One of the ideas behind GFAL was also to have a POSIX like interface on top of SRM. Whereas SRM provides a FTP-like put and get functionality including space reservation and pinning etc., GFAL would give the user the ability to have a "Grid open, read write, close" functionality. Cheers, Heinz On Mon, 1 Nov 2004, Peter Elmer wrote: > [Add xrootd list] > > Hi Alvise, > > As far as I can tell GFAL is just one of these "I'm the interface layer" > things, whose main purpose is theoretically to shield the user from the > fact that various file access protocols may be used behind the scenes. > It doesn't replace xrootd (dCache, rfio/Castor, ...) but simply interprets > the protocol from the filestring and loads the right (real) software behind > the scenes. It looks like it also accepts LFN's and GUID's in which case > it has to look things up in a catalog before using xrootd/dCache/rfio... > > In ROOT (for example) the PluginManager plays more or less the same role. > Any real framework can provide such a thing relatively easily. (At CHEP Nobu > Katayama of Belle got in a nice joke about the complexity of the similar > layer that they have in their software.) > > Anyway, maybe I've missed something about GFAL, so I've CC-d the xrootd list > in case someone else wants to add something. > > > Pete > > On Tue, Oct 26, 2004 at 09:24:16AM +0200, Alvise Dorigo wrote: > > sono al ws grid-infn a bari... > > Ti sono arrivate voci about gFAL ? al CERN è spinto parecchio come > > sistema di accesso remoto a file (via dcache o rfio).... ci stanno > > lavorando molto. Spero che non sorpassi xrootd... > > > > Alvise > > > > ------------------------------------------------------------------------- > Peter Elmer E-mail: [log in to unmask] Phone: +41 (22) 767-4644 > Address: CERN Division PPE, Bat. 32 2C-14, CH-1211 Geneva 23, Switzerland > ------------------------------------------------------------------------- >