Thank you all for the valuable explanations. It does sound like a feature that should be handled at the gateway.
On Wed, Mar 19, 2014 at 3:20 PM, Miller, Mark <[email protected]> wrote: > I agree with Borries that this is a feature that each Gateway must have > complete control over. > > If I understand how Airavata/SciGap architecture is planned, the Gateway > retains responsibility for the database of results it generates. > In this context, it seems simpler to me to let the Gateway set/manage > permissions for its own users according to their needs. > CIPRES has records of data, tasks, and results that could be shared. > > That said, we have not implemented a way of doing that, it is something we > hoped to do. > > If there is a way of making sharing possible within Airavata > 1) without disturbing the CIPRES results DB, and > 2) gives CIPRES users control over it > 3) could be modified on a per Gateway basis > > It would be something to discuss. The description Borries gave is what I > would like to have in CIPRES, and intuitively, anyhow, it seems like > something that could be left to the Gateway. > > Mark > > > -----Original Message----- > From: Borries Demeler [mailto:[email protected]] > Sent: Wednesday, March 19, 2014 12:10 PM > To: [email protected] > Cc: users > Subject: Re: Experiment Sharing > > Dear Sachith, > > It is quite possible to share experiments through the UltraScan gateway, > and we have carefully thought about this problem. Of course, it is > important to recognize data ownership and to protect it as much as possible. > > In our gateway people can identify selected users of their own gateway > instance with whom they want to share their data. On a first level, this > only permits access of the analysis results, visualizations and metadata. > Another flag (=user-level) controls if they should have access to the > primary data. User levels are decided by the administrator of the gateway > instance. We chose to assign individual gateway instances for each > institution. Each also has their own MySQL DB backend, so data can never > get mixed up or misappropriated. > > So we allow the user pretty much fine grained control over who can access > what portions of their data. This is one case where you really want to > micromanage access rights to safeguard people's research data and possibly > proprietary information for corporate clients. > > I think this question is best handled by the specific implementation of > the gateway, and probably not necessarily something that should be handled > on the level of Airavata. > > Regards, -Borries > > > On Wed, Mar 19, 2014 at 02:56:18PM -0400, Sachith Withana wrote: > > Hi all, > > > > I'm exploring the use cases of allowing experiment sharing through > > Airavata. > > It would be wonderful if the science community can help me understand > > the real world use cases of Experiment sharing. > > > > I initially thought of having groups in a community and allow sharing > > within group(s) or make it public ( within the gateway). But it could > > be different. > > > > Is anyone using this now? and how do you do that? > > > > -- > > Thanks, > > Sachith Withana > -- Thanks, Sachith Withana
