Service Request Receiver for JVM Garbage Collection Notification

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Garbage collection operations in Java Virtual Machines (JVMs) and similar run-time environments disrupt service performance by pausing threads, leading to increased response times and reduced throughput, which is problematic for services with stringent performance and scalability requirements.

Innovation Solution

Implementing a redundant set of virtual machines that provide advance notifications of impending garbage collection phases, allowing service requests to be redirected to a spare machine, thereby avoiding performance disruptions.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Device complexity

If garbage collection operations are performed in a single virtual machine, then memory management is simplified, but service performance deteriorates due to thread pauses and increased response times

Engineering Contradiction:
Improvememory management complexityVSAvoidservice throughput
Core Design Contradiction:
Device complexityVSProductivity

Solution Approach 1:

The patent divides the service processing workload across multiple virtual machines (e.g., first virtual machine and second virtual machine), each capable of independent garbage collection. This segmentation allows one VM to perform GC while others continue processing requests, resolving the contradiction between simplified memory management and maintained service throughput.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent implements advance notification mechanisms that alert the service request receiver before a garbage collection phase begins. This preliminary action enables the system to proactively redirect service requests to alternative virtual machines, preventing performance degradation before it occurs and maintaining high throughput during GC operations.

Inventive Principle:
Principle #10Preliminary action

2Quantity of substance

If garbage collection operations pause threads to free memory, then memory availability is improved, but response time increases and throughput decreases

Engineering Contradiction:
Improveavailable memoryVSAvoidresponse time
Core Design Contradiction:
Quantity of substanceVSLoss of time

Solution Approach 1:

The patent introduces a service request receiver as an intermediary component that sits between clients and virtual machines. This receiver uses advance notifications from virtual machines about impending garbage collection to intelligently route requests, acting as a buffer that decouples request processing from GC interruptions and maintains low response times while memory is being freed.

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

By receiving advance notifications before garbage collection starts, the system can preemptively redirect active requests to other virtual machines. This preliminary action ensures that no requests are lost or delayed during the GC pause, maintaining response time performance while still achieving the memory freeing objective.

Inventive Principle:
Principle #10Preliminary action

3Productivity

If multiple virtual machines are used to avoid garbage collection disruptions, then service throughput is maintained, but system complexity increases

Engineering Contradiction:
Improveservice throughputVSAvoidvirtual machine configuration
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The patent implements self-service mechanisms where virtual machines automatically generate and send advance notifications about their own garbage collection schedules to the service request receiver. This eliminates the need for complex external monitoring or manual coordination, allowing the system to maintain high throughput while keeping management complexity low through automated self-communication.

Inventive Principle:
Principle #25Self-service

Solution Approach 2:

The system establishes a feedback loop where virtual machines provide advance notifications about their GC status to the service request receiver, which then adjusts request routing accordingly. This feedback mechanism enables dynamic load balancing without requiring complex predictive algorithms or manual intervention, maintaining throughput while managing complexity through simple, automated responses to GC notifications.

Inventive Principle:
Principle #23Feedback

Data Source

PatentUS11748260B1Service performance enhancement using advance notifications of reduced-capacity phases of operations
Publication Date: 2023.09.05 AMAZON TECH INC
  • US11748260B1 patent drawing
  • US11748260B1 patent drawing
  • US11748260B1 patent drawing

AI summary

A first run-time environment executing a first instance of an application, and a second run-time environment executing a second instance of the application are established. An indication of an impending commencement of a reduced-capacity phase of operation of the first run-time environment is obtained at a service request receiver. Based at least in part on the indication, the service request receiver directs a service request to the second instance of the application.