Ensemble Event Scheduling for High Precision Timing

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing event scheduling systems in cloud-based service provider networks face limitations in precision due to jitter, which restricts the ability to schedule events with high granularity, particularly for users requiring precise timing in managing distributed resources.

Innovation Solution

The implementation of an ensemble of event schedulers that coordinate to generate ticks within a specified jitter delay period, increasing the likelihood of a tick occurring at the desired time by calculating the optimal number of schedulers to include in the ensemble based on user-defined confidence levels and threshold periods.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Device complexity

If a single event scheduler is used to schedule events at scale, then processing power consumption is reduced and system complexity is lowered, but scheduling precision deteriorates due to jitter in schedule ticks

Engineering Contradiction:
Improvescheduling system complexityVSAvoidevent scheduling precision
Core Design Contradiction:
Device complexityVSMeasurement precision

Solution Approach 1:

The patent divides the scheduling system into multiple independent schedulers (e.g., first scheduler, second scheduler, third scheduler) that operate in parallel. Each scheduler generates ticks independently, and the system selects the earliest tick that falls within the desired time window. This segmentation allows the system to achieve high precision without requiring a single complex scheduler, as each individual scheduler remains simple while their collective output provides precision through statistical aggregation.

Inventive Principle:
Principle #1Segmentation

2Measurement precision

If multiple schedulers are used to improve scheduling precision, then event timing accuracy improves, but processing power consumption increases

Engineering Contradiction:
Improveevent scheduling precisionVSAvoidprocessing power consumption
Core Design Contradiction:
Measurement precisionVSUse of energy by moving object

Solution Approach 1:

The patent combines multiple scheduler outputs into a unified selection mechanism where the system evaluates ticks from multiple schedulers and selects the earliest one within the desired time window. This merging approach allows the system to achieve high precision through statistical aggregation of multiple simple schedulers, rather than requiring a single complex high-precision scheduler, thereby maintaining lower processing power consumption while improving precision.

Inventive Principle:
Principle #5Merging (Combining)

3Measurement precision

If high precision event scheduling is implemented, then user requirements for granular timing are met, but system scalability deteriorates

Engineering Contradiction:
Improveevent scheduling precisionVSAvoidsystem scalability
Core Design Contradiction:
Measurement precisionVSAdaptability or versatility

Solution Approach 1:

The patent introduces a configurable parameter (desired time window or threshold) that allows the system to adapt to different precision requirements. By adjusting this parameter, the system can provide high precision when needed while maintaining scalability for less critical events. The selection mechanism remains consistent, but the threshold for accepting ticks can be dynamically adjusted based on service requirements, enabling the system to scale effectively across different use cases.

Inventive Principle:
Principle #35Parameter changes

Data Source

PatentUS10942773B1Ensemble event scheduling by executing a plurality of schedulers having a probabilistic distribution of ticks to ensure a tick within a threshold time
Publication Date: 2021.03.09 AMAZON TECH INC
  • US10942773B1 patent drawing
  • US10942773B1 patent drawing
  • US10942773B1 patent drawing

AI summary

This disclosure describes techniques implemented by an event scheduling system for using an ensemble of event schedulers to achieve higher precision when generating a tick at a desired time at which an event is scheduled to occur. Event scheduling systems that schedule events at scale may utilize event schedulers that have relatively low precision in order to provide the scheduling services at scale. By utilizing an ensemble, or group, of event schedulers to generate a tick at a particular time for a scheduled event, the ticks generated by the ensemble of event schedulers may be distributed throughout jitter delay such that likelihood that one of the ticks is close to the scheduled time for the event is increased. Thus, the techniques described herein may provide higher precision scheduling techniques for use by event scheduling systems.