Self-Learning Event Broker for Adaptive Pub-Sub Message Delivery

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Conventional Pub-Sub messaging systems face inefficiencies in network traffic and latency due to the asynchronous nature of message delivery, especially in large-scale distributed computing environments with varying subscriber capabilities.

Innovation Solution

A self-learning cloud broker that predicts optimal message parameters, such as batch count, payload size, and time gap, using Thompson sampling and a multi-armed bandit model to ensure successful delivery to individual subscribers, adapting to changes in network capacity and subscriber capabilities.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If conventional Pub-Sub messaging paradigm is used for asynchronous communication, then information can be transmitted from publishing clients to subscribing clients based on triggering conditions, but network traffic inefficiencies occur including reduced latencies and traffic bottlenecks

Engineering Contradiction:

Improveinformation transmission efficiencyVSAvoidlatency

Core Design Contradiction:

ProductivityVSLoss of time

Solution Approach 1:

The system dynamically adjusts message delivery parameters (batch count, payload size, time gap) based on real-time network conditions and subscriber capabilities. The broker learns optimal parameters through continuous feedback from delivery outcomes, adapting the messaging strategy to changing system conditions rather than using fixed asynchronous delivery patterns.

Inventive Principle:

Principle #15Dynamics

Solution Approach 2:

The invention changes key message delivery parameters including batch count, payload size, and time gap between messages. These parameters are optimized based on learned subscriber characteristics and network conditions, transforming the traditional fixed-parameter Pub-Sub model into a flexible, adaptive system that resolves latency and efficiency trade-offs.

Inventive Principle:

Principle #35Parameter changes

2Adaptability or versatility

If the number of subscribing clients and publishing clients increases, then the system can scale to handle more clients, but determining optimal message parameters becomes more complex and resource-intensive

Engineering Contradiction:

Improvesystem scalabilityVSAvoidparameter determination complexity

Core Design Contradiction:

Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The broker performs self-learning by automatically observing message delivery outcomes and updating its understanding of subscriber capabilities without external intervention. This self-service mechanism allows the system to scale to more clients without proportionally increasing configuration complexity, as each new subscriber is automatically characterized through learned interactions.

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

The system implements feedback loops where message delivery results are observed and used to update the distribution models for each subscriber. This continuous feedback enables automatic adaptation to new clients and changing conditions, reducing the manual complexity of managing large numbers of publishers and subscribers while maintaining optimal performance.

Inventive Principle:

Principle #23Feedback

3Reliability

If message delivery parameters are optimized for each individual subscriber, then successful message delivery is improved, but the computational overhead for determining and updating parameters increases

Engineering Contradiction:

Improvemessage delivery success rateVSAvoidcomputational resource consumption

Core Design Contradiction:

ReliabilityVSUse of energy by moving object

Solution Approach 1:

The system uses sampling-based parameter determination rather than exhaustively optimizing all possible parameters for each subscriber. By sampling from learned distributions to determine message parameters, the system achieves sufficient delivery success rates without the prohibitive computational cost of complete optimization, balancing reliability with resource efficiency.

Inventive Principle:

Principle #16Partial or excessive action

Solution Approach 2:

The broker pre-learns subscriber characteristics and message delivery patterns during initial interactions, building distribution models that predict optimal parameters. This preliminary learning reduces the real-time computational burden when actually sending messages, as the heavy lifting of parameter optimization is performed incrementally during the learning phase rather than at message send time.

Inventive Principle:

Principle #10Preliminary action

Data Source

PatentUS20250350573A1Techniques for a self-learning scalable event broker

Publication Date: 2025.11.13 ORACLE INT CORP

AI summary

Techniques are disclosed for implementing a self-learning cloud-based message broker are disclosed. The message broker can receive an event trigger that includes information usable to identify a subscribing client of a publisher-subscriber messaging system. The message broker can determine message parameters for one or more messages by sampling a distribution. The message broker can determine the message parameters in response to receiving the event trigger. The message broker can send the one or more messages to the subscribing client. The one or more messages can be characterized by the message parameters. The message broker can receive a response status from the subscribing client and, based on the response status, update the distribution.