A technically interesting presentation related to Grid computing was the presentation by Alexander Reinefeld from ZIB Berlin. At ZIB they are creating a data management system, called ZIB-DMS, that works at a basic level: it tries to augment files with simple metadata typically atribute/value pairs to give information on what is inside a file. The metadata can then be used for searching on locating the file system.

ZIB is paying attention to integration with the existing file systems. So you can do things like "cd name=knuth" to go to a file that contains an article authored by Knuth. The system has some advanced data handling features, like collections, bi-directornal links and virtual files. It is compatible with current file systems through the NFS API and it is Grid enabled, in that it is scalable, and extensible. According to Reinefeld the system, although still experimental, works already. You can find more info on: http://zib/de/csr/