Recommendation Engine for Batch Job Submission Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional systems fail to prevent batch job failures due to system overload, resource contention, and deadlocks, as they cannot proactively manage user submissions when the system is in an undesirable state, leading to further deterioration.
Innovation Solution
Implementing a recommendation engine that processes application server properties to advise users on whether to submit or schedule batch job requests, providing notifications to refrain from submitting new jobs until the system recovers, thereby preventing further load and allowing resources to focus on existing jobs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If a system limit is imposed on the number of batch jobs, then the quantity of jobs is controlled, but the system cannot prevent users from submitting new jobs when overload or resource contention occurs, worsening system state
Solution Approach 1:
The system continuously monitors system state (overload conditions, resource contention, deadlocks) and provides real-time feedback to users through notifications. When the system detects an undesirable state, it actively communicates this to users, enabling them to adjust their job submission behavior accordingly, thus preventing further system deterioration.
Solution Approach 2:
The system evaluates system state before allowing job submission and proactively notifies users of potential problems before they occur. By assessing current system conditions and warning users in advance, the system enables preventive action rather than reacting after failures occur.
2Productivity
If users continue to submit new jobs unaware of system state, then job submission volume increases, but system performance deteriorates and requires administrator intervention
Solution Approach 1:
The system provides continuous feedback to users about current system conditions through notifications. When the system is in an undesirable state (overload, resource contention, deadlocks), users receive alerts that inform their decision-making, enabling them to adjust submission behavior and avoid exacerbating system problems.
Solution Approach 2:
The system empowers users to make informed decisions about job submission by providing them with system state information. Users can independently assess whether to submit jobs based on notifications they receive, eliminating the need for administrator intervention to manage submission volumes during problematic periods.
3Productivity
If the system accepts all job requests without restriction, then user productivity is maintained, but system overload and resource contention occur leading to job failures
Solution Approach 1:
The system maintains high user productivity by accepting job requests while simultaneously providing feedback notifications about system state. This feedback mechanism allows users to understand when system conditions are problematic, enabling them to voluntarily adjust their behavior without explicit restrictions, thus maintaining productivity while improving job success rates.
Solution Approach 2:
The system takes preliminary anti-action by warning users of potential system problems before they submit jobs that would fail. By providing advance notice of overload conditions, resource contention, or deadlocks, the system prevents users from submitting jobs that would inevitably fail, thereby improving overall job success rates while preserving user autonomy.
Data Source
AI summary
Described herein are systems and methods for receiving a recommendation before submitting a work request. As described herein, an indication of a work request, a recommendation request and a set of application server properties are received at a recommendation engine. The recommendation engine processes the recommendation request, and based on the set of application server properties, determines a recommendation on whether to submit the work request and/or whether to schedule the work request for a later time. Thereafter, the recommendation engine generates a recommendation notification that indicates whether to submit/schedule the work request to provide for a proactive approach to submitting the work request.