Implement S3 bucket and Snowflake architecture for raw PeMS data

[ian-r-rose]

Fixes #7, fixes #8

This creates a publicly-readable bucket (I can change that for development if folks feel nervous about that), as well as a service account with write access to that bucket. I've also created our reference Snowflake architecture here.

[ian-r-rose]

I should note: I've done everything here in AWS us-west-2 (both the S3 bucket and the Snowflake account)

Caltrans has expressed some interest in using Azure in the long-term, and we may tear down this infrastructure and rebuild it there. However, ODI is not currently set up with Azure, so in the short term I feel it's best to move forward with AWS with the knowledge that we can change that in the future. Does that seem reasonable @jasonlally?

[jasonlally]

Yeah that seems reasonable, what follow up should I have on the Azure request. Does that still need budget for approval?

[ian-r-rose]

Yeah that seems reasonable, what follow up should I have on the Azure request. Does that still need budget for approval?

I pinged the relevant issue. It sounds like we still will need a budget estimate