A Grid Service Broker for Scheduling Distributed Data-Oriented Applications on Global Grids

Title: A Grid Service Broker for Scheduling Distributed Data-Oriented Applications on Global Grids

Research Question: How can a Grid Service Broker be designed and implemented to effectively schedule and manage distributed data-oriented applications on global grids?

Methodology: The researchers developed a Grid Service Broker called the Gridbus Broker. This broker is designed to handle the complexity of grid computing by transforming user requirements into jobs that are scheduled on appropriate resources, managed, and results collected when finished. It has capabilities to locate and retrieve data from multiple data sources, select the best data repositories, and redirect output to storage for downstream processes.

Results: The Gridbus Broker was applied to the high-energy physics domain, specifically the Belle Analysis Data Grid. Experiments were conducted on the broker, resulting in successful deployment of Belle experiment data analysis jobs on a grid testbed with distributed resources.

Implications: The research implies that the Gridbus Broker is an effective tool for managing and scheduling distributed data-oriented applications on global grids. It provides a declarative and dynamic parametric programming model for creating grid applications, making it easier for users to access distributed datasets and computational resources. This can significantly enhance the efficiency of data-intensive computing environments and facilitate collaboration among researchers in scientific experiments.

Link to Article: https://arxiv.org/abs/0405023v1 Authors: arXiv ID: 0405023v1