Extending NFSv4 for Petascale Data Management

Abstract

Designed with Internet data management in mind, NFSv4 is meeting the needs of widely distributed collaborations. Anticipating terascale and petascale HPC demands, NFSv4 architects are designing pNFS, a standard extension that supports parallel access to cluster file systems, object stores, and massive SANs. In combination with advances in optical networking, pNFS decouples computing from storage, even for petascale computations. CITI's GridNFS project integrates NFSv4 into the ecology of Grid middleware with critical virtual organization management: Globus support, name space construction and management, fine-grained access control with foreign user support, and high performance secure file system access for jobs scheduled in an indeterminate future. GridNFS and pNFS combine and integrate standard Internet protocols, promising near-universal compatibility with standards-compliant desk-top and enterprise network services.