VOResource 1.1: Mirrors?

Markus Demleitner msdemlei at ari.uni-heidelberg.de
Tue Jun 14 09:58:49 CEST 2016


Dear Registry WG,

I took away from our mirrors discussion that

(a) people want the concept of mirrors in the registry,

(b) modelling each mirror as a separate resource is not appropriate
for all the cases (in particular not for VizieR)

(c) there is no obvious candidate for metadata beyond the plain URL
of the mirror.

If you disagree with one of these conclusions, please shout.

Based on this, I've introduced a new mirrorURL element of type
xs:anyURI as a child of vr:Interface in Volute rev. 3444 (and 3445,
when I realised that mirrorURL should be xs:anyURI).

Apart from redaction in the text body, the change is that the schema
annotation for access URL is now:

  Although the schema allows multiple occurrences of accessURL,
  multiple accessURLs are deprecated. Each interface should have
  exactly one access URL. Where an interface has several mirrors, the
  accessURL should reflect the "primary" (fastest, best-connected,
  best-maintained) site, the one that non-sophisticated clients will
  go to.

  Additional accessURLs should be put into mirrorURLs. Advanced
  clients can retrieve the mirrorURLs and empirically determine
  interfaces closer to their network location.

and the mirrorURL is documented as:

  This is intended exclusively for true mirrors, i.e., interfaces that
  are functionally identical to the original interface and that are
  operated by the same publisher. Other arrangements should be
  represented as separate services linked by mirror-of relationships.

mirrorURL is just xs:anyURI rather than vr:AccessURL because mirrors
must be functionally identical and thus cannot have different @use
values (which perhaps suggests that @use shouldn't have been on
accessURL in the first place, but IMHO this is not a time to fix
this).

I feel a bit bad for doing this, as I don't see that any software
might use this by the time I'd like VOResource 1.1 go to REC, and
nothing to go to REC that's not been tested in use.  So, if a client
author volunteers, I'd be happy to work with you.

Otherwise: Opinions?  Proposals?

Cheers,

         Markus


More information about the registry mailing list