Grid Parallel Execution with Failure Risk Assessment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Grid computing systems face challenges in efficiently executing applications in parallel across multiple nodes due to varying failure rates and resource availability, which can lead to increased risk of application failure and suboptimal resource utilization.
Innovation Solution
A method is implemented in a grid network to maintain a database of application execution history and a list of capable compute nodes, allowing parallel execution of applications only when the risk of failure is below a customizable threshold, with automatic termination of execution on successful completion by one node.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If an application is executed in parallel on multiple grid compute nodes, then productivity is improved, but the risk of application failure increases due to varying failure rates and resource availability
Solution Approach 1:
The system performs preliminary actions by maintaining a database of execution history and capability information before parallel execution. The grid manager queries this database to assess node reliability and determine appropriate parallel execution strategies, preventing failure rather than responding to it afterward.
Solution Approach 2:
The system changes execution parameters dynamically by adjusting the degree of parallelization based on assessed failure risks. When risk is high, the system may limit parallel execution or select more reliable nodes, while allowing full parallelization when risk is low, thus optimizing both productivity and reliability.
2Productivity
If parallel execution is implemented across multiple grid compute nodes, then processing efficiency is improved, but resource utilization becomes suboptimal due to varying node capabilities and failure rates
Solution Approach 1:
The system applies local quality by matching specific application requirements with specific node capabilities. The database stores detailed information about each node's resources, reliability, and suitability for particular applications, enabling the grid manager to assign work to the most appropriate nodes rather than using a uniform approach.
Solution Approach 2:
The system enables self-service by allowing the grid manager to automatically query the database, assess node capabilities and reliability, and make intelligent allocation decisions without external intervention. The execution history and capability information in the database serve the system itself in optimizing resource distribution.
3Reliability
If execution history and capability information is maintained in a database, then reliability assessment is improved, but device complexity increases
Solution Approach 1:
The database serves multiple functions: storing execution history, maintaining capability information, assessing node reliability, and guiding parallel execution decisions. This multi-functionality justifies the added complexity by providing comprehensive reliability assessment and optimization capabilities in a single centralized component.
Data Source
AI summary
A method includes, in a grid network, maintaining a store containing a history of specific application executions, maintaining a list of grid applications in parallel, executing an application capable of executing in parallel on one or more grid compute nodes contained in the list, terminating execution of the application on the one or more grid compute nodes upon a successful completion of the application in any of the one or more grid compute nodes, and transmitting a result to a client.