2021-02-10 FEB


continuation of last week's agenda ...

SAS BASE Plan B

Since we have decided on Plan B, I would like to continue our Prep meeting discussion of processing that can happen in GCP prior to downloading tables to SAS

Discussion points

  • API as an enabler and as architectural sanity

  • Martin: still important to know where SAS and RStudio will be for different analysts.

  • Hui (?): what are the boundaries of the Connect Platform. (question about "Cloud2")

  • The serverless conversation from an industry perspective: https://www.capitalone.com/tech/cloud/serverless-first-mindset-pros-and-cons

  • Gerlanc, Nicole (NIH/NCI) [C] to Everyone

https://cloud.google.com/bigquery/providers/simba-drivers
Gerlanc, Nicole (NIH/NCI) [C] to Everyone
see
SAS serving humanity since 1965 ;-)


BigQuery

Demo BD from R, and from SAS

The continuum json file, firestore, BigQuery

Testing with other JSON APIs such as AoU.

Post BigQuery Tables a better point of entry? As single files (G buckets, Box.com?) or index files (DBs) ?

WebAssembly (conclusion of preliminary tests)

Jeya / Jonas / ...

for discussion:

1 - Is there anymore a good reason to distribute containerized code? (No, wasm runtimes are containerized too)

2 - If some collaborators in a project require old-school command line executables can we in principe provide it to them? (Yes)

3 - Are there any major barriers, of interoperability or numerical support, to JS weaving distributing computing ecosystems contained by sandboxes? (No)

For wasm runtimes - for example https://wasmer.io

Rust Resources:


Data Platform

Jonas/ Bhaumik / Daniel / Lorena / ...

PLCO

Lorena / Jonas / Jeya / ...

https://exploregwas-qa.cancer.gov/plco-atlas/#/gwas/summary

Serotracker

https://observablehq.com/@episphere/sero

Time series

Daniel ....

Image Data Commons

Halo and other clients ... Praful

...