The SAM (Sequential Access through Meta-data) system is being built as a distributed cache management and data access layer for the D0 experiment at Fermilab. The project has taken an innovative approach to data management by designing and implementing the system, from the outset, as a distributed data server system deployable worldwide. The goal is to facilitate physicist access to data, while ensuring that ways of dealing with limited resources (such as robot, tape drive or disk availability or network bandwidth), are tightly coupled to physics groups and to experiment priorities and policies. Central to the SAM system is an Oracle meta-data catalog and database system that allows for sophisticated definition and selection of datasets, and for meticulous tracking of the entire data processing chain. The central Oracle database is expected to reach 0.5 TB in size as it tracks about 1 petabyte of D0 data over the next two to three years. Given these volumes of ! physics data, SAM interfaces with, and relies on the services of, a storage management system for the primary storage of all data. At Fermilab this system is Enstore, with alternate systems planned for at Lyon, France and Nikhef, Holland and Lancaster, UK. SAM deals primarily with data files and controls the storage, delivery, disk caching and presentation of files (to the analyzing programs) in various access modes: - raw data logging, farm production processing, monte-carlo data import, bulk organized summary data processing, condensed summary data analysis. One access also allows for individual event selection. Movement and replication of data files between different storage locations and disk caches is handled through communication between SAM's CORBA servers; this has some elements of a data grid.

Back to Program

Back to ACAT Web page