Asynchronous Bot Detection in Multi-Instance Cloud Systems

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

In cloud computing environments, distinguishing between human and bot activities becomes increasingly difficult as data volumes grow, leading to inefficiencies in resource utilization and billing, as current methods fail to efficiently detect and classify bots within multi-instance systems.

Innovation Solution

Implementing an asynchronous non-blocking transaction bot detection mechanism that executes only once per session, utilizing session attributes and an in-memory cached list of bots to enhance detection speed and efficiency, allowing for differentiated handling of human and bot interactions.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If traditional bot detection methods are used in cloud computing environments, then bot detection capability is provided, but resource utilization efficiency deteriorates and billing accuracy worsens due to inability to efficiently distinguish human and bot activities

Engineering Contradiction:

Improvebot detection accuracyVSAvoidresource utilization efficiency

Core Design Contradiction:

Measurement precisionVSProductivity

Solution Approach 1:

The bot detection system is segmented into multiple independent bot detection instances, each handling specific sessions. This allows parallel processing of bot detection across different instances without blocking other transactions, improving both detection accuracy and resource utilization efficiency simultaneously

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The system performs preliminary bot detection by setting session attributes during session creation and executing bot detection logic before main transaction processing. This preliminary action enables early identification of bots, allowing optimized resource allocation and accurate billing decisions to be made upfront

Inventive Principle:

Principle #10Preliminary action

2Reliability

If bot detection is executed for every transaction, then comprehensive bot detection is achieved, but transaction processing speed deteriorates due to repeated detection overhead

Engineering Contradiction:

Improvebot detection completenessVSAvoidtransaction processing speed

Core Design Contradiction:

ReliabilityVSSpeed

Solution Approach 1:

Bot detection is implemented as periodic action executed only once per session at the beginning, rather than for every transaction. The system sets a session attribute indicating bot detection has been performed, and subsequent transactions within the same session skip detection, maintaining reliability while improving processing speed

Inventive Principle:

Principle #19Periodic action

Solution Approach 2:

Inventive Principle:

Principle #10Preliminary action

3Measurement precision

If synchronous bot detection is used, then accurate bot classification is achieved, but system responsiveness deteriorates due to blocking transactions during detection

Engineering Contradiction:

Improvebot classification accuracyVSAvoidtransaction delay

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

An intermediary session attribute mechanism is introduced to store bot detection results. The asynchronous bot detection process writes results to this intermediary storage, allowing the main transaction flow to continue without waiting for detection completion, thus maintaining both accuracy and responsiveness

Inventive Principle:

Principle #24Intermediary (Mediator)

Solution Approach 2:

The patent replaces the mechanical synchronous blocking detection mechanism with an asynchronous non-blocking detection system. Bot detection runs in the background while transactions proceed independently, with results communicated through session attributes rather than direct blocking calls

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

4Reliability

If multiple bot detection instances process the same session, then detection coverage is improved, but system efficiency deteriorates due to redundant processing

Engineering Contradiction:

Improvedetection coverageVSAvoidprocessing overhead

Core Design Contradiction:

ReliabilityVSLoss of energy

Solution Approach 1:

Multiple bot detection instances are merged through a shared session attribute space. All instances read the same session attributes and bot detection results, eliminating redundant processing while maintaining comprehensive detection coverage across the distributed system

Inventive Principle:

Principle #5Merging (Combining)

Data Source

PatentUS11258860B2System and method for bot detection and classification

Publication Date: 2022.02.22 SERVICENOW INC

AI summary

A computing system includes a server. The server is communicatively coupled to a data repository and is configured to store a data in the data repository. The server is further configured to create a server instance, wherein the server instance is associated with a user. The server is additionally configured to create a session based on an external entity requesting a resource from the server instance, and to execute a bot detection logic to determine if the external entity is a bot. If the external entity is a bot then the server is configured to perform a bot-based action, wherein the server is configured to provide for multi-instance support to a plurality of users.