A framework supporting the development of a Grid portal for analysis based on ROI

Area of Science:

• Neuroscience
• Computational Science
• Grid Computing

Background:

• Brain function analysis requires simultaneous interactive and batch processing.
• Users need distinct resource allocation for region of interest (ROI) analysis versus entire dataset analysis.

Purpose of the Study:

• To propose a Grid portal for flexible resource assignment in brain function analysis.
• To integrate interactive and high-performance batch processing within a unified Grid environment.

Main Methods:

• Developed a Grid portal integrating interactive and batch processing.
• Implemented a job steering mechanism prioritizing interactive jobs based on user-defined priority.
• Utilized a priority-based result delivery mechanism to manage data significance.

Main Results:

• Ensured rapid turnaround times for interactive processing through priority-based job control.
• Provided quality of service (QoS) for interactive analysis, prioritizing user access to results.
• Achieved high-performance computation for Magnetoencephalography (MEG) analysis using batch processing on the Grid.

Conclusions:

• The priority-based job control mechanism effectively assigns computing resources based on user requirements.
• High-performance computation significantly advances brain science research.
• The portal offers users flexible access to substantial computational power for their analyses.