Hi Alan, On 28/08/15 14:40, Sill, Alan wrote:
While I think it is worth exploring new paradigms in this area, it is definitely the case that the EGI Federated Cloud makes explicit mention of using GLUE in all of its public talks, so keeping discussion and support going with them in this area would seem to be a good idea.
For what it's worth, the INDIGO-DataCloud project also includes explicit mention of GLUE in their design documents. I hope the approaches EGI and INDIGO-DataCloud take for service discover can be sufficiently similar that we can use the same cloud extensions within GLUE.
More troubling to me is that the Open Science Grid has apparently abandoned use of BDII and seems to be going in a new direction for its related tasks. OGF's role and interest in these areas is always just to provide a forum and arena to document and to encourage exchange of information among participants on best practices and methods, including development and support of standards where appropriate.
Is a discussion underway with OSG about its needs and current approach? Can we understand what led them to stop using BDII and GLUE (or did I misunderstand their current approach and get this wrong)?
At a somewhat higher level, should we restart or recommission a community or working group to look at overall practices in this area, with a focus on new technologies?
I think it would be good to attempt this, but I'm unsure how successful it will be. This move away from GLUE apparently took people within WLCG by surprise, but OSG people claimed they have been planning this for years. Certainly there's been a disconnect here: GLUE 2.0 was developed to overcome deficiencies in GLUE 1.3, yet a major user decided to abandon GLUE altogether rather than move. Cheers, Paul.