• RecurringMeetings
• 2014-11-19

Recurring Meeting of Cleveland Clinic - Stanford University

Date: November 19, 2014

Time: 4:00 PM EST

Means: Conference Call

Attendees:

1. Ahmet Erdemir (Cleveland Clinic)
2. Joy Ku (Stanford University)
3. Henry Kwong (Stanford University)

Agenda:

1. Update on SimTk.org upgrade.

2. Progress on cloud computing interface.
3. Open Knee(s) - Generation 1 on cloud computing.
4. Other.

Immediate Action Items:

• Stanford University (Joy & Henry)

  • Provide a written update in December about SimTk.org upgrade progress.

• Cleveland Clinic (Ahmet & Craig)

  • Start preparing Open Knee(s) - Generation 1 model for cloud computing.

Notes:

1. Update on SimTk.org upgrade.

   • The Stanford University team is moving along with the upgrades. Yet, the planned launch date during February may not be possible. The group will have a better idea of the timeline in early December.
2. Progress on cloud computing interface.

   • The focus has been on the upgrade of SimTk.org framework. There has not been much progress in the cloud computing interface. The team seems to have a good idea about what is required for an administration interface. Nonetheless, some models may have to be hard coded to start. Ahmet noted that the team should interact by e-mail to capture some of the conference deadlines to showcase the cloud computing interface. Ahmet will inform Joy for conference submissions and vice versa. Target conferences can be biomechanics oriented or those related to high performance computing. Submissions for the general approach on cloud computing infrastructure or its specialized case for the Open Knee(s) are possible.

3. Open Knee(s) - Generation 1 on cloud computing.
   • Ahmet has been working on upgrading and testing the Open Knee(s) - Generation 1 model to use with FEBio version 2. Ahmet described Joy and Henry the requirements of this model (2 GB memory, 1 hour simulation time using 16 cores, etc.) to see if the current infrastructure is adequate or not. It may be possible to lean towards XSEDE. Henry asked if it would be possible to have a library of solutions to provide, in case a simulation with a specific input set has already been completed by someone else. Ahmet reminded his intentions to do this, particularly for abridged result sets.
4. Other.

   • The group discussed potential mechanisms for management of access to cloud computing. Logistically, it may not be feasible to provide unlimited access. It may be possible to provide a flexible business structure to let exploratory simulations and a pay-as-you-go model for large scale simulations. Ahmet noted that a user should be able to download the model and run it in their own computer architecture. Joy also noted the possibility to provide downloadable virtual machines for this purpose. Active users of SimTk.org (those who contribute, share models, review models, document, etc.) may earn points to run models at the infrastructure. It is also a possibility to provide points for new registered users of SimTk.org to test cloud computing. There may be quotas for a given project and project developers and users may be able to donate cloud computing points. This may be similar to NIH's Data Commons business model. The group is not sure at what level the community will be involved. It will be an exciting social experiment to try out different operational models to provide this feature.