needed for upcoming queries, rather than moving all data. This partial action approach reduces unnecessary network communication while still achieving the latency reduction benefit for the specific data that matters
Solution Approach 2:
The system dynamically adjusts data placement decisions based on changing parameters such as predicted query patterns, current data locations, and network conditions. This allows optimization of the balance between pre-moving data (reducing latency) and network communication usage under different operating conditions
3Loss of time
If compute operations are allocated among geographically distributed computing devices, then data exchange duration is reduced, but system complexity increases
Solution Approach 1:
The patent employs feedback mechanisms where the system monitors query patterns, data access trends, and performance metrics to continuously optimize compute operation allocation. This feedback-driven approach automates the complexity of managing distributed task allocation, making it adaptive rather than statically complex
Data Source
AI summary
Latency in responding to queries directed to geographically distributed data can be reduced by allocating individual steps, of a multi-step compute operation requested by the query, among the geographically distributed computing devices so as to reduce the duration of shuffling of intermediate data among such devices, and, additionally, by pre-moving, prior to the receipt of the query, portions of the distributed data that are input to a first step of the multistep compute operation, to, again, reduce the duration of the exchange of intermediate data. The pre-moving of input data occurring, and the adaptive allocation of intermediate steps, are prioritized for high-value data sets. Additionally, a threshold increase in a quantity of data exchanged across network communications can be established to avoid incurring network communication usage without an attendant gain in latency reduction.