[GFS-WG] RNS specification

[Andre Merzky]

Hi Christopher, 

that clarifies the issue, many thanks.  

We do not have any particular motivation to include symbolic
links right now, that is, none of our use cases mentions
them explicitely.  Its at this time considered to be a
'nice-to-have' feature.

Thanks, 

  Andre.


Quoting [Christopher Jordan] (Oct 06 2006):
> 
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> Andre,
> 
> I don't believe the RNS specification ever made any specific  
> provision for symbolic links, and there are no plans to provide  
> special handling of "symbolic links", but I know that in previous  
> versions of the draft spec, it was explicitly mentioned as an analogy  
> for an RNS entry that pointed to another RNS entry.
> 
> Basically, an RNS entry is just a human-readable name, organized in a  
> hierarchy of other human-readable names, which are associated with a  
> WS-Address-compliant address (which can be a direct reference to a  
> data resource or a reference to an abstract name, allowing for the  
> three-tier naming scheme used in OGSA). This can include a reference  
> to another entry in the same RNS namespace, which would be analogous  
> to a symbolic link in a standard Unix filesystem, or it could be a  
> reference to an EPR which is also referenced by another RNS entry,  
> analogous to a "hard link".
> 
> It's possible that the GFS profile for RNS will include operations to  
> create these kinds of RNS entries, to provide analogous functionality  
> to other Unix filesystems; I'd be interested to know how much  
> interest there is in this.
> 
> I hope this clarifies things for you, at least WRT how RNS  
> implementations can function to provide symbolic link-like  
> functionality.
> 
> Thanks.
> 
> On Oct 5, 2006, at 1:57 AM, Andre Merzky wrote:
> 
> >Hi GFS-WG,
> >
> >I seem to remember that the RNS specification did, at some
> >point in the past, mention symbolic links.  The last draft
> >(sent by Osamu on Sep 21st) does not mention links anywhere.
> >
> >Does my memory trick me? (very likely)
> >
> >Is there any intention to introduce special handling of
> >symbolic links in RNS, or in GFS?
> >
> >Thanks, best regards,
> >
> >  Andre.
> >
> >-- 
> >"So much time, so little to do..."  -- Garfield
> >--
> >  gfs-wg mailing list
> >  gfs-wg at ogf.org
> >  http://www.ogf.org/mailman/listinfo/gfs-wg
> >
> 
> Chris Jordan
> Data Reliability Manager
> San Diego Supercomputer Center
> ctjordan at sdsc.edu
> 858.534.8347
> 
> "If we had a keen vision and feeling of all ordinary human life, it  
> would be like hearing the grass grow and the squirrel's heart beat,  
> and we should die of that roar which lies on the other side of  
> silence." - George Eliot
> 
> 
> 
> 
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.4.1 (Darwin)
> 
> iD8DBQFFJY9XPCVtcXn6kg8RAkwvAKCqQoltNBkOz3C7/bfjmce3gWBbOQCgs8rN
> V77u1QCLykozcMX1qNvP3u8=
> =57ZK
> -----END PGP SIGNATURE-----



-- 
"So much time, so little to do..."  -- Garfield