Skip to content

March 17 2022

Jump to bottom
Zac Adelman edited this page Apr 14, 2022 · 1 revision

LADCO Cloud Modeling Survey Results

Discussion

  • Zac: working w/ parallel cluster. Perhaps in the future have a topic on each call. Perhaps present an application, tooltip, or trick. Practical information on what is working for them.

Priorities?

  • Volunteer on each call to share quick tip or trick
  • David healy: reiterate his priorities; Ability for modeling centers to share data without having to ship hard drives; Data security. Data being backup up; Presentations last week spoke to that
  • Margaret mccourtney: survey results? (Zac to share); As someone who will set up sys in future but hasn’t yet. State has contract w/ azure; Big hurdle to get over is to determine what best route to go is. If epa put all data on aws system, then cost a lot to download; Important for them to know to make decisions for those types of things; Cloud data centers in different locations
  • David healy: Semi-regular walkthrough of github site. Mention what is new, reminder of what’s there
  • Byeong kim: Discussion about bringing vendors to share what can offer?; AER. If anyone interested in working w/ them, that might be possible; AER could set up system, or they could provide space under their aqcast system
  • Zac: support from CSPs for sci computing; Support for linux admin support. Upgrading OSes, patching, managing user accounts; Told might be required to still hv jumpbox; don’t expect that would be allowed to have full control. Don’t want to discuss IT responsibilities too muchA
  • Zac: EPA work in this area. Looking at cloud to supplement internal modeling capabilities? Outward facing? things people would like to see shared via github? Eg., scripts
  • Byeong: would like to share, but
  • Zac: address joey Huang comments in github
  • Kristen foley: cmas has built step by step directions. Testing. Can share w/ this group and test
  • Liz: benchmark domain for two days. $10
  • Byeong kim: Include other models: moves? Blue sky framework. Learned that is running on linux. Might be a good candidate system to run in cloud
  • Zac: learned a lot from aws contact about recommendations for running wrf
  • Barron: Geoschem has put all data in S3 bucket thru open data. Running base simulation for any particular yr.
  • Liz: hv benchmark tutorial in an s3 bucket. CMAS center was granted thru open data program. Wl eventually use tt. Including annual met and emissions; Would be interesting to see hw cmaq south american colleagues could benefit from similar platform
  • Zac: also check survey results for ideas for priorities

Consolidated Ideas for Priorities:

  • Vendors talk about what can offer
  • Cloud service providers—e.g., what support they provide for scientific computing
  • Information about linux admin support
  • AER
  • Ramboll?
  • How set up systems in cloud
  • EPA vision for cloud and support for different CSPs
  • Ability for modeling centers to share data without having to ship hard drives
  • Data security. Data being backed up
  • GitHub walkthrough/updates
  • Compare local security requirements/approaches (e.g., requirement to have jumpbox)
  • Other modeling systems and tools: Blue Sky, Moves, AMET
Clone this wiki locally