Magda - Manager for grid-based data

Magda is a distributed data manager prototype for grid-resident data. It makes use of the MySQL open source relational database, Perl, Java and C++ to provide file cataloging, retrieval and replication services. For data movement, gridFTP, bbftp and scp can be chosen depending on available protocols. Third party transfers are supported. Magda currently catalogs 321K files with total size of 85.8 TB. It was successfully used in the Data Challenge 1 production for the Atlas experiment. It has been used to replicate more than 7 TB of data between BNL and CERN mass stores. We present here the design and implementation of Magda, and how it has been used as a file catalog and replication tool.