Toward a Data Analysis Grid for Biometrics Research

Advanced research in biometrics presents new opportunities and challenges in grid computing. Biometric workloads are both data and computation intensive, and have the potential to be accelerated by employing large numbers of machines. Many large workloads follow a common high level structure and could be made robust and scalable by custom high level abstractions. However, these workloads also have non-trivial security constraints to ensure the privacy of participants, and present new problems in large scale storage management. In this paper, the authors give an overview of the needs of a biometric research group, and discuss the research problems in grid computing that result.