Dynamic Throttling of Hard Delete Tasks in Data Centers

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

The challenge in data center management is the spike in hard delete requests, which leads to scaling issues and data integrity concerns due to the sudden influx of deletion tasks, potentially deleting incorrect data without proper validation.

Innovation Solution

Implementing dynamic throttling of hard delete tasks to match the rate of soft delete requests, ensuring that data deletion is synchronized with customer requests and reducing the risk of incorrect deletions by smoothing out traffic spikes and incorporating rate limiting mechanisms.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If batch hard delete tasks are executed periodically to reclaim capacity, then resource reclamation efficiency is improved, but traffic spikes and data integrity risks increase

Engineering Contradiction:
Improveresource reclamation efficiencyVSAvoiddata integrity
Core Design Contradiction:
ProductivityVSReliability

Solution Approach 1:

The system dynamically adjusts the hard delete task execution rate to match the soft delete request rate. Instead of fixed periodic batch processing, the hard delete operations are throttled based on real-time monitoring of soft delete incoming rates, creating a dynamic balance that prevents traffic spikes while maintaining efficient resource reclamation

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The system implements feedback mechanisms by monitoring soft delete request rates and using this information to control hard delete task execution. The throttling rate is adjusted based on feedback from the soft delete request stream, ensuring that hard delete operations do not exceed the rate at which soft delete requests are received, thereby preventing data integrity issues

Inventive Principle:
Principle #23Feedback

2Productivity

If hard delete tasks are processed in batches, then capacity reclamation is accelerated, but scaling issues and operational complexity increase

Engineering Contradiction:
Improvecapacity reclamation speedVSAvoidsystem scaling complexity
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The system changes the parameter of delete task processing from fixed batch intervals to rate-based throttling. By adjusting the execution rate parameter to match soft delete request rates in real-time, the system maintains high productivity without the complexity of managing large batch operations and their associated traffic spikes

Inventive Principle:
Principle #35Parameter changes

3Productivity

If soft delete requests are immediately hard deleted, then resource reclamation is maximized, but risk of incorrect data deletion increases

Engineering Contradiction:
Improveresource reclamation rateVSAvoiddeletion accuracy
Core Design Contradiction:
ProductivityVSReliability

Solution Approach 1:

The system performs soft delete as a preliminary action before hard delete. This preliminary soft delete stage allows the system to queue deletion requests and throttle subsequent hard delete operations based on monitored rates, ensuring that deletions are performed at safe, validated rates rather than immediately, thus preventing incorrect data loss while maintaining reclamation productivity

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentUS11106386B1Dynamic throttling of capacity reclamation
Publication Date: 2021.08.31 AMAZON TECH INC
  • US11106386B1 patent drawing
  • US11106386B1 patent drawing
  • US11106386B1 patent drawing

AI summary

Techniques for dynamic throttling of capacity reclamation are described. A method of dynamic throttling of capacity reclamation may include obtaining a plurality of deletion requests, the plurality of deletion requests including client device-originating deletion requests and service-originating deletion requests, generating a plurality of deletion tasks corresponding to the plurality of deletion requests, determining a deletion request rate associated with the client device-originating deletion requests based on request metadata, and executing the plurality of deletion tasks based on the deletion request rate.