Notes from team discussion (Ben, Craig, Mike)

 

EventStopStartContinue
DDAUnprepared organizers

Shared data

Shared services (geoserver)

Early engagement

Charging real costs

Being choosy

Supporting commercial providers

Don't underprovision

Slack is helpful

pre-allocated logins

github for event

GW

Running Jupyter in WB

Living in a bubble

More engagement

Being smarter about our market segment

Being good a JupyterHub

Embracing our uniqueness

Engage w/ Jupyter community

Cultivate community

Globus Personal in JupyterHub?


ESIP/INCOREDoing too much

Improve "Zonca" method

Support Rob's script

"Dask"

JupyterLab

KubeAdm

Using Kubernetes (don't move to swarm)


Things we do well:

  • Complex stacks
  • Training/education

Things we could do in next 8 months:

  • K8 1.8+
  • GCE/AWS support
  • Focus on NCSA/ISDA
  • Focus on Earthcube case


CHEESE

  • AWS/GCE/Azure
  • Requests with credits/votes and leaderboard
  • Docker security: authentication plugin, no root access, network isolation



  • No labels