2018-01-04 Meeting notes

Date

Attendees

Goals


Continue review of roadmap


Discussion items

TimeItemWhoNotes
Continue review of roadmap document

https://opensource.ncsa.illinois.edu/confluence/display/NDS/2017-10-19+Roadmap


6) Earthcube (Illya)

  • Suave earthcube building block. Load up simple dataset, it gives you a way to visualize the data http://suave.sdsc.edu/gallery/
  • Run some packages, analyze some data, how do you push back to the original data source
  • How does workbench work in the data management pipeline
  • Interaction between Suave and Jupyter and back to the original data

Relates to item 1) taking things out of workbench. 

Many challenges here:

  • Packaging up the data
  • Interfacing with different repositories

Universal interoperability pilot.  RDA has some interest in these things, but not sure on timelines.

Sandra working on REST webservices for pushing metadata back to repositories. DS3 API a layer on S3.   Spectra logic REST API for deep storage. 

https://www.slideshare.net/insideHPC/introducing-ds3-an-interface-to-deep-storage

TODO: Sandra to present description of what is being proposed. 


  •  
Deeper dive on point datasets index 

Dataset page includes affiliated (through pilots) datasets and datasets that are being hosted

These datasets are used to demonstrate capabilities. Learning from the needs of people we're working with. 

Datasets all different protocols. Different descriptions.  Very big. Etc.

Ideally would morph into the federated data repository. 

1 click button to access dataset in the workbench. Binder-like model. 

MDF - developing Forge package for working with MDF data. Workbench could host the forge environment. 

Dataset page is different from the FRDR project.


Re3data - repository of repositories

https://www.re3data.org/

Most of this is human readable. Would like to have machine-readable profiles that could be used to match data with potential places. 


NDS considered a repository recommender service: https://opensource.ncsa.illinois.edu/confluence/display/NDS/Data+Repository+Recommender



Prioritization goals
  1. Build the beginnings of FRDR-like model. Relationship to open storage network. 
  2. Need to support publication of data that doesn't belong elsewhere. Cobble together sites. Open storage network could again be useful. NDS can provide glue to make things work better. 
  3. Workbench sustainability model

Action items

Next meeting invite someone to explain FRDR

Sandra to explain proposal

Repository recommender

  •